GOTM 16 Statistics

[Karasu]

This is a first sample of the kind of information that we will be publishing as an extra to the GOTM.
Rather than flooding the forum with graphs and tables, I would start with a few obvious things, such as scores and victory types.

The main reason behind this is clearly the anticipated change in scoring system: I found it very interesting to see how the finish types and dates were distributed in GOTM 16, and wonder how the picture may change in GOTM 17.
Well, GOTM 16 was admittedly an 'atypical' game -but from now on, I sort of suspect that Cracker will only throw a "conventional" game at us every once in a while, as a surprise...

From my professional point of view, I must say, I am delighted to see that most of you decided to launch their spaceship...

This month, we had 144 submissions. A remarkable 99 wins were obtained, with an average score of 4850 points (in spite of my score in this game, that's above my average... ), with 1781 AD as the average finish date.

More in detail:

Score  -            Overall        Losses          Wins
Highest              20,860          3,639       20,860
Lowest	                456            456        2,526
Average               4,850          1,850        6,213

Finish Date -       Overall        Losses          Wins
Earliest	     280 AD        280 AD       1200 AD
Latest              2050 AD       2036 AD       2050 AD
Average             1781 AD       1758 AD       1792 AD

Victory Conditions

Distribution of Scores

Finish Dates (Internal Civ3 Year)

For those of you who were wondering why the "turns per session" figure does not appear in the results table any longer...

We believe it is definitely more useful to present it this way, so that everyone can see where they fall in the plot and have a better understanding of their own playstyle and of the general trend, rather than discussing any single player's count and related playing habits (be assured that Cracker's watchful eye is ever alert for potential frauds... ).

I would add that it will be particularly interesting to see how this plot evolves with time.
A few numbers before the plot:

Mini turns per session: 3
Max turns per session: 216
Average: 31

 

[civ_steve]

An interesting set of statistics! I appreciate and enjoy the expansion of information and experiences that the GOTM staff have added.

Would you also have a distribution of the player's scores using the new scoring system? This may give participants a better idea of how their game-style fits with the new system.

 

[TedJackson]

It will be even more interesting if you develop this further to show the month on month trends, especially the victory types and score distribution.

Nice job.

regards

Ted

 

[IceCreamEmperor]

How about a breakdown of Scores-by-Victory-type?

This might be particularily interesting once the new scoring system comes in, since each victory type would be measuring itself against a different 'best date.'

 

[Ribannah]

Thank you, Karasu. I'm glad that I liberated you.

 

[RufRydyr]

Kewl. I love seeing performance statistics in pretty graphs.

Do you plan on doing any individual comparisons? Showing how much each play improved or got worse between GOTMs?

 

[ControlFreak]

Wonderful work and very nice presentation.

One might think you had a copy of excel with you in that volcano .

I second RufRydyr's request for improvement tracking. I think this stat would be fun and useful if measurable!

 

[Karasu]

Thanks everyone for the feedback -and, of course, for liberating me from the volcano... (and yes, of course I had a copy of excel with me; I even kept it when some slavists among you forced me to clear acres and acres of jungle...).

Yes I had in mind some changes to implement with the new scoring system. Finish date will very likely be replaced by a breakdown by victory type.

I like the idea of tracking some players' performance over time, if that's what you are suggesting. Of course, with about 800 registered players (and counting), I wouldn't want to do that for everyone...

Of course, there may be several ways to make individual comparisons -I am open to specific suggestions....

 

[Zachriel]

You guys are fanatics! Really.

What would we do without you.

 

[jeffelammar]

Neat stats here.

The turns per session stat is a little strange though. It is so dependent on many things. For example a warmonger is more likely to have a low stat than a classic builder. Someone who milks all the way to 2050 is going to probably have a lower number too. This is because of the time each turn takes when you have a large empire or must move lots of military units.

I would bet that early wins will have a much higher average turns per session. I know that I can often get though to 1000 BC in a single session (not always), but to get 2 or 3 turns in 1928 AD can take an entire session.

PS: Is there any way I can find out what my total turns per session is.

 

[Karasu]

Zachriel, yes, I guess so (unless of course you meant "fantastic", which is also true... )

Jeffelammar, the idea behind the TpS overall plot is to avoid public discussion of a single player's behaviour. The data is not top secret, though: I can PM it to you if you want.

 

[a space oddity]

Hi Karasu,

great job on the stats. One remark though, I had a diplo win with 1861 points, which is lower than the minimum winning score you mention.
Ok, not something to be proud of perhaps, but hey, recognition as a recordholder counts for something ...

Earth to A Space Oddity -These statistics are from Gotm16-Rome. Your 1861 points is from Gotm17 which has not even closed yet. - cracker

 

[a space oddity]

eh... thanks cracker, I guess I am so excited to enter the GOTM crowd that I got carried away. Methinks I return to my odd space now.

 

[Megalou]

We believe it is definitely more useful to present it this way, so that everyone can see where they fall in the plot and have a better understanding of their own playstyle and of the general trend, rather than discussing any single player's count and related playing habits

Fine, but how do I know my own number turns per reload?

Very neat figures, thank you.

 

[Freiherr]

Karasu,
Thank you for the fine work on the statistics. I look forward to seeing the numbers for GOTM 17. I would also like to add my request to those of jeffelammar and Megalou for an explanation of how the value of turns per session/reload are determined.
Thanks!

 

[Karasu]

Yes, GOTM 17 will open up new dimensions for stats -it also looks like a new submissions record, btw.

The GOTM staff approach regarding the Turns per Session figure is not to encourage public discussions of single player data -and I personally agree with it, as we have seen in the past that this may lead to misunderstandings that do not add anything to this forum.

Having said this, there is no reason to withhold the data. I think that most of us will have a fair idea of how many reloads they had, so that they can place themselves in the general picture. In any case, I can PM the number to you upon your request -we may have to change this approach if I get flooded with PMs, though...

 

[cavemanf16]

Originally posted by Karasu
Zachriel, yes, I guess so (unless of course you meant "fantastic", which is also true... )

Jeffelammar, the idea behind the TpS overall plot is to avoid public discussion of a single player's behaviour. The data is not top secret, though: I can PM it to you if you want.

I would love to have all the spreadsheets and data you have! I work with statistics and analysis in my job on a daily basis, and perhaps I could evaluate a couple other 'interesting' things that you may not have seen in the data yet. Extra eyes on the data always help!

Anyways, private message me the spreadsheets if you would. Thanks!

Also, I'm wondering if the 'hours of play' can be extracted from the PTW games? That might show some differences in scoring vs. real world time spent playing, thereby giving more of us a true idea of how much real world time needs to be spent on the game to achieve scores of 20000+.

 

[Freiherr]

Karasu,
Thanks for your reply. What I would like to know, however, is not the number for any individual game, but rather how one calculates the statistic in the first place. Thanks again!

 

[Karasu]

Right, sorry for misunderstanding you.

The calculation is made with a tool that was developed last year after some discussions on ways to spot "peculiar" games. It turned out, then, that most "cheated" games exhibit a distinctively low number of turns per session.
This, of course, does not mean that few turns per session automatically means that a game is not ok: if you have the patience of browsing through several old threads, you will see how the discussion developed.

I may seem a bit evasive here: the point is that I am not the best person to go into this, as I am a relative newcomer myself and I have had no role in developing that tool -the other staff have more knowledge and experience on this issue.

The one thing that I can say is that this information is read from the save file, where the number of times you load it is stored (but don't ask me how... I don't know!!).

 

[Freiherr]

Karasu,
That's what I wanted to know. I had no idea that the number of loads was stored in the save file. Thanks!