GOTM scoring formula

[noughmaster]

This is a thread developed from conversations in the topic: "A Possible Scoring Anomaly". Its purpose is to discuss formulae for the GOTM score which reward people who win by conquest or spaceship, while also capturing differences in quality of gameplay of players who don't win.
Starlifter has suggested a step function (which btw is discontinuous by definition). It don't think continuity should be a prerequisite for a function, since we're thinking of dividing winners from losers/drawers, so a step function is OK.
I'll propose a set of properties for such a GOTM function. When we agree on these, we should be able to derive a class of functions. I don't know what treatment this topic has had in the past as I've only been contributing for a few days. If anyone could post references to these, it would be useful.
I know this topic won't be for everybody (can any topic be?) as it might get quite mathematical, but hopefully the contributors will get something out of it. I'm looking at it as an academic discussion. If the scoring system actually chages as a result, good, but that's a separate, political issue.

Properties:
(1) Divide the set P of all possible scores into 4 parts: scores obtained by conquest wins C, scores obtained by spaceship wins S, scores obtained by the game ending in 2020 E, scores obtained by AI spaceship win A. When the domain of the function is restricted to any of these single subsets of P, the function should by continuous and increasing. This means that if two people finish in the same way, the higher Civ score gets the higher GOTM score.
(2) A score from A cannot beat a score from E, which in turn cannot beat a score from C or S.

These properties are meant for discussion. We'll need a few more to derive a function.
Should C and S be treated the same (I think so)?
Should scores from the same domain be linearly related to the Civ score? The current GOTM function gives extra weight to an earlier finish. I agree with this. How convex should the function be? Maybe we could look at some spreadsheets here.

 

[starlifter]

I will quote some of the highlights from the thread "GOTM Stays a monthly game", and add some things to get all the thought and discussion into one spot.

This is mainly a discussion of the Global Ranking formula, which is more seriously flawed IMHO, so I wrote more about it than the GOTM base computations.

posted June 18, 2001 04:41 PM

The more I have thought about it the more I've realized it is crazy to include another person's performance into one's own Global Ranking (GR) score.

If we think about it... why should a new player enter a GR with a bias better than all those who have actually earned their GR position? A player should have only their own performance counted in a GR.

So as a new Civ Fanatics member and GOTM player, I am completely opposed to using a "median" player 's score, or a score of Zero, or any other score that the player did not actually earn, as part of the player's GR.

Perhaps the first GR could have a minimum 2 or 3 GOTM requirement. In pro tennis (to follow Smash's analogy), you cannot be ranked "globally" until you enter 3 pro tournaments. Check out Martina Hingis' performance before she was ranked. Imagine telling pro tennis players that new tour players would come into the rankins with a 70% bias to the median player on the tour.


So a suggested GR should be:

0.6 * (AVG(MAX [use top 2 of last 4 GOTMs])) + 0.4 * (Most recent GOTM)

NOTES:

1. The exact numbers could of course be tweaked, e.g., MAX[3 of last 5].

2. Other Medians or arbitrary scores are not used. A player would have to play another GOTM to use as the history before being ranked. Until then, the new player would not be ranked with a GR.

3. If a recent GOTM was missed, the GR score would not be impacted.

4. If a person played a GOTM with an alternate strategy, such as OCC (with the attendant low score), no worries.

5. As long as a person plays at least one "full" GOTM every couple months, the pressure would be off to play a long GOTM every month, and people could even pursue more non-traditional styles.

6. A "bad" GOTM (for a particular individual, that is) would not impact the GR.

7. A miss (or even a defeat) in the GOTM would not reduce the GR. It would take more than one defeat or miss to begin reducing the GR. This is particularly helpful if a person goes on vacation for a month.

posted June 18, 2001 06:26 PM by Matrix:

When you're new to the GOTM, you'll start with the median score of the Global Ranking, so that's in the middle of everyone: no disadvantage.

The disadvantage we're talking about is for people who missed a GOTM after they've already played one before, because then you drop in the Global Ranking.

It is still dead wrong, mathematically and intuitively, to use anything but the player's own performance in computing the GR. It is wrong to use Median, Mean, Mode, and/or Zero for historical scores when a new player enters the GOTM. Look at all the blanks (zeros) in the spreadsheet results. It is distortion on a massive scale for everyone exept the long-time players.

I thought about this but came to the simply conclusion that you cannot, or actually may not pick the best of certain GOTM results, because you just can't compare separate GOTM's. Really, I'm a bit reluctant in disagreeing with experts like starlifter or stormerne, but picking the best GOTM's...does not seem correct in any way, to me.

The main reason for using the top 2 of the last 4 is so a person can miss a GOTM and not destroy their GR. Or to allow the person to try a different strategy that may fail, or score low (like OCC).

Since everyone would have their best two historical GOTM scores used for the historical computation, the GR would reflect a player's actaul performance more accurately from month to month.

And starlifter, using the median score for people who are starting just, looks exactly the most reasonable thing to do. First of all, because it would be unfair to give starting people a disadvantage above others. But to fight your statement that it's crazy to include another person's performance into an individual score: you not using another person's performance, but the performance of every player who ever played, which is the exact way to insert someone into this "system", so to speak. What could be more neutral than the middle of the list?

Sorry Matrix, but your reasoning is mathematical hogwash. You pose a faulty delimna... you state an incomplete list of outcomes, all of which are unacceptable, e.g. "what is more neutral"... the question is not about neutrality, but about deteriming a player's cumulative GR. All these problems have been solved in the real world, like in sports. An I do not know of a single sport anywhere that uses medians of other players to compute the player's standing. Imagine if they did what you suggest in Golf or Tennis! Ugh!

It is my contention that the player's GR should be based solely on that player's own scores, not on the game score that a Mr. Median player (or Mr. Zero player) achieved.

An alternative is to not compute a GR for a person until they have at least two games of their own, for example. The newbie could be required to either play a GOTM for a couple months to get a GR, or else go back and play a previous GOTM so that score can be used for the historical part of the GR formula.

Almost everyone's concerns (posted in prior threads) can be addressed in one fell swoop if you use a max function, and take the top 2 score out of the last 4 games, to compute the historical part of the GR.

posted June 21, 2001 01:11 PM by Big Wheel:
quote:

Originally posted by Matrix:
...newbies start from the bottom in the Global Ranking?

-----------------------
Bitg Wheel:

Agreed, but reluctantly. I am against using the median, but also see shortcomings with using a 0 for a newbie's history.

Instead, why not just acknowledge that we have no history for that player and score accordingly.

For example, a (remarkably consistent) player who starts getting a normalized score of 10 every game will only have a global ranking of 8.8 after six months of playing. This happens because we have effectively said that the player scored 0 on his/her first game.

Instead for the very first game, why not say that that game's score is both GOTMscore and OldScore? That is:

NewScore = 0.7*if(OldScore=0,GOTMscore,OldScore)+0.3*GOTMscore.

Note that once a player has a history, the formula reverts to its current form. Our same consistent player will start at a10 ranking and will stay there. How 'bout it?

posted June 21, 2001 01:39 PM by Lord Oden
Bigwheel has a good point. I think that is the best way for "newbes" to get in to the ranking with out to much displacement from starting with a old score of "0" or "avg"

posted June 21, 2001 03:56 PM by Starifter:
Quote by matrix:
>Well, you could be right about the GR. A
>newbie might have to start with a score of 0
>instead of the median score. But my original
>question remains: what about people missing
>a GOTM? Do we have to be so harsh and just
>calculate with a score of 0?

No, I never suggested that newbies have a score of 0 used as their historical score when new to the GR. The choices are not either 0 or the median, LOL...

The way to compute a newbie's first GR is either:

A. as Big Wheel suggests, or

B. require at least one historical game score in order to obtain the first GR.

Mathematically, Big Wheel's suggestion is the same as using the newbie's first GOTM score as both the current and the past GOTM. This is both an acceptable method and an easy method to implement.

I personally favor taking it one step more and saying that a player should have at least two GOTMs submitted (even a past GOTM, like #4) in order to have a statistically valid (well, semi-valid) sample on which to derive the GR.

In fact, I am currently playing GOTM #4, which I will submit for a score when I finsh in a few days... But GOTM 4 is a much larger map than GOTM 5 was, so it may take a few days!


>But I still disagree about the
>space-ship-loss-end. I will use a finishing
>date of 2020 AD.

LOL, no worries... how many people actually lose the space race anyway? Probably almost never happens... I know I've never personally heard of anyone losing the space race.

>Everyone, do you agree letting newbies start
>from the bottom in the Global Ranking?

*** I disagree with inserting a 0 into the past score as much as using median score *** my reasons are the same as what Big Wheel outlined.

So my input is either use the current GOTM score as the past entry too, or require that a newbie also submit a second GOTM game to be used for the past entry.

posted June 21, 2001 06:06 PM by matrix:
No, no. That won't work - I've discussed that before. That's because that first game will be far too important. In your example for instance, the older GOTM has a 70% influence on the GR score, against 30% of the new one. A month later it's: 49% - 21% - 30%. A month later: 35% - 15% - 21% - 30%. After five months the influence of te first GOTM will be less than that of the last GOTM. Does this make more sense than the median score? Or zero? I'd say even less. I mean, you must be an idiot to send in your first GOTM while having a normalized score of 5. That is, if you give a damn about the GR in the first place.

posted June 21, 2001 06:43 PM by Big Wheel:

quote:

Originally posted by Matrix:
No, no. That won't work - I've discussed that before. That's because that first game will be far too important.


So how did you start the scoring for us players who have been playing since GOTM1? Didn't we get the same start where the first game got a 70% weighting in the second month?

The downside of the 70/30 formula that is in use is that you have to pick something to apply the 70% to. The only question is what. When a player starts you really have only one piece of personal data that works. Anything else is meaningless and shouldn't be considered.

Let's assume the following HYPOTHETICAL data:

Apr GOTM Score 50
May GOTM Score 0 (did not finish, vacation!)
Jun GOTM Score 40
Jul GOTM Score 4 (played as OCC)
Aug GOTM Score 70 (current most recent score)

In my earliest post, I suggested the formula:

0.6 * (AVG(MAX [2 of last 4])) + 0.4 * (Most recent)


In this formula, here is the computation:

0.6 * ((50+40)/2) + 0.4 * 70 = 55 (final GR)

As you can see, in effect, a 30% weight is given to each of the 2 older games, and 40% to the newest. Also very important, judging from the posts of those that voiced a concern, is that it is not essential to submit a high-scoring GOTM every month in order to maintain a relatively steady GR. Missing a GOTM, or playing it with a low scoring strategy, or even getting whomped by the AI would not sink the GR.

Personally, I think it is a good idea to allow a person to have a bad or missed game, hence the reason for using a MAX(2 of 4) function.


Here is an example for a new player:

Aug GOTM Score 70 (Only GOTM played so far)

Option 1: No GR (GR not computed until at least 2 historical GOTMs are on file).


Option 2: Use Current GOTM for historical data (Big Wheel's suggestion):


0.6 * ((70+70)/2) + 0.4 * 70 = 70 (initial GOTM)


Option 3: Use 0 for historical data:

0.6 * ((0+0)/2) + 0.4 * 70 = 28

IMHO, this is a needless penalty for new players!


Option 4: Use Median for historical data (assume 20 is median):

0.6 * ((20+20)/2) + 0.4 * 70 = 40

Once again, not much reflection of reality for the new player, and arbitrarily assumes the player did historically better than half the players who have entered the GOTM.

So those are the basic options. Since the best solution is for a player to have at least one (or two) historical games that can be used in the GR computation I have outlined, the new player should either wait for at least 3 months (until they play several GOTMs), or else download, play, and submit a prior GOTM.

Naturally, if a player chose to play a recent prior GOTM for a basis in the GR computation, the player would not be eligible for any award or recognition for playing a prior month's GOTM, nor would their score used to adjust the original standings... it would be simply used to establish the baseline historical data for a new player's GR.


Anyway, that's my input... you can of course adopt any GR method you desire. At a minimum, however, I'd recommend to use something that:

(1) does not penalize established players who miss a game (or lose/have a low score) from time to time and

(2) neither penalizes nor rewards new players, esp. on an arbitrary basis.

The current system evidently hits a player hard for missing a month of GOTM, and using a Median score (or zero) for new players is quite arbitrary, esp. with a 70% weight.

[This message has been edited by starlifter (edited July 06, 2001).]

 

[noughmaster]

Starlifter,
I agree with your assessment of the global ranking formula. The general technique proposed is called exponential smoothing.
Let's says we have a formula

GR(n) = t*S(n) + (1-t)*GR(n-1)

where S(n) is the nth game score.
It's wrong to use 0 as GR(n-1) for new players because the 0 score will forever affect all future GR's. The only sensible way is to wait two games.
If we want to have all newer results more important than old ones, we must have t > 0.5, although this places a lot of emphasis on the most recent game.
I think exponential smoothing, which takes into account every single past game isn't the best method. Perhaps something like the tennis (as you mentioned) where game influence has a finite life is better.

 

[starlifter]

Perhaps something like the tennis (as you mentioned) where game influence has a finite life is better.

Using a sliding "window" of say, the most recent 5 GOTMs, is the best reasonable way.

My simple (but strong) suggestion is to look at the most recent 5 GOTMs, and use the best 3 results. Accurate, simple, fair. Trival with a spreadsheet. Easily understood by all, quite sensible, and also a very common approach in "real" life for similar computations.

Take those best 3 of 5 GOTMs, average them straight up: (1st Best + 2nd Best + 3rd Best)/3

This is the GR!

No "weighting"... Simple, stable, fair, time-tested.

(Note: Completely discard the remaining two.)

Discarding the lowest 2 scores of 5 solves all (count 'em... all) the numerical complaints and issues that the GOTM participants have voiced in posts.

It's wrong to use 0 as GR(n-1) for new players because the 0 score will forever affect all future GR's.

LOL, how very true. So painfully obvious that I had assumed it was a truism! <IMG SRC="http://forums.civfanatics.com/ubb/smile.gif" border=0>

The only sensible way is to wait two games.

I think it could be OK to use 0's for new players, **IF** a Sliding Window was used (e.g., take the Best 3 of the 5 most recent GOTMs).

Further, a new player should be able to submit prior GOTMs for historical computations (not eligible for GOTM awards or recognition or re-writing of prior GOTM results).

If we want to have all newer results more important than old ones, we must have t > 0.5, although this places a lot of emphasis on the most recent game.

Since there is no logical imperative in such an informal computation as a Civ II GR that we need a more complicated weighting, I agree that we don't need a recent-game emphasis. Nothing really justifies emphasizing the last game, if you take the best 3 of 5. Remember, there is no real-life money, promotions, lives, etc. at stake.

As others have noted, the only people that ever really have a chance for a decent GR score (in the current system) are the ones that have been doing GOTMs for the longest. Numerically, equal players can never achieve an equal GR, unless they have an equal number of GOTM submissions.

If a person plays a single GOTM as an OCC, their GR is trashed forever (look at Smash's GR). A single "bad" game, or the birth of a child, or a vacation, or an accident will forever wreck the GR, as it is now computed. Might as well rename the GR to the OBGR (Old Boys' Global Ranking), because all new players are forever shafted, and prevented from ever having a contending OBGR for all time (assuming the Old Boys continue to play).

Let's restate the obvious: If you are not in the top few players right now, you never will be no matter how good you are, or how much you improve... as long as the current players at the top continue to submit decent game scores. In tennis, it would be like telling all new players you cannot move up until the top players get out (as long as the top players just continue to do "OK"). In toher words, what matters the most is senority, not skill.


In real life, we rate the performance of our enlisted Military people numerically. We use a standard "sliding window" (with a weighting for more recent reports) when computing the single number that represents the "average". Reports outside the "window" have no effect, numerically. A simplified version of this process is very similar to the way the GR should be computed.

As can be seen from posts over the last couple months, unfortunately, some people have become discouraged because of low scores, poor performances, etc. Though the GOTM is a fantastic idea (and has a defect in it's computation, too), the Global Ranking is currently computed in a flatly illogical and unfair manner. Ironically, the current GR computation penalizes almost everyone to varying degrees.

From a meaningful numerical analysis perspective, the current GR yields pure rubbish. It is just plain silly to compute the GR as it is currently done (no personal offense to Matrix or Thunderfall)... the fix is easy, and should be done with deliberate haste.


EDIT: Added info, fixed spelling.

[This message has been edited by starlifter (edited July 09, 2001).]

 

[noughmaster]

It's wrong to use 0 as GR(n-1) for new players because the 0 score will forever affect all future GR's.

LOL, how very true. So painfully obvious that I had assumed it was a truism!

Unfortunately, it seemed to be the topic of a fair bit of discussion among a few people. I suppose this illustrates the problem with formulae: many people don't really understand how they work (especialy if they are recursive) and may feel excluded.
I agree 100% with using an average of the best of the last few games. The bulk of discussion required is do we take 3 of 5, 4 of 6 etc.? This could be decided on with a poll.

Time periods like this take account of recent form well enough without any need to weight the individual games differently. People can also feel more equal participants because they can understand the process fully.

I think it could be OK to use 0's for new players, **IF** a Sliding Window was used (e.g., take the Best 3 of the 5 most recent GOTMs).

Yes, I agree. New players could start off at the bottom and work their way up the list, exactly as in the tennis. It wouldn't take them long.

I know Matrix and Thunderfall put a hell of a lot of effort into this site

and I hope they don't take it the wrong way when people make strong criticisms (especially new people). However, those who can see mathematical problems in ranking systems which affect the enjoyment of the whole venture need to point them out and try to fix them.
If we can get Matrix and Thunderfall onside, do you want to start a new thread which acts as a poll to change the global ranking formula to the simple averaging method?

Also, what do you think of not using the Civ Bonus Score for early finishes in the GOTM score. I didn't know about this and got 14 instead of 78. Although I'll get over it

, if there's going to be a competition, people will compete and get upset if they don't see their result as a fair representation of their performance.
I also think the GOTM score should use the Civ percentage, not the base score, as low level games will have more influence on the Global Ranking (by any calculation method).

Should we get political and seriously campaign to change the scoring system?

It needs improvement to be fairer. I'm willing to do a bit of work with data etc. if Matrix and Thunderfall don't have time.

 

[noughmaster]

Two things I forgot to mention in my previous submission:

(1) With regard to using best 3 of last 5, best 4 of last 6 or whatever: I've seen comments on other threads that it wouldn't be a true indication because people can not count bad games. It's hard to reconcile this with the situation where genuinely good players might not have time to submit some months. We could use best 4 of last 5 or change the rules to allow early retirement (although this last suggestion has a problem because people wil retire before a war which they lose). I don't think early retirement should be allowed. I suppose if you don't submit often enough, your ranking will drop. You can always pick it back up again, especially if we don't use too many (or any) large maps.

(2) Conversations about the Global Ranking don't require much mathematics if we use the straight average method you suggested (or even just the sum). People can vote on alternatives, the consequences of which they can easily understand.
For the GOTM score, detailed discussions will only be accessible to a few people. However, people could vote on a series of CRITERIA (see first topic in this thread) from which a GOTM formula could be derived. Everyone can discuss the fairness of the criteria if we present them in non mathematical language.

 

[starlifter]

...
Should we get political and seriously campaign to change the scoring system?
...

LOL, I'm not interested in making a big campaign or anything along those lines, because this stuff is really up to Thunderfall and Matrix. I may not agree with some things, but hopefully they will eventually rethink things on their own, or maybe get some help from the help in these threads. But if they don't, I'll live.

The GOTM (and GR) is an informal thing; no lives or money or careers are at stake, LOL. Yes, the current system has some flaws that are glaring and profound, yet I'm actually more amused than 'upset'.

The funny thing is that the current system will favor players such as myself over the long run, but fariness dictates I continue to point out the ways to fix it, even if Matix is in denial. It is a shame, because even though most people don't really have a full handle on the various mathematical implications, many posters have posted varying complaints... and all of them stem from the illogical and mathematically disheartening way the 'Informal' GOTM is set up.

The stated goals of the GOTM are in extreme conflict with the "death penalty" method that the GR currently uses, and the funny thing is many players know or suspect something is wrong, but can't really express what it is... and they probably don't read boring math threads.

Another way of stating the problem is this: the current GOTM GR system is Drakonian... it pummels anyone who is in one of the following categories:

1. Misses a GOTM.
2. Loses a GOTM.
3. Gets low score in a GOTM.
4. Tries an OCC.
5. Is a new CivFanatics member (since April).
6. Goes on vacation (and rushes/misses a GTOM).


It is my bet that many players are in at least one of those categories. If so, bend over and take it with a smile, because the GR is going to give you the shaft.

The simple and fair way to kill all unfairness with one fell swoop is to take the best 3 of the last 5 GOTMs. Stunningly simple, amazingly elegant, supremely easy to implement, and it addresses and solves all the shortcomings of the GR.

As long as the GOTMs are informal and meant to help, learn, and exchange ideas, it is nuts to ruthlessly punish someone who misses a GOTM. Absolutely insane.

No one wants a crappy score, especially one that they can never have the slightest hope of ever erasing or recovering from.

The scoring method should focus on POSITIVE results, not exaggerate and inflate negative results. Occasional missed/bad games should never be allowed to rape

the GR like they currently do. And all new player are pre-raped for life, with no hope at all under the current GR method.

So if for no other reason, the GR needs to be completely redone (using the 3 of 5 method) to encourage, not DISCOURAGE, players... especially newer players, and players that are trying to improve.

So bottom line, no, I don't want to start a big campaign; but I will respond to threads and comments

, and continue to make inputs along the way. And of course continue to point out that which is logical and fair, and defend the newbies and underdogs

.

EDIT: Added & Spelling

[This message has been edited by starlifter (edited July 11, 2001).]

 

[noughmaster]

I was being a little bit facetious when I said "political campaign", but I'd seriously like to encourage Matrix and Thunderfall to endore a sort of "official" poll thread where people could vote on the GR scoring system they wanted (after some discussion so everybody understands what they're voting for).
It's easy (especially for academics like me) to get carried away with theoretical issues, however there is a real and very important issue with the scoring system: the participation of the community attached to this website.
I've enjoyed contributing to and learning from these forums. It would be a shame if new members lost interest and stopped contributing because they felt the system was inherently unfair and an old boys' club.
Many people, while admitting they believe the scoring system to be unfair, also say "I don't really care about the GOTM score Global Ranking". This is true in a sense, because no-one is going to kill themselves over it (I hope!). However, if people don't feel they're getting fair reward for effort, they'll stop participating in the GOTM, from which will quickly follow not participating in the forums. I'd hate to see something which I'm glad I found and have enjoyed go downhill.
I'm sure from the effort they've put in to this site, Matrix and Thunderfall would like to see a vibrant Civ forum community. If this is to happen, the community needs to feel that the site is partly "theirs", so when people point out something which is obviously wrong and how to fix it, we need to actively encourage Matrix and Thunderfall to do something about it.
In summary, I don't want to see new members or members who don't get high scores discouraged and lose interest. Matrix and Thunderfall, please take Starlifter's, mine and others' suggestions to heart and start a poll thread to change to scoring system. I think we all have a common interest and goal here.
Best wishes.

 

[starlifter]

posted July 12, 2001 12:20 AM
I was being a little bit facetious when I said "political campaign", but I'd seriously like to encourage Matrix and Thunderfall to endore a sort of "official" poll thread where people could vote on the GR scoring system they wanted (after some discussion so everybody understands what they're voting for).

Evidently, they already did that, sort of... way back when. A few weeks ago, a thread was asking for input, but most suggestions were discarded. Although the current GR system is greatly flawed, I agree that someone at the top must decide. I really don't think running life with polls is a good option, so even though I'm apalled at the GR computing method, I agree with the summary decision making process. The buck should stop with Matrix, not a poll. So since Matrix has at least read the way the GR should be done (even though he discarded it), the decision should still be his.

LOL, who knows, one day he may read a statistics book or look at some real life examples and suddenly realize the right way to compute the GR for an "informal" GOTM that is suppoesed to promote gameplay and discussion.

I've enjoyed contributing to and learning from these forums. It would be a shame if new members lost interest and stopped contributing because they felt the system was inherently unfair and an old boys' club.

I think the GR is integral, essential, and entertaining. Personally, if were able, I'd simply fix things and quit rationalizing why things can't be fixed, like Matrix has done in other threads with mine and others' suggestions. This whole issue is so trivial to fix, but the real crux seems to be a difference in stated GOTM philosophy (see main GOTM page), and the "punish the player" approach for computing the Good Old Boys GR.

As anyone can see reading the posts of others, some folks are already discouraged and talking about not playing GOTMs. Matrix actually told one guy that someone had to finish last, and it may as well be that dude, LOL! Not exactly a great encouragement, and of course many others simply read that stuff and make no post or comment.

When one looks at the balance of GR's on a scatterchart, it is quite obvious the GR system is whacked. It is ironic that I personally favor awarding a score of 0 to anyone who loses (std def), and Matrix would prefer to give the loser a significant number of points. On the other hand, I do not want to count the occasional loss or low acore against a player, but Matrix wants to pummel them mercilessly, and for life.

In effect, this method tells a player that if you ever miss a game, or are new to CivFanatics, or if you ever play OCC or score low, your GR is ruined for all time (assuming the regular top players perform at least average).

GOTM Rules page:

"GOTM is meant as a casual contest."

Assuming a player would like a nice GR, there can be very little that is casual about a contest where one sub-par performance sneds your GR down the toilet for life.

In summary, I don't want to see new members or members who don't get high scores discouraged and lose interest.

What you dread has already occurred and been noted by Matrix, judging from some posts.

When edge became discouraged and realized he should not have even entered (poor score) and decided to skip GOTM 6:

From Matrix, to Edge posted July 07, 2001 05:46 PM:

"... Someone has to be last! ..."

Deftly handled, LOL!

Many people, while admitting they believe the scoring system to be unfair, also say "I don't really care about the GOTM score Global Ranking".

People may say that sometimes, but no one wants to see a hopeless GR. Everyone looks at them, and it looks pretty crappy when some score 50 and 100 times more than others.

However, if people don't feel they're getting fair reward for effort, they'll stop participating in the GOTM, from which will quickly follow not participating in the forums. I'd hate to see something which I'm glad I found and have enjoyed go downhill.

Wise words which will probably go unheeded. The GOTM will continue, but participation will not be as broad as it could be, and some will lose intrest when they realize they are still going to be permanent losers (in scoring), even if/when they improve. And to me, a good thing about the GOTM is helping others improve their own game.

... the community needs to feel that the site is partly "theirs", so when people point out something which is obviously wrong and how to fix it, we need to actively encourage Matrix and Thunderfall to do something about it.

Matrix has said that Thunderfall does not even like the GR. But the GR is a good thing, and could simply be fixed to reflect the supposed relaxed nature of the GOTM. There should be no "pressure" to turn in a game, much less a "top" game, every month. The way to do this is to count 3 of the last 5 GOTMs.

3, because 3 is a good minimal number for a meaningful statistical average. 5, because would only allow one variation in gameplay (like OCC or fast finish) or miss every 5 months. 6 would be too long, but 6 might work if it was a 4 out of 6. However, this means new players must play for 4 long months before even being theoretically able to attain a full GR score. So mathematically and considering human nature, 3 of 5 seems a very good balance, and hence that is what I support.

Well, I doubt this thread will be read by many GOTMers, but at least it is here for those that aren't daunted by the hint of math and formulas

.

 

[shadowdale]

Well I just thought that I would say a few words in what seems to have become a talk between two people who basically agree!!!

I agree with the fact that it seems rather harsh to reduce the GR with 30% if a player misses a game or by another large amount if he/she plays a bad game or a OCC - I would say that a 10% reduction if you don't play would be more fair. As to what should be done about players who play OCC and therefor score lower than they would have had they played a "regular" game - well I can't really see any satisfactory solution to that problem.

But concerning what the historical GR for a new player should be, well I must say that I think that everybody should begin in the same way, with 0 as the historical GR. You might say that this affects new players unfairly! But after they have played some 6-8 games their GR would most likely be the same even if the historical GR had been the median or the first game!!! The system is made in a way where every new game doesn't mean all that big a change - well in the beginning it does but after having played some 4-6 games the GR doesn't change all that much!!!

But I do think that it should be possible to see how many games a player has participated in, in order to get a better feel for his or her GR - so I suggest that we add a new feature to the GR which says how many games a player has played.

And one last thing, I think that it would be cool to add one more feature to the GR, and that would be the "all time high" of each player - so even if you take a 3 month break and you GR drops accordingly you would still be able to see how high a GR that players has had!!!

------------------
We are species 8472 - assimilation attempts are futile - the weak shall perish

No wait we are species 5618 and we got beer...... don't harm us!!!!!!

 

[starlifter]

posted July 12, 2001 06:49 AM
Well I just thought that I would say a few words in what seems to have become a talk between two people who basically agree!!!

Very glad to have more discussion! I figured no one else even looked at this thread because of the boring nature of it.

I agree with the fact that it seems rather harsh to reduce the GR with 30% if a player misses a game or by another large amount if he/she plays a bad game or a OCC - I would say that a 10% reduction if you don't play would be more fair.

Mathematically, there is no reason to penalize a players at all. Philosophically, it just plain stupid, given the fact this is "supposed" to be an "informal" fun game. Just to be clear, this rule does not affect me since I will not be missing any GOTMs (fingers crossed, LOL!). I streneously argue this point on behalf of players who cannot or do not, for whatever reason, submit a GOTM every month.

As to what should be done about players who play OCC and therefor score lower than they would have had they played a "regular" game - well I can't really see any satisfactory solution to that problem.

I've already made a viable suggestion (inspired by Matrix, ironically!) to alleviate this somewhat. Give a category "star" award. Under the point system, this would give a small recompense for this style of play. But you are right, IMHO, there is no fair and logical way to make OCC viable GOTM score-wise, against a big "normal" style of play. But my point is that we should not *penalize* an OCC (or a skipped month).

But concerning what the historical GR for a new player should be, well I must say that I think that everybody should begin in the same way, with 0 as the historical GR.

As currently implemented in the existing falulty GR, that is illogical in the extreme... both mathematically and in common sense... because you can NEVER overcome this albatross! It is a permanent penalty, and lifetime stigma. And it reeks.

Ironically, I favor using 0 for the score in a new player's unplayed games, but those zeros are rapidly erased as the new player finished a 2nd and 3rd game. At the 3rd and later games, a new player is on equal footing with all Good Ol Boys.

Once again, to be clear, what I'm advocating is what's fair for everyone, especially the currently raped new folks. And BTW, in my own case, the Matrix's flawed current system will be to my personal advantage as time goes on. But it's not a fair system, so it should be changed. <IMG SRC="http://forums.civfanatics.com/ubb/wink.gif" border=0>

The fair (and stunningly self-evident) solution is to institute a sliding window of games that are used to compute the GR. Here is how:

1. Examine the last 5 GOTMs of each player.
2. Use their top 3 GOTM game scores.
3. Use straight mean ("mean" is a specific mathematical method of averaging) of those top 3 results.

====== EXAMPLE ========

For a brand new player, here is month one with a GOTM of 60:

(0+0+60)/3 = GR of 20

Month 2 (GOTM=90):

(0+60+90)/3 = GR of 50

Month 3 (GOTM=75):

(60+90+75)/3= GR of 75

===== END EXAMPLE ======

As you can see from my example, I actaully do think the "fair" way is to use zero for new players, but zeros are replaced as the player submits completed GOTMs. In Matrix' faulty system, a player never rids him/her self of the unfair influence of the "Good Old Boys" penalty for not discovering and playing GOTMs from the very start.

You might say that this affects new players unfairly!

Even if I say nothing at all, the unfairness of raping new players remains. <IMG SRC="http://forums.civfanatics.com/ubb/cry.gif" border=0>

But after they have played some 6-8 games their GR would most likely be the same even if the historical GR had been the median or the first game!!!

The entire current computation method is illogical and would not be tolerated in "real" life, if applied to other things like pro sports or personnel evaluations or systems design. The field of mathematics is mature and vast, and the silly thing (I'm not trying to create a schism here) is that all this is so obvious as to be a truism for those with some related professional experience or university training. Or even just good plain common sense.

The system is made in a way where every new game doesn't mean all that big a change - well in the beginning it does but after having played some 4-6 games the GR doesn't change all that much!!!

Of course the score does not change a lot, because the new player is pre-shafted with scores of zero in 70% of his/her GR. However, the purpose of a GR is not to make it unchanging, LOL!! There are other (even more stupid) ways to do that.

Nay, the objective of developing a GR should be to have it reflect the relatively recent performance of the player, and not Matrix' mathematical exercise in averaging lots of zeros into a recursive formula.

In the current flawed form, the GR will change wildly if a player missed the current GOTM. 30% of the GR will simply and suddenly be Zero, and the GR will plumment. And it can never fully recover, either.

But I do think that it should be possible to see how many games a player has participated in, in order to get a better feel for his or her GR - so I suggest that we add a new feature to the GR which says how many games a player has played.

A nice idea! At least this way, eveyone can see just who is raped the most.

And one last thing, I think that it would be cool to add one more feature to the GR, and that would be the "all time high" of each player - so even if you take a 3 month break and you GR drops accordingly you would still be able to see how high a GR that players has had!!!

A very good idea; this is done in the professional world, too.

Under the current GR, a player that takes a 3-month break will come back ridden hard and put away wet. The equivilent of a gang-rape with the police watching and laughing.

A player should not be punished in the long term for taking a break. Using your example, a player with a 3 month break might have scores such as these:

0 40 60 25 0 80 30 70 45 0 0 0

Before a the player comes back and plays a game, the GR would be:

(0+45+70)/3= 38.3

After the player submits a GOTM of say, 80, the GR would be:

(45+70+80)/3=65

And bingo, recovery complete. The rape of the current method does not exist! And all scores used in the GR were actually played and earned by the player! <IMG SRC="http://forums.civfanatics.com/ubb/goodwork.gif" border=0>

EDIT: Fixed quote boundaries.

[This message has been edited by starlifter (edited July 12, 2001).]

 

[starlifter]

For long-term consolodation, I am going to summarize a short thread started by Cornmaster:

Topic: Hall of Fame Question. CornMaster Moderator

posted July 07, 2001 12:35 PM
I got a question about how you do the scoring for the Hall of Fame???

In my opinion Catcus Pete should be higher than Smash (Nothing against you Smash). Even though Smash has a Gold, Pete has 2 silvers, and star too..... so I would think he should be higher.

I would do scoring something like this:
Gold = 3 pts
Sliver = 2 pts
Bronze = 1 pts
Stars = .5 pts
Who ever has the most points would come 1st, then 2nd, etc....

So what do you do???

--------------------------
REPLY BY: starlifter
posted July 07, 2001 01:14 PM
Your concept is a good idea, IMHO! I can already hear the "No No No's" from Matrix, LOL .

However, the specific progression you suggested favors lower performances in general, and unnecessarily involves fractions .

With a point sysem, a function such as ((prior value)*2+1) might be more appropriate.

To keep it brief, this is how such a function would look when translated into points:

1 - Stars (high score, fastest finish, etc)
3 - Bronze
7 - Silver (three bronze surpasses a silver effort)
15 - Gold (three silver surpasses a gold effort)

Note the pattern. It takes over two of any prior award to equal or exceed the value of the next higher award. This rewards good work, and in practice means that it is better to win a game every other month, rather than always finish second.

Here is another progression that I personally find to be a better balance overall (key difference is that the bronze is worth 2 points, hence 2 stars equals a bronze effort):

1 - Star
2 - Bronze (two stars equals a bronze effort)
5 - Silver (three bronze surpasses a silver effort)
11 - Gold (three silver surpasses a gold effort)

I might suggest we 'lift' the number of the "star" awards (is this somehow a pun with my 'starlifter' name? ). For instance, two new star categories (value: one point) might be strongly considered:

- Fastest OCC finish (since OCC is a game of speed, not score,no award for OCC high score is necessary)

Cornmaster has an awesome point, and excellent idea, which can be easily tweaked for reasons in my reply.

Using a point system properly will add credibility to the Hall of Fame, and encourage a variety of gameplay, like OCC and Fast Finishes, even if such strategies do not gain the highest GOTM score. I am still revolted at the thought that an excellent game like Smash's GOTM 4 OCC is treated like a resounding loss, score and recognition wise. With some easy tweaks, the HOF can be drastically improved. Thanks, Cornmaster, for "thinking outside the box" and pointing the way!

SUMMARY:

Points awarded for computing the HOF standings:

1 - Star
2 - Bronze (two Stars equals a Bronze effort)
5 - Silver (it takes three Bronze surpasses a silver effort)
11 - Gold (need three silvers to surpass a gold effort)

New Star Categories:

1. Fastest finish.
2. Highest score.
3. Fastest OCC finish (New!).
4. Survivor Award (New!).
5. 1500 AD Retirement (New!).
6. Lo-Boy (New!).

Notes on #3:
- Since the purpose of OCC is to finish in the earliest possible year, an additional category fo high scoring OCC award is not neessary

Notes on #4:
- This is for the best "Survivor" (highest non-victory score) in a given month.
- Players must survive until 2020!
- If the AI eliminates a player, they are not eligible for this award.
- This will primarly be a consolation award for those who cannot win via a Spaceship or Conquest by 2020.

Notes on #5:
- This is for those that cannot or do not choose to finish a GOTM.
- If someone submits a game in which they retired in 1500 AD (exact year, not approx), they are eligible for this one-point Star award.
- A 1500 AD retirement is not eligible to win any other category or medal.
- A player cannot submit a 1500 AD retirement plus some other game ending; it is one or the other.
- The purpose of this is to recognize, in some small way, the efforts of those that wish to play fast and furious, and get it done quickly.
- A 1500 AD retirement is not treated as a victory, and such a GOTM score is still offically computed using 2020 AD.

Notes on #6:
- This is for the lowest-scoring player who still achieved a standard victory with more than one city (landed SS or conquered world) in a given month.
- Players must achieve a Civ II Victory!
- If the AI wins, the player is not eligible for this award.
- If the AI lands a SS first, the player is not eligible for this award.
- The player must have more than one city; this is not an OCC strategy award.
- This will primarly be a consolation award for those who can struggle to a win, but do it with a LOW score!.

 

[Matrix]

Thanks for the summary, starlifter. I don't feel like reading all this stuff if you don't mind. Even if I'm the moderator here. <IMG SRC="http://forums.civfanatics.com/ubb/wink.gif" border=0>

First of all, I must say I'm really into shadowdale's ideas.

Secondly, I am however, quite conservative. And with reason: we can't keep changing the rules on and on. Then the meaning and the values will fade away...<HR>Anyway, a price for the fastest OCC finish would be a good idea. Only there are two problems: first of all, it's inconsistent with the GR: a OCC type of game is not good for your GR score... Second: until now there has only been one player (Smash, GOTM IV) who played a OCC game. Do you think that's going to increase?

I do not like the other prices for the following reasons:
Survivor Award and Lo-Boy: those are just results within the scale of all the results. I don't see why someone who barily won or who survived the best way should be awarded; doesn't make sense to me. I understand the fun of the idea, but basically it's nonsense.

1500 AD Retirement: I don't want to include a pre-retired game in any possible way. That's harsh, but I've excluded all my unfinished games as well, with reason.<HR>Shadowdale's ideas:
I will show how many GOTM games people have played in the past.

Concerning the decreasion of the GR score with a lower percentage then 30% when they haven't played a GOTM: what if (s)he got killed? Should the GR score decrease with (almost) 30% then? And if not, is it fair that someone who survived till 2020 with only a few cities decrease more in the GR then people who got killed (or didn't play in the first place)?

When inserting the GOTM results into the GR spreadsheet, I suddenly got the idea as well that it's not...right to let newbies start with a certain score - whatever score (in this case the median score). While I was against this at first, I have now changed my opinion. Now, because CivFanatics is a dictatorship <IMG SRC="http://forums.civfanatics.com/ubb/groucho-marx.gif" border=0> <IMG SRC="http://forums.civfanatics.com/ubb/lol.gif" border=0> and I make the decisions - seriously, I jumped in when GOTM started, because I just found it a terrificly good idea <IMG SRC="http://forums.civfanatics.com/ubb/wink.gif" border=0> - I want to know whether there are any complaints when we change this to starlifter's idea: let them start with 0. Therefore, I've started a new topic about this. Only, do we change this <u>from now on</u>, or do I change it for past results too? I say we do, by the way. <IMG SRC="http://forums.civfanatics.com/ubb/wink.gif" border=0>

------------------
<IMG SRC="http://home.hetnet.nl/~maartencl/tmp/MatrixBW.gif" border=0>
Game of the Month administrator.

[This message has been edited by Matrix (edited July 12, 2001).]

 

[noughmaster]

Secondly, I am however, quite conservative. And with reason: we can't keep changing the rules on and on. Then the meaning and the values will fade away...

Any new system takes time to settle into its best form. You have to expect things to change a bit early, as with anything that's being tested.
The meaning and values won't fade away; they are in the process of being established. I think there's good reason to change the current rules.
What you're saying is true of a tried, tested and well established system. Of course, rules can change there too.
I've put my suggestions for a new GR system in the new thread.

PS: What do you think of using the Civ percentage (or making our own level dependent function of the Civ score) instead of the actual score in the GOTM score? It's easier to get a big score on a lower level, so lower level games will have more effect on the GR. If we're planing changes, they may as well be all done at the same time.

 

[starlifter]

My Comments on the EARLY FINISH BONUS:

This post is to place my comments from teh GOTM V thread into a location where I can find them 6 months from now.

-------- Posted by me in antoher thread: -------

OK, I understand the Early Bonus issue now, and can give a little more light on the subject. In Civ II, Brian Reynolds (main programmer) decided with Sid Meier that there would be the "normal" score, but realized that rapid conquest should be recognized. After some tweaking, they decided to compute a floor value ("bonus"), and compare it against the normal score. The player was then awarded the higher of the two. Plotted on a chart, it looks like a curve crossing a line. Where the two intersect is where it is no longer advantageous to use the bonus value.

Naturally, they did not expect a group of Civ II Fanatics (huuurah!) five years later to superimpose another scoring system on top of the one the game uses. So naturally, since the GOTM balances speed and size, the bonus has the effect of placing a severe inflection in the GOTM scoring curve. On the early side of the inflection (e.g., before the curves cross), the GOTM "assumes" the player has "built" the civilization to that score, when in fact Civ II simple spits out a flat score. By earning this flat score early and placing into the (50^PNP) adjustment of the GOTM formula, the reward is huge for achieving this bonus early.

Since the bonus is a flat function, there is no Civ II score differentiation. But in the GOTM formula, using a constant (the bonus) greatly unbalances Matrix's (50^PNP).

So, I agree the bonus should not be used in the GOTM score. The reason is that the GOTM score already rewards the time domain (early finish). The normal Civ II score does not reward the time domain; the bonus does. Understanding this will allow everyone to also understand why the SS bonus is both acceptable and necessary to include in the GOTM score... namely, the SS is not a reward for performance in a time domain. It is simply part of accumulating the overall Civ II score, like the FT bonus, or the Barbarian bonus, or even the Wonder bonuses.

Bottom line: Good job, Matrix, and everyone else who must have looked into this at one time. The right thing to do is use the "normal" Civ II score, not the early finish flat bonus. Equal kudos for including all the other bonuses, like the SS bonus. This aspect of the GOTM, together with the (50^PNP) concept is a job well done!

Last note: Since so many play the GOTM, and no telling who might be browsing and decide to play (like I did 3 weeks ago), I do think the time has come to make a proper and complete set of rules, just so even newbies know things like this. They don't have to understand why (necessarily), but it is good to know, and only fair. The only way to really know all the basics is to search and read quite a few threads. Personally, I had no idea about this whole GOTM bonus issue (since early finish is not my usual style

) until yesterday, and had no idea where it was documented.

 

[goodbye_mr_bond]

Originally posted by Matrix:
I want to know whether there are any complaints when we change this to starlifter's idea: let them start with 0.

I'm for it, and past results too.

But I wanted to highlight noughmaster's last comment as well... I have wondered myself about the discrepancy between different levels and their overall effect on the GR score. It is easier to get higher GOTM scores on lower levels because it's easier to win earlier and expand more quickly at lower levels, thus maximizing the GOTM multiplier.

And of course, the PNP part of the equation is greater because there are more turns.

 

[starlifter]

posted July 12, 2001 05:29 PM
Thanks for the summary, starlifter. I don't feel like reading all this stuff if you don't mind. Even if I'm the moderator here.

You can do as you see fit, but I will point out that myself and others put a lot of time into trying to explain this stuff in writing, and it is mostly for your benefit. If it is too much to read and comprehend at once, by all means take a few days. Some of us have spent years learning the underlying concepts in the University, Government, and Corporate worlds. The usual consulting fee for such advice exceeds $100 USD per hour. There is no charge for you <IMG SRC="http://forums.civfanatics.com/ubb/wink.gif" border=0>.

First of all, I must say I'm really into shadowdale's ideas.

Shadowdale has some good ideas; he's an intelligent guy.

Secondly, I am however, quite conservative. And with reason: we can't keep changing the rules on and on.

In World War I, officers told their enlisted something strikingly similar when sending hundreds of thousands to certain deaths in pointless and flawed trench warfare. No one wanted to change the rules. Inertia.

Anyway, a price for the fastest OCC finish would be a good idea. Only there are two problems: first of all, it's inconsistent with the GR: a OCC type of game is not good for your GR score... Second: until now there has only been one player (Smash, GOTM IV) who played a OCC game. Do you think that's going to increase?

I will first say that the reason you are having such difficulty with this is probably because logic is hard for many to follow, in a sustained fashion.

I will donate another $100 dollars of time and try to help you understand. What is the goal of the GOTM? Assuming it is what you and Thunderfall have espoused, this whole concept is supposed to foster learning, new play styles, sharing of ideas, encouragement to play Civ II, etc. Hopefully, we agree on that.

If so, here is the next step. What are the major styles of play? Each should be recognized, and each should have some small incentive. OCC is a legit (and I might add, very helpful) style of Civ II play. So is early conquest. So is empire building. Naturally, only one will emerge as the overall winner each month.

The next step is to determine the best way to encourage varying styles of play, while realizing that all styles will not result in a top score. Your method of awarding "stars" is excellent, IMHO. It needs to be quantified, and BTW Cornmaster came up with an excellent concept that I tweaked a little bit 4 posts earlier in this thread.

Your logic completely broke down when you tried to associate OCC with being a style that was "inconsistent" with the GR. With all due respect, Matrix, take a shovel and pitch that pile of manure behind the barn. Do not leap to conclusions, esp. when they break your own chain of logic. Since you are the moderator, myself and others cannot do this for you. You have to learn to recognize its fallacy and avoid it.

Now about OCC, I am not advocating a huge mathematical correction (though I actually could) to make it possible for an OCC finish to beat the normal "Shadowdale" games. And the GR itself has nothing whatsoever to do with awarding a Star for the best OCC finish. The GR is a separate issue, and I won't sidetrack you with it for the moment.

I am here to tell you that I, and doubtless others, would like to play OCC from time to time, especially if a measly Star was at stake. Let's remember that all this is just a game, and we do it for fun. Further, I know I am raped for life if I ever turn in an OCC score. In fact, I am currently double-raped (maybe gang-raped?), because I am a new player and not a Good Old Boy. 70% of my GR score is 0, and all the prior GOTMs will forever pummel my GR score. So I'm not about to play an OCC, but it would be nice, esp. if I knew Smash and a few others were going to. Smash is one of the best OCC players around, from what I've read.

And if I played an OCC one month guess what? Someone else will have the chance to finish in the top 3. With Shadowdale, Smash, Kev, Cactus Pete, Me, Dimitris, goodbye_mr_bond, etc. at the top every month, others will not have as good a shot at a higher finish. So you see, OCC is a valid play style, and it is compatible with the GOTM, and it is just fine with the GR... as long as you use a GR computation along the lines that I have suggested in other posts (e.g., best 3 of last 5.

Second: until now there has only been one player (Smash, GOTM IV) who played a OCC game. Do you think that's going to increase?

No, not under your current GR system. Everyone that knows what OCC is sees how Smash was raped on his GR score. Although he knew what the result would be, he chose to play it anyway. He is a great OCC player. But most of us are not great OCC players. And if you're going to use an OCC game to shaft me in the GR, I doubt I'll play an OCC, though I'd like to once every 4 months or so, particularly if several other players would informally like to do so also. It will allow others a good shot at a monthly GOTM medal, since my OCC score will likely be the the bottom 10% of the GOTM scores.

Hopefully, you can see past your prior predispositions and see this is yet another reason why the GR should be a average of the best 3 of the last 5. It will help everyone, and there is no incentive for the "usual" high scoring players to play a high-scoring game every month, since the GR can be maintained even with occasional GOTM scores less than 10! You really need to walk to the top of the hill and see the big picture here, dude.

I do not like the other prices for the following reasons:

Survivor Award and Lo-Boy: those are just results within the scale of all the results. I don't see why someone who barily won or who survived the best way should be awarded; doesn't make sense to me. I understand the fun of the idea, but basically it's nonsense.

I have no idea what you do, or even how old you are in real life, but you have an extraordinarily narrow view. You are not demonstrating the remotest grasp of understanding the basics of human behavior or psychology, and I am positive you have never led or motivated people in the corporate or military world. And that's OK, I'm not dogging you. But you are going to have to look at what is in the best overall interest of those that like to play the GOTM, and have an open mind to continually improving it. Your presumptions and attitudes really show through, but I'm glad that you're up front about it.

You do not necessarily need to do the things I suggest because I (Starlifter) suggest them, but it would greatly help both CivFanatics and the GOTM to accommodate the masses (understand the concepts of rewards, not punishment).

You should reward and encourage players to participate. I have seen you insult them in some of your recent posts, even telling one guy something like "someone's gotta lose and it might as well be you". That is crap. Now I'm the last one that will proclaim someone who loses a game as the "winner" (As you know from one of our prior exchanges), but that does not mean that you cannot award a "Lo-Boy" Star to the lowest scoring 2020 winner.

Recall my system, which is based on Cornmaster's awesome idea for the HOF rankings? Here it is:

Points awarded for computing the HOF standings:

1 - Star
2 - Bronze (two Stars equals a Bronze effort)
5 - Silver (it takes three Bronze surpasses a silver effort)
11 - Gold (need three silvers to surpass a gold effort)

Put the thinking cap on, and let's get real. You yourself have been begging people to submit low scoring games, and even bad losses. You have (at time) tried to "encourage" this by apparently insulting them and pounding them with a big stick (your current flawed GR). Why not give a special award, the Star? Your reason is contradictory within its own context "it doesn't make sense to me. I understand the fun of the idea, but basically it's nonsense."

With due respect as the Moderator, you have no clue about fun. You also have no grasp between real life and this game. Most of the people in the world would describe this game of Civ II as "nonsense", and say "it doesn't make sense to me". The same words you used. So get real. This is a game, we play it for fun, and it is nice so see more than the top 3 players get awards. So get off your high horse, and lighten up. We even give informal recognition for screw ups in life and death combat in the real world (I've been in 3 wars). And combat training is dead serious, yet even my tough-as-nails sergeants have a better grasp of giving little awards to people, and expanding recognition, more than you're demonstrating.

I have a task for you, if you're up to it. Sit down and think of several new categories to award Stars in. Think of 10 new categories, even if it takes you a week. They don't have to be the ones I've suggested, even though I'm a good Civ II player and put some thought into my suggestions, AFTER carefully reading the posts of lots of other people. So why don't you sit down and think of some REWARDS (e.g., Star Awards) to hand out each month, and help encourage the players that are not yet highly experienced in Civ II. Let us know what you come up with, but don't come back empty handed.

After you think of 10 new categories of Star Awards (you're welcome to "borrow" my suggestions, or those of others <IMG SRC="http://forums.civfanatics.com/ubb/wink.gif" border=0> ), post them, and let us pick 6 or so, even if you think "it's nonsense."

1500 AD Retirement: I don't want to include a pre-retired game in any possible way. That's harsh, but I've excluded all my unfinished games as well, with reason.

This says volumes, and also reveals that you are the most nay-saying negative person I have seen in CivFanataics before. I've met others that I disagree with on issues more sharply, especially in the Off-Topic threads, but even those people are not so negative as you are.

OK, I'll go through the basic logic again with you. The suggestion was for an Award, not a request for you to recite what you have done in the past. I know what you have done in the past, that is why I made a suggestion to change it. Hence, submitting "but I've excluded all my unfinished games as well.." as a reason for not listening to the suggestion is illogical and circular. In other words, go shovel the manure behind the barn, LOL! <IMG SRC="http://forums.civfanatics.com/ubb/smile.gif" border=0>

"I don't want to include a pre-retired game in any possible way." This is not a reason for discarding a suggestion. We already know you don't want it. I do want it, as a player. Personally, I will never use it, but you need to read Sun Tzu's writings, Matrix. Heed the part about knowing yourself, and knowing your opponent. You do not know the very members of this Forum. Many have expressed a desire to submit a partial game. I will reprint my suggestion, in the hopes that you will actually read it and quit dismissing it out of hand.

Notes on 1500 AD Retirement:

- This is for those that cannot or do not choose to finish a GOTM.
- If someone submits a game in which they retired in 1500 AD (exact year, not approx), they are eligible for this one-point Star award.
- A 1500 AD retirement is not eligible to win any other category or medal.
- A player cannot submit a 1500 AD retirement plus some other game ending; it is one or the other.
- The purpose of this is to recognize, in some small way, the efforts of those that wish to play fast and furious, and get it done quickly.
- A 1500 AD retirement is not treated as a victory, and such a GOTM score is still officially computed using 2020 AD.

As you can see, it is a small kudo to those that cannot finish a full game. As we earlier noted, even OCC is not a realistic option at the moment. But nevertheless, many players want to participate in the GOTM, but cannot find the time for a full game every month. Your solution thus far has been to chastise and insult them, and in your own words, be "harsh". Let me repeat something... this is a game, and we the players play it for fun. We want to play. As the moderator, you need to be more accommodating, or even more people will leave the GOTM and CivFanatics. This is not a crucible for you to pummel hapless players. Get the big picture, dude.

If you don't want my specific suggestion, take a course in logic and discuss why not rationally, and then suggest a viable alternative. But I demand you consider a reward method (not a punishment method) of accommodating players who cannot finish the tedious end-game every single month.

Shadowdale's ideas:

I will show how many GOTM games people have played in the past.

Good news! So maybe I need to get Shadowdale to submit the ideas for change, LOL!

Concerning the decreasion of the GR score with a lower percentage then 30% when they haven't played a GOTM: what if (s)he got killed? Should the GR score decrease with (almost) 30% then? And if not, is it fair that someone who survived till 2020 with only a few cities decrease more in the GR then people who got killed (or didn't play in the first place)?

This is not fully clear to me what you mean.

I will simplify. Forget 30%. Your current GR system should be discarded. Not to be insulting (really), but you just don't have a grasp of how things are really done in the real world, and you've said you are not going to read the stuff I and others have written to help you. I find it amazing that you are so wrapped around the axle with a flawed system. You simply cannot see the forest for the trees, dude. <IMG SRC="http://forums.civfanatics.com/ubb/wink.gif" border=0>

All (count 'em... every single one) of your objections and hang-ups that you express arise only because of your current flawed system.

None of the "problems" you express even exist if you take a straight average (mathematical mean) of the best 3 of 5 (or 4 of 6) recent GOTMs. Are you not understanding how simple this suggestion is? I am not going to help your tweak you current "30%" system, because the only way to fix it is to get rid of all of it.

In my earliest posts two weeks ago, I tried to help you "fix" your current "30%" system, but to no avail. I should never have bothered, as the whole system was and will always be unfair to the majority of players (Ironically, your current system already favors me personally).

If you can think of any serious disadvantages to the proposed best 3 of 5 system, perhaps you can articulate them. Rapid fire adjectives do not count, BTW. There are many scholars and corporate executives that would be interested in your findings, since most of the real world uses this sort of system with great effect.

When inserting the GOTM results into the GR spreadsheet, I suddenly got the idea as well that it's not...right to let newbies start with a certain score - whatever score (in this case the median score).

Congratulations, they taught us that in our first year of numerical analysis. If you will go back and reread all my posts, I've been explaining this to you for 3 weeks.

I want to know whether there are any complaints when we change this to starlifter's idea: let them start with 0

You have taken me WAY out of context.

I am opposed to using ANY score that a player did not actually earn!!

HOWEVER, if using a sliding window (best 3 of 5), it is OK to use Zero for the newbie's unplayed games... until the newbie either plays 3 future GOTMs, or plays the current one and two prior ones (for that player's initial GR only; the player's "catch up" games would not affect concluded GOTMs in any way).

I absolutely oppose using zero in your current "30%" GR method. Likewise, I oppose using the mean. I have written many posts about this.

Only, do we change this from now on, or do I change it for past results too?

Just like Pros Sports, when the rules change (and all rules MUST be able to change!), they apply from that point forward. All awards and standings as you have computed them in prior months will never be changed, even if new rules applied retroactively would cause a change.

Beginning on 1 Aug 2001, the posted results for August would be computed using the new rules. For instance, all GR's would be based on the 3 of 5 method, and the Hall of Fame order would be computed with the 1-2-5-11 point system. In six months, or at some time in the future, if rules change again, then the results from that point forward are computed under the new system. Don't be scared of change. The one constant in life is change.

Now, because CivFanatics is a dictatorship and I make the decisions ...

That may be, I actually am not sure who "owns" CivFanatics, but I leave you with a thought that I've tried to teach many new officers and NCOs:

"Lead, Follow, or Get Out of the Way"

You are the Moderator, and your job is to lead. You've done some great things, but seem to be a prisoner of your own limitations and predispositions. Take the lead, and be positive, not a nay-sayer. The world is full of those who "cannot" (and you certainly always seem to have reasons that you "cannot", LOL!), but the world (and GOTM players) always needs someone who has a "can-do" attitude.


<IMG SRC="http://forums.civfanatics.com/ubb/beerchug.gif" border=0>

EDIT: Spelling.

[This message has been edited by starlifter (edited July 13, 2001).]

 

[starlifter]

There are 3 main parts to this summary:

1. Global Ranking Suggestions
2. Global Ranking Example Computation
3. Hall of Fame Suggestions


1. GLOBAL RANKING SUGGESTIONS

As currently implemented, the existing GR is illogical and faulty in the extreme... both mathematically and in common sense... because a new player can NEVER overcome the "penalty" of being a newbie. The bias is a permanent. And it reeks.

Veteran players are ridden hard and put away wet if they dare miss a GOTM, or (heaven forbid) try a lower-scoring strategy like OCC.


The fair (and stunningly self-evident) solution is to institute a sliding window of games that are used to compute the GR.

The Sliding Window is simple and professionally accepted (in the corporate, academic, government, and military worlds). It merely averages the highest X in Y scores. A good model to use is take the top 3 GOTM results of the last 5 GOTMs offered, and average those 3 to obtain the Global Ranking (GR). This means a player can miss a game, get a low score, try a new strategy, even lose a GOTM in spectacular fashion, and not blow their GR. This system emphasizes the players best efforts, and discards the worst results. And the history is never older than 5 months, which means any results from 6 or more months ago are stale and no longer used. This is excellent news for player that are improving (they are not stigmatized for life, like the current system does).

With this mathematically fair Sliding Window method, I favor using 0 for the score in a new player's unplayed games. However, those zeros are rapidly erased as the new player finishes a 2nd and 3rd game. At the 3rd and later games, a new player is on equal footing with all Good Ol Boys.

I am opposed to using ANY score that a player did not actually earn (like Zero or Median in the current, flawed GR method)!! HOWEVER, if using a Sliding Window (best 3 of 5), it is OK to use Zero for the newbie's unplayed games... until the newbie either plays 3 future GOTMs, or plays the current one and the two immediately prior GOTMs (results used for that player's initial GR only; the player's "catch up" games would not affect concluded GOTMs in any way).



2. GLOBAL RANKING COMPUTATION

1. Consider the last 5 GOTMs. GOTMs older than 5 months will not be considered anymore.
2. Determine the player's highest 3 GOTM scores.
3. Use straight mean on those top 3 results ("mean" is a specific mathematical method of averaging).
4. For newbies, 0 will be used for unplayed GOTMs.

====== EXAMPLE ========

For newbie, here is month one (assume GOTM of 60): (0+0+60)/3 = GR of 20

Month 2 (assume GOTM=90): (0+60+90)/3 = GR of 50

Month 3 (assume GOTM=75): (60+90+75)/3= GR of 75

===== END EXAMPLE ======

As you can see from my example, I actually do think the "fair" way is to use zero for new players, but only in a best 3 in the last 5 method (not in the current GR method). Zeros are replaced as the player submits completed GOTMs. In Matrix' faulty system, a player never rids him/her self of the unfair influence of the
"Good Old Boys" penalty for not discovering and playing GOTMs from the very start.

Fair. Simple. Flexible. Questions?

3. HALL OF FAME SUGGESTIONS

Cornmaster has an awesome point, and excellent idea, which can be easily tweaked. Using a point system properly will add credibility to the Hall of Fame, and encourage a variety of gameplay, like OCC and Fast Finishes, even if such strategies do not gain the highest GOTM score.

Thanks, Cornmaster, for "thinking outside the box" and pointing the way!

This system is primarily based on Cornmaster's awesome idea for the HOF rankings:

HOF Points (for computing the HOF standings):

1 - Star
2 - Bronze (two Stars equals a Bronze effort)
5 - Silver (it takes three Bronze surpasses a silver effort)
11 - Gold (need three silvers to surpass a gold effort)

HOF Star Categories:

1. Fastest finish.
2. Highest score.
3. Fastest OCC finish (New!).
4. Survivor Award (New!).
5. 1500 AD Retirement (New!).
6. Lo-Boy (New!).
7. Player's Choice (New!).

Notes on #3 (Fastest OCC finish):
- Since the purpose of OCC is to finish in the earliest possible year, an additional category fo high scoring OCC award is not necessary

Notes on #4 (Survivor Award):
- This is for the best "Survivor" (highest non-victory score) in a given month.
- Players must survive until 2020!
- If the AI eliminates a player, they are not eligible for this award.
- This will primarily be a consolation award for those who cannot win via a Spaceship or Conquest by 2020.

Notes on #5 (1500 AD Retirement):
- This is for those that cannot or do not choose to finish a GOTM.
- If someone submits a game in which they retired in 1500 AD (exact year, not approx), they are eligible for this one-point Star award.
- A 1500 AD retirement is not eligible to win any other category or medal.
- A player cannot submit a 1500 AD retirement plus some other game ending; it is one or the other.
- The purpose of this is to recognize, in some small way, the efforts of those that wish to play fast and furious, and get it done quickly.
- A 1500 AD retirement is not treated as a victory, and such a GOTM score is still officially computed using 2020 AD.

Notes on #6 (Lo-Boy):
- This is for the lowest-scoring player who still achieved a standard victory with more than one city (landed SS or conquered world) in a given month.
- Players must achieve a Civ II Victory!
- If the AI wins, the player is not eligible for this award.
- If the AI lands a SS first, the player is not eligible for this award.
- The player must have more than one city; this is not an OCC strategy award.
- This will primarily be a consolation award for those who can struggle to a win, but do it with a LOW score!

Note on #7 (Player's Choice):
- All of that months GOTM participants can vote on their favorite 3 saved games of another player.
- Voting period for 14 days after the zipped .SAVs are posted.
- Other Star/Medal winners ineligible (?)
- Can consider a player's included text narration (.TXT or .DOC file with the .ZIPs)
- This category could be expanded to Favorite Narration, Closest Call, Most Impressive Comeback, etc.
- Player's Choice Award(s) would be one Star (one point for HOF)

 

[noughmaster]

I like the idea of a few different categories, but I don't think they should all get stars for the HOF. The highest scoring 2020 retirement can't be seen in any way on a similar level to highest score or fastest finish.
I'm not sure about option 6: lowest scoring win. It would be easy to get a large negative score by nuking everything and polluting the world.
Has anyone else ever done this? It's quite funny. If the ice caps melt 6 or 7 times, virtually the whole planet is either desert or swamp. If you've already built a lot of military units and are in fundamentalism, this strategy actually helps to conquer the world by throwing the other Civs into chaos (assuming there are any left after you've let off about 50 nukes).
Maybe it would be good to have a category to see who can **** up the planet the most, but I don't think it should get a HOF star.

 

[starlifter]

posted July 16, 2001 02:08 AM

I like the idea of a few different categories, but I don't think they should all get stars for the HOF. The highest scoring 2020 retirement can't be seen in any way on a similar level to highest score or fastest finish.
I'm not sure about option 6: lowest scoring win. It would be easy to get a large negative score by nuking everything and polluting the world.

Good point about #6. The main idea about the 2020 thing is to have some sort of way to both encourage players to finish their games and submit them, yet to have a little recognition for the one who did the best in that situation. Everyone should know by now that no 2020 finish is likely to contend for any of the main honors (the medals).

In general, it is good to find ways to recognize people, esp. when you have 50 to 90 players. This is an informal game (the GOTM), and we should try and think of some creative ways to encourage/recognize various results. I'm not suggesting more that 3 medals, nor any way for no-medal winners to accumulate multiple stars (thus "beating" medal winners in HOF points). In fact, if the Gold is 11 points, it would take about a year of stars to get the points to equal just one top finish. The differential also means a "top" player can take a month "off" and not be worries about some non-winner getting a start (one point) and overtaking his medal effort.

So hopefully, all you creative types (I'm more of an analytical type) think up some things that might work, and hopefully we can discuss 'em and find a few new good star categories that aren't flawed (like I now realize my #6 is).