GOTM Score - how it should be!

[Diogen]

Hi there!

I am a newbie here. Please do not push me very much.

I looked at the GOTM14 scores, and found that it is a final s#xx!!! Anyone can wait until 2050, developing higher and heigher score. It is ridiculous!

Analising that table, I came to a suggestion: add to the game score (2050-year_of_victory)*30. Then it would be fair to all.

Any comments?

 

[Yndy]

Let’s say you disband your settler in 4000BC. Score = 16,200

Say only victories would get this bonus. The problem would somehow alleviate the problem but not totally. The milking will stop when score increase per turn drops below 30 so you will see a range of finishing dates but the scores will still be milked.

For example: In the GOTM 15 milked game I got an initial 60 points per additional turn and by 2050 this dropped to 25, I think. So with your scoring method I would have ended the game around 2000 with a score 250 points higher.

Read these thread if you want to know more about proposed scoring changes (it’s a long read)
Old GOTM Scoring Discussion
Read this thread for milking discussions (pretty long as well)
Ethics of milking (Edit: saw you read that one)
A new scoring system will probably be implemented in the near future that would still allow milking but would minimize its effects. Still chances are that you will see the same people in the top table as they really are good and it’s not just an effect of the scoring system.

 

[Diogen]

Yes, U'r write! ONLY the victory game will be using this scoring. And one more point: not (2050-yotv), but Turns to the 2050 - what I ment. That is ending (victory!) in 1800 will add 7500 to your score.

 

[Megalou]

I believe there is an early win bonus which will help the best player win most GOTM:s, or at least get a very high rank. Note that GOTM14 was extremely hard and therefore no one was apparantly able to finish the game early enough. (It's still exceptional that they won at all, isn't it?)

I think the formula for early win bonus is Difficulty level * (2050 - year of victory). Please correct me if I'm wrong because I'd like to know. If it were 2050 - number of turns left that counted, then it would be LESS favorable to win early than it is now, so I didn't understand your last comment.

Sincerely,

 

[Diogen]

My last comment is based on the note, that actaully all milkers (4 in the gotm14) get their highscores just different by turns(2050-yotv)*30. Of course, this is for Monarch level ONLY! For others one should look at the similar statistics. edit: I mean, different from similar types of victory.

 

[Bamspeedy]

Using a strict # as the early win bonus unfairly favors one victory condition over others depending on the map. If the bonus is too high, then spaceship victories cannot compete with an early conquest/domination. If the bonus is too small, then milking will still win.

GOTM12-Kemal had a very early spaceship launch. He won at 1385 A.D., so his bonus would be 8,190 (273 turns*30). I got conquest at 540 A.D., so the bonus would be 11,130 pts (371 *30). Could Kemal have 'milked' 3,000 more points during those 98 turns, while still trying to launch his spaceship as early as possible? Before the game added on it's own early win bonus he didn't have 3000 pts, let alone having that many more points higher than my base score. Did I play a better game than him? It's hard to say, but I don't think so. We had two completely different goals, so you can't compare.

There are many factors that go into how much an early win bonus should be given that any strict number doesn't take into account (map size, land %, # of opponents, land formations i.e. islands vs. pangea, which victory condition you are going for, etc.). The proposed scoring system that Aeson has posted in the other threads look much closer to what the optimal system should be. There are different bonuses based on the map specifics and which victory condition you go for. There may be some situation early on where it may look unfair to some victory conditions, but hopefully we will get those more evened out as we get more results from various playing styles and results of games. No formula will be perfect, but Aeson's looks like the closest we are going to get for awhile.

 

[Ribannah]

The alternative and very simple scoring system, which I proposed several months ago, would be to give the fastest victories in each category 100 points, and rate the other games according to the percentage of turns 'wasted' compared to the fastest finish in the same category (so a finish in 2050 AD would get ZERO points, or a minimum for having won at all). This way no examination of the map is needed. There could always be extra prizes and/or bonus points for building the most wonders and the like, as well as for the highest in-game score (so milking would still earn an award, but not dominate the ranking).
A mix is also possible: 50 points max for finish per category, plus 50 points max for score (over all the categories). This would cover all aspects of the game and rate losses, too.

 

[Aeson]

If you look at the GOTM results, often there are victory conditions which only a couple of people go for, and in some cases, none at all. This can result in the 'fastest' victory of one type or another not being fast at all. Consider an extreme example... 2 players launch spaceships, one at 2049, and the other at 2050. Should one of them score 100, and the other 0?

 

[Ribannah]

That's a definite yes. Teaches the 2050 launcher to go for 2048 the next game.

Seriously: people will adept their game to the rules, just like they are doing now.

 

[Aeson]

The tournament uses that sort of scoring system... 'fastest take all'. If that's the sort of competition you're interested in, it's already available. The only reason it works IMO, is because the victory condition is set the same for everyone. If you have players choosing which condition to go for, often it's going to be a scoring system which rewards those who go for a victory condition fewer people chose, not the skill in which they obtain their victory.

 

[Moonsinger]

Originally posted by Ribannah
Seriously: people will adept their game to the rules, just like they are doing now.

Exactly! The skillful players won't complain about the challenge, rule or restriction that comes with the game. One way or another, it would be no problem at all for them to adapt.

 

[Ribannah]

Originally posted by Aeson
If you have players choosing which condition to go for, often it's going to be a scoring system which rewards those who go for a victory condition fewer people chose, not the skill in which they obtain their victory.

I expect this to balance out until each victory condition has roughly the same number of players, on average.

But what do you think about a mix, with say 50 points for speed and 50 for score? This is still very easy to implement.

We would see a lot of different plays, with some people trying to maximize their empire in 1900 AD, others choosing different times. You'd have to plan ahead on when you will seek victory as well as how, weighing early war against prioritizing infrastructure or research etc.

Will you wipe out your neighbour first and then start the tech race with a bigger but ancient empire, or will you start small but advanced?

--

I don't prefer the Tournament because you're kinda stuck at one difficulty level. Maybe, once I have played a lot of games, I will stick to deity forever (just as with Civ2), but for now I'm enjoying the variety, taking my time.

 

[Aeson]

I think that regardless of the ratio used, it's adding an arbitrary modifier to the scoring system, with variation based not on skill, but on who else does what. The player base isn't wide enough to make it statistically likely that a 'good' early victory for each type will be submitted. If you take the top 10% (just theoretical) of 80 players (8), and spread them between 5 or 6 victory conditions (20k and 100k culture really being seperate victory types), there often will be a victory types that aren't represented in their submissions. If Kemal is launching in the 1400's and gets 100 points, it's not the same as if none of the top players go for Diplomatic that month, and someone 'lucks' into 100 points with a 1800's Diplomatic victory. If Kemal instead calls a vote when the UN is built in 1150AD (usually ~40 turns before a spaceship can be launched), then the 1800's victory is going to be given ~75 points. In both cases the skill involved is the same, and this variation (100 vs 75) is based soley on the choice made in another game.

If you look at the GOTM results, almost every month there are victory conditions which just don't get represented by a top flight submission. It becomes a guessing game for just about everyone. "Which victory condition will the players better than me not go for this month?" Guess right and get a good score. The only players it isn't arbitrary for are the ones who can consistantly turn in top flight games in their chosen category.

If the goal is to implement a scoring system which somehow gives even weight to different playstyles, this doesn't fulfill that.

A better method is to base the date/score that each game is weighed against off the map, as that isn't going to change. It's more complex, but doesn't show variation based on what other players do. What matters is what each individual player does in their own game.

 

[Bolka]

I see an interesting element in this thread, which is to have a fixed percentage of the score based on the finishing year. I like the incentive of finishing early, it also make for shorter games, which is good if you want to submit one a month...
I like the idea that the Civ score would only be a percentage of the gotm score.

The scoring system should stay simple, and I don't think a bonus that varies depending on the victory condition fulfills that, since the player will have to think a lot in order to maximize his score and will want to compare with already existing result of the current game (which he might even get from this forum).
But the idea of a generic bonus for early win sounds good still.

One question related to Bamspeedy's comment: if you finish with a space race, don't you get an advantage on score compared to the early domination/conquer win since you played longer, and therefore spent more years with higher population?

Any new scoring needs trial anyway, and will have to be applied to existing situations, like Aeson has done with the new scoring system.

 

[Bamspeedy]

Bolka- Yes, space race will score a higher 'base' score (score before early win bonus is added) because of having more territory for a longer period of time. The game's default early win bonus just is not high enough on most maps. Kemal, if he was anywhere near the domination limit would have a higher score by delaying his launch until 2050 A.D. I don't know, but I may have scored higher by milking until 2050 A.D. instead of winning at 540 A.D.

The later on in the game you play, the more advantageous it is to milk. Around 500 A.D, each turn represents 10 years, so 10(years) *4 (monarch level)=40 pts/turn I'm losing out on the early win bonus if I hold off victory. After 1950 A.D. each turn loses out on just 4 pts/turn (1 year *4). So with the game's scoring system an early conquest/domination can only win if it is very early in the game when each turn offers so many more bonus points. If you change it and the early win bonus is way too high, then you'll have the complete opposite and early conquest/domination will always prevail.

In that game I used horsemen extensively and knights just to kill off the last AI (it was a pangea map). Had I needed to get to Navigation to reach some of the AI, I wouldn't have won that early and lost out on lots of points.

Cracker scored 10,000+ pts by milking to 2050 A.D.. For the game to give him a higher score than that by an early victory means he would have had to win around the year 210 B.C.!! (most likely earlier than this, I'm just guessing what his base score would have been at that point). At 210 B.C. the early win bonus for monarch level is 8,960 pts.

On tiny/small maps there just isn't enough land to support the population that an early victory is almost always better. The more land available/larger map the easier it is for the milkers (or late finishes/victory conditions). The worst extreme is huge maps. If you cheat and win at 4000 B.C. on deity, you get around 36,000 pts. Aeson in his deity HoF game milked until 2050 A.D. and scored over 60,000 pts.

 

[Bolka]

Thank you very much for the thourough explanation Bamspeedy!

So it looks like there is room for an early win reward, as long as it makes it even for milkers and swift conquerors, so that space launchers are confortable in between...

 

[pterrok]

Not to open a can of worms here, but there was a bit of a thread in one of the spoilers for this month's game about how to LOSE ASAP when you felt the game was hopeless so as to keep a better score. On the Deity games especially, there are going to be a LOT of losers!

Do you think a tweak could be used to adjust losing players scores for staying in the game longer when they lose? i.e. adding/subtracting points to their score for number of turns survived?

I didn't submit the last Deity game because I lost so quickly (and replayed it and lost quickly, and replayed it and lost quickly, etc. <g>) but with the new global ranking system I WILL submit future Deity games--but I'd like to not have to worry about making a decent sized empire to get my score up as much I think I'll be able and then just giving it all away to get killed to lock-in that score...

Am I just being silly here?

I DID submit the Deity game prior to last month's--I got whittled down and kept playing until the AI won in the 1500s or whatever. I DID have to wait for them to go through their turns to fight their wars which got tedious, so maybe forcing players to never say die is NOT a good thing...

So I guess, Aeson, Bamspeedy, et al, start thinking like losers and see how you would break the system for your benefit if you couldn't win Deity games...

 

[Aeson]

Something like that is in the works pterrok.

'Hanging in there' should definitely be rewarded and not punished. A couple of suggestions in that regards have been tried, but they need some tweaking. The problem is getting it so that losing the game will never score more than if you were to win in the same position. So far there have been loopholes of this nature in the suggestions.

In the context of the scoring system in the 'Old Scoring System discussion' thread, the initial suggestion was to take the square root of PlayerScore/MaxScore and multiply that by 10000 (to format the score), which does solve many of the problems. When applied to some games though, losing would be more rewarding than winning. A bit of tweaking should get rid of those loopholes though.

 

[DaveMcW]

I don't like giving bonus points for 'hanging in there'. You can survive indefinitely by blocking the coast of an island.

If you have been beaten down to an OCC and are 2 eras behind, just suicide and take the loss.