Grand Strategy Victory Stats and Summary

rkr1958 · Post by **rkr1958** » Tue Feb 16, 2010 7:20 pm

In order to better assess play balance I would like to collect statistics from games played with the Grand Strategy Expansion. I've constructed a simple spreadsheet, which is shown below, for this. Also, a side benefit of this is that may help players to match up for new games. The more participation we get the better the statistics will be, which will let us know if one side or the other has an advantage. You can provide your data to me via PM, post or email.

Also, if a player resigns then estimate your victory.

rkr1958 · Post by **rkr1958** » Tue Feb 16, 2010 8:48 pm

Added Supermax.

rkr1958 · Post by **rkr1958** » Wed Feb 17, 2010 3:46 am

Added Borger and Tor.

rkr1958 · Post by **rkr1958** » Thu Feb 18, 2010 3:15 am

Added Joe Rock and corrected Borger's entry.

rkr1958 · Post by **rkr1958** » Fri Feb 19, 2010 12:19 am

Added Ray and summary plot.

rkr1958 · Post by **rkr1958** » Fri Feb 19, 2010 4:33 pm

To statistically evaluate the balance of CEAW - GS properly we do need as many samples as possible. For example, applying standard statistical techniques to the data above the margin for error for 9 samples (i.e., won-lost-draw) is +/- 27%. This means that we CANNOT say that the expansion is UNBALANCED (with 90% confidence) if the axis win percentage (or the allied win percentage) is between 23% and 77%. The confidence that I used to calculate the margin of error was 90%. There are two ways that one can make the margin of error smaller and that's to add more samples or decease the confidence level used to calculate it.

In the end we'd like to get the margin of error small enough such that we can say with 90% confidence that the chance to win as the axis is X% and the chance to win as the allies is Y% ASSUMING matched opponents. Hopefully X% and Y% will be within the margin of error of 50% or close enough to 50% to be considered balanced.

Anyway, you know what they say about statisticians.

There are liars, dang lairs and then there are the statisticians.

supermax · Post by **supermax** » Sun Feb 21, 2010 9:56 pm

Great job Ronnie

rkr1958 · Post by **rkr1958** » Sun Feb 21, 2010 10:30 pm

supermax wrote:Great job Ronnie

Thanks. Hopefully we'll get more samples so we can get more meaningful statistics.

supermax · Post by **supermax** » Sun Feb 21, 2010 11:28 pm

Whatever statistics, i find it hard to believe that you can loose often with germany, its really difficult for the allies ot get both capitals.

ncali · Post by **ncali** » Mon Feb 22, 2010 6:56 am

ncali · Post by **ncali** » Mon Feb 22, 2010 6:56 am

EDIT: Sorry for the triple post. The server seemed to be locking when I was trying to post the message below.

ncali · Post by **ncali** » Mon Feb 22, 2010 6:58 am

supermax wrote:Whatever statistics, i find it hard to believe that you can loose often with germany, its really difficult for the allies ot get both capitals.

The Italians will surrender without the Allies needing to take Rome. The new surrender conditions for Italy are much more realistic. Not that Mussolini could last so long anyway, once the Germans are conquered.

pk867 · Post by **pk867** » Mon Feb 22, 2010 10:22 am

The Italians will surrender when three cities are taken outlined in the players manual, but you still have to take Rome
for the victory conditions. Rome switches to Germany ownership when Italy surrenders by the three city method.

rkr1958 · Post by **rkr1958** » Tue Feb 23, 2010 4:06 am

Added Paul and margin of error calculation.

rkr1958 · Post by **rkr1958** » Sat Feb 27, 2010 4:00 pm

Stats and plots updated.

rkr1958 · Post by **rkr1958** » Sat Mar 06, 2010 6:23 pm

Stats and plots updated.

marcelscheele · Post by **marcelscheele** » Sun Mar 07, 2010 8:04 am

I'm a bit puzzled on the stats that are posted.

Why don't the wons and lossed & draws per side add up to a 100%?

In my opinion a game not won, is either lost or drawn, but now this set adds up for the axis to 91,7% of all games and for the allies to a stunning 108,3%?

Or am I misreading the numbers?

Peter Stauffenberg · Post by **Peter Stauffenberg** » Sun Mar 07, 2010 11:47 am

The reason you don't get losses + won = 100% is because the people who tend to report their games are the ones who win. Those who lose don't do it. So the report is biased because of that.

marcelscheele · Post by **marcelscheele** » Tue Mar 09, 2010 6:38 am

I'm not sure I follow that.

The list of total wins, losses and draws does add up to a 100%, so that data would be fully reported and the split between axis and allies wouldn't?

And in the case of the allies, how can non-reporting add the total over 100%.

Again, I might be missing something, but it does not seem to be non-reporting.

Peter Stauffenberg · Post by **Peter Stauffenberg** » Tue Mar 09, 2010 7:24 am

Let's say I play against you and win. I report to Ronnie my result and the chart is updated with another victory for me. You, however decide not to report your loss. That means we don't have your data in the chart. So we have added the victory, but not the loss.

Remember that the chart shows the results of the players who have reported their games NOT the results of all the games.