GOTM 17 and 16 Statistics Summary

Karasu · Apr 14, 2003

As we did last month, I want to share with you some of the statistics we have reviewed to help put the results fo Gotm17-Carthage into perspective.

Choices of Finishing Conditions

Out of the 162 submissions that are already listed in this month's results table, 141 were winning games.
Your general choice favoured once again going for space with a remarkable number of highly-skilled diplomats...
And no histographic victories: now that Jason rules, everyone seemed afraid to milk a game

Out of the 21 defeats, 18 were space race and 3 diplomatic. The truly separated archipelago style map at least prevented the AI from military and cultural wins...

From this month on, we are also looking at the different software versions, and here they are:

Code:

                 Total     Lost      Won
Civ3               65        5        60
PTW                78        8        70
Civ3 Mac           7         0         7

So, while the "PC Civ" groups gave a nice display of skill by submitting ~90% victories, the intrepid minority of Mac players really blew us off by winning all of their games. Congratulations!

Finally (on finishes...), and as both an experiment and an attempt at encouraging discussion, here is a look at how the victory conditions we selected changed from GOTM16 to GOTM17, as a percentage of the total number of victories.

Karasu · Apr 14, 2003

To the scores now.
The new system makes it hard to compare these results with last month -we'll have to wait until GOTM18 is processed for that...
In this sense, we'll record for future reference that

Code:

                   Overall   Losses    Wins
Highest            9877      1870     9877
Lowest              527       527     1547
Average            3898      1126     4311

The first thing one notices is that the scores are indeed less spread out than previously (of course, were it not for DaveMcW who insists in beating the system "best" date...

). The large gap between the high scoring games and the general distribution of all players capped off with scores by well played quick victory games has been reduced to provide a more understandable set of results.

Following your collective suggestions, this month we are including an overview of game scores by victory type. The plot is a bit compressed for publishing, but you'll be able to find both the jpg files and the source data zipped at the end of all this...

It is interesting to note how space race and diplomatic victories, although the more abundant, were amassed at the bottom end of the scores... better luck next time!

Karasu · Apr 15, 2003

Same lot of information for the finish dates here.

In summary,

Code:

            Overall      Losses      Wins
Earliest    970 AD       1862 AD     970 AD
Latest      2050 AD      2014 AD    2050 AD
Average     1826 AD      1952 AD    1808 AD

The earliest victory belonging to the often mentioned DaveMcW, who is really biasing our statistic here...

.
The average finish date is clearly high, due to the map characteristics -still, the few 2050 victories may mean that Jason did not scare everyone from a milking attempt...

Here too, in a valiant attempt at anticipating your requests, is the distribution by victory type.

Karasu · Apr 15, 2003

And now, the one thing we are really interested in... the infamous "turns per session" plot.

As usual, some numbers first:

Code:

             GOTM17   GOTM16
Max         153        216
Min           2          3
Average      31         31

The plots tend to speak for themselves, but you may want to pay particular attention to the fact that well over 90% of the games fall into the category of being played in sessions that average well above turns per session. We always seem to have 2 or 3 very low games that require special review but for the most part the statistic represents that most players do tend to play the game in blocks of time.

cracker · Apr 25, 2003

My apologies to all for being slow to get the graphics and text for this great statistical analysis out to all of you. Karasu did all the hard work as usual and then we just had to wait on me to get my fumbling act together from other firefights.

I wanted to make sure that everyone had this information available to them before you began your pregame planning for the upcoming Gotm19-Ottomans game.

a space oddity · Apr 25, 2003

Wow, great job again guys! :goodjob:

Where on earth do you find time to do all this??? (esp. given the fact you're Civfanatics too

)

edit: Apart from the shift away from histographic, it seems that the map and the added time pressure made people shift from Space Race to Diplo. But there is also a shift from Dominition to Conquest. My guess is that this map made Conquest actually quicker than Dominition, because dom. needs expanding borders = buildings = expensive on a map like that.

Darkness · Apr 25, 2003

I have no clue whatsoever as to where you people find the time to do this and still play the GOTM... I'm having enough problems finding enough time to finish before the end of each month...
Great work!

LeSphinx · Apr 25, 2003

I'm realy very impressed by all the work all the guys have done here! Waoua! Score and players Statistics, interesting GOTM, QSC, lot of players, a lot a update informations about civ world...
It's the best civ series sites I ever see!

I'm looking forward to the pre-discussion about GOTM19!
What do you reserve to us... Let's wait a couple a day...

LeSphinx

Karasu · Apr 25, 2003

Originally posted by cracker
My apologies to all for being slow to get the graphics and text for this great statistical analysis out to all of you. Karasu did all the hard work as usual and then we just had to wait on me to get my fumbling act together from other firefights.

"Slow"... "Fumbling act"... :lol:

(admirable word choice, by the way).
You are just too efficient -and a gentleman for taking the blame on yourself. :worshp:

Time is not a problem when you enjoy what you do -the real problem is to find myself at work on holidays (like today... :mad:

).

mad-bax · Apr 25, 2003

I'm no statistician, but some of these distibutions really strike me as unexpected. The high proportion of Space race games I can understand because I personally find it the easiest way to win when I am stuck in the modern age.

But the high proportion of Diplo games amazes me. When asked if I want to hold an election, If there was a "Not likely Bud and don't ask again" I'd click it.

It also surprises me that conquest and domination wins score so highly. But maybe thats because these types of wins are acheived more often by the more experienced players and (ex) milkers. Now if there were a set of distributions for win type against global ranking.........(only joking).

Nice stats. Need a bigger data set for conquest/cultural/domination wins but that will come.

I'd like to meet the person who could play 200 turns per save.

tao · Apr 25, 2003

Originally posted by mad-bax
I'd like to meet the person who could play 200 turns per save.

Meet the Macintosh players.

Since currently the saves get corrupted sooner or later, we just never quit Civ III. My gotm 18 is now running since 3 weeks. The good thing is: this is no problem with Mac OS X. And I can even play another game with 1.29 beta in parallel.

EMan · Apr 25, 2003

......You guys are performing a tremendous job for the CFC community! :goodjob:

I really tried in GOTM17 to keep the number of turns per reload to a minimum BUT was "gotcha...d" for 3 different reasons:

1. Early game AND late game had game-crashing bugs (no complaint intended) that caused several reloads while I figured out what was causing the bug and adjusting the play accordingly.
2. My notebook computer, for whatever reason, seems to get "CPU-hog-tied" and the game starts grinding to a halt.....have to reload to "free" the memory.
3. While traveling, I can't hibernate the Notebook when Civ is running....so the game has to be exited and reloaded.

Anyway, I 've decided that turns per reload is an interesting statistic BUT that I'm never gonna win an award for "low turns"!

(UNLESS............I buy a MAC!!!)

EMan (HOF Milker & Former GOTM Milker).
P.s. As to the post re Conquest v Domination, I had the Domination victory "for the pickins" many turns before Conquest BUT figured the Conquest would score higher under Aeson Scoring, so went for the "Gold', so to speak!

Dianthus · Apr 25, 2003

It might be interesting to know the correlation between turns per session and score. I would expect that a very low/high turns per session would suffer in overall score due to :

1) Low, forgetting what was occuring since the last session and what was planned for the future.
2) High, rushing too much and loss of concentration.

I only discovered GOTM with 4 or 5 days to spare before the submission deadline and 2 of those I was going to be away for. As a result I did some pretty long sessions for the 2 or 3 (working) days I had available. Part of the reason I did so poorly (2nd from last) is that I had to rush this, though to be honest I'm nowhere near the standard most of the rest of you are at.

It would also be interesting know how long players take per turn. This stat would obviously change throughout the game due to a combination the size of the Civ increasing as time progresses and the amount of micro management decreasing as the game progresses. I guess this would only apply to PTW though as I don't think the total game time is available in the .SAV for previous versions.

Moonsinger · Apr 25, 2003

Karasu,

Nice works! :goodjob:

I was trying to do one in Microsoft Excel last week (where I have a set of final scores and years and I want a graph to show the Jason curve for each type of victory) and after about 4 hours, it still won't come out right.

------------

Originally posted by mad-bax
It also surprises me that conquest and domination wins score so highly. But maybe thats because these types of wins are acheived more often by the more experienced players and (ex) milkers. Now if there were a set of distributions for win type against global ranking.........(only joking).

Not really!

Early Conquest and Domination wins aren't always give us the best score. I'm sure we will talk more about this topic on another thread after May 10th.

jaheira · Apr 25, 2003

I will apply for GOTM19.. but speaking about 17.. I did play, and comparing the number. I must say that the fact turns are just too long .. and makes it difficult to have one session with lot of turns. I am travelling and usually I close while my laptop is lacking battery (average 2 hours).

Wait for result of GOTM18 as I am on the way to finish it. and then I hope to find time to be part of the next GOTM19 list..

HookEmHorns · Apr 25, 2003

But the high proportion of Diplo games amazes me. When asked if I want to hold an election, If there was a "Not likely Bud and don't ask again" I'd click it.

I was late in taking out the Egyptians so planned to just stay on that group and ride out the game, trading with all as a proper Carthaginian. Once I saw the lack of coal, Diplomatic just made sense since I would never have the production capability without railroads. I led the rest of the game in science, selling every six turns or so to rake in the bucks and only built two factories (one for the forbidden palace in Thebes and in Carthage for the U.N.) and only one hydro plant (Carthage). Near the end I gifted a lot of techs to help my global image.

mad-bax · Apr 25, 2003

Originally posted by Moonsinger

Not really! Early Conquest and Domination wins aren't always give us the best score. I'm sure we will talk more about this topic on another thread after May 10th.

You went for a cultural win, as did Ribbanah. But there must be a reason why so many people go for dip and Space wins compared to the others. There must also be a reason why those victories score less on average. Is it because of the scoring system, or is it because a higher proportion of experienced players play for these victory conditions.

As I said, I can understand Space race, because as a bad player I can usually contrive an SS win if I am in trouble. But Diplo? It's so much harder to tell if you're going to win (IMO) and so clicking "YES" is a :nono:

for me, unless I'm forced into it in a tourney game.

Moonsinger · Apr 25, 2003

Originally posted by mad-bax
You went for a cultural win, as did Ribbanah. But there must be a reason why so many people go for dip and Space wins compared to the others. There must also be a reason why those victories score less on average. Is it because of the scoring system, or is it because a higher proportion of experienced players play for these victory conditions.

I have a few things I want to share, but I can't do it until after Cracker publishes the result of the GOTM18. Just because something work best on one type of map or diffcult level, that doesn't always mean it will work well on another type of map.

As I said, I can understand Space race, because as a bad player I can usually contrive an SS win if I am in trouble. But Diplo? It's so much harder to tell if you're going to win (IMO) and so clicking "YES" is a for me, unless I'm forced into it in a tourney game.

Yes, Diplomatic win is most difficult one of all. Even the top player like Sir Pleb and Aeson often have trouble with it, but not any more. Sir Pleb just found a key for it as he had demonstated in his GOTM18.

For the record, I have never been able to win by the UN ever.

Therefore, I will try to make it as part of my goal for the GOTM19.

civ_steve · Apr 25, 2003

Great Stats!! :goodjob:

The Gotm crew continues to impress, especially (in this case) Karasu who is one of the later spoiler posters, but still finds the time to put these detailed charts together!!

I think it makes sense that the Diplo and Space victories score lower overall, because the Jason scoring system levels the playing field between the milked games and early finish games. For the average player, the Diplo and Space victory will come later in the game, and if the player is not performing good 'milking' techniques in conjuntion with the victory attempt, their score will suffer. (These games are neither early finish nor milked.)

a space oddity · Apr 25, 2003

Speaking for myself (with a diplo win): I just wanted to win my first GOTM game on a difficulty level I had not beaten before. And rather sooner than later. The UN was just my first chance of winning (I hadn't even eliminated the Egyptians). I do know how to find out how well I am doing in the opinion of the AI and as I play always honourable (comes naturally, not conciously trying to and a hindrance to scoring really well

) I mostly do well in these elections.

I think there are more players like me that do not normally play the higher levels (yet) took this chance to enter the GOTM. Early conquest and dominition require planning and will be done by more experienced players IMHO. (But maybe that's just me, and naturally there are players that have conquest as their natural game style.)

GOTM 17 and 16 Statistics Summary

Wanderer

Wanderer

Wanderer

Wanderer

Gil Favor's Sidekick

Deity

Shadow creature

Bachogwa

Wanderer

Deity

Deity

HOFer: Milk-Cow?

Small but hardy

Settler

Chieftain

Chieftain

Deity

Settler

Deity

Deity

Similar threads