1) Is there any existing data compilation / analysis of the win rates of various countries (for this site)? I know that such a breakdown exists for games played on the vdiplomacy website, but the overall sample size is very small (only 18 games).

2) Is there any existing data compilation / analysis of the various openings? I.e. a quantification of each possible opening by each country. (e.g. Moscow moves to Ukraine)

3) Is there a way to access cross-tables or datasets of game results from this website so that an entrepreneurial individual could theoretically create such an analysis if it did not currently exist? Collecting the information from the game search interface is obviously not feasible.

I did a preliminary search and came up empty-handed, but maybe people more involved with the site might know things I don't! :)

Thanks!

This might help from the 18 games of it finished on vdip. http://vdiplomacy.net/stats.php?variantID=2

I think I did an un-scientific look at this many years ago. Jmo would be able to help you much better than anything I could dig up.

From memory though, the results were that Kenya was preeeeeetty good.

Yeah, my hypothesis is that Quebec, Kenya, China (for example) are pretty good draws, whereas Antarctica, USA, (and especially Pacific Russia) are not, but there really isn't enough data points on vdip to make that conclusion more scientifically.

It also seems like there are a much bigger variety of openings per country than most variants, but only for some countries.

Ideally it would be great to exclude gunboats and games with CDs, etc. or at least theoretically be able to.

My personal assessment:

A tier:
Western Canada
Brazil
Ghana
Kenya
India

B tier:
Quebec
Argentina
South Africa

C tier:
Europe
Russia
Pacific-Russia
Near East
Oz

D tier:
USA
Frozen

F tier:
Libya
China

Wow, interesting! Why so low for Libya & China?

My list is more or less the same as swordsman's. I'd probably swap Argentina and Brazil, as well as China and USA. I seem to remember lots of solos for Argentina, but a worse draw rate than Brazil. Basically meaning if Argentina can beat Brazil, it's really good, but more often than not Brazil kills it.

Libya is sandwiched between an ocean and two of the best draws on the board. It has little to no chance of taking control of the Med or tackling the Near East. By far the worst nation on the board.

China I wouldn't put on F tier bad. But I'd definitely rather have any one of it's neighbors

I think Libya is somewhat akin to Austria. Definitely a high death rate, but great expansion possibilities if it can survive and lots of variability. Lower tier but not the worst.

I guess I was more wondering about China; it seems like a pretty good country to play. Highly defensible and at an advantage versus both PR & Oz in terms of potential conflicts.

You can tell just looking at the board that Libya is at an enormous disadvantage when it comes to picking up neutral SCs.

I once made a data compilation of 50 games and put them on the forum. That was about one or two years ago. You can try to find it back. I remember that in FP DSS Ghana appeared to be the best country, and China the worst. Western-Canada, Quebec and Kenya scored high as well, while Libya, South-Africa and Brazil scored low. Unfortunately, the document in which I saved all the data is lost due to a laptop crash.

How did you create the compilation Mercy?

China is a really bad country because it is by far the most logical target for Pacific-Russia. And while Pacific-Russia is not a logical target for any other country, it is really tempting for India to attack China when he knows that Pacific-Russia will attack China, too.
I think I would be comfortable to play any country except China. I wouldn't know what to say to convince Pacific-Russia and India not to kill me.

@Friendly Sword
I put the data from the game search interface by hand in Wolfram Mathematica and wrote a line of code to give each country a score.

Hi Mercy! I've copied your previous analysis below in case anyone is interested. I am surprised about China primarily because Pacific Russia is at a disadvantage in a 1-on-1 fight, and Russia has even *more* of an incentive to snatch the only other centre PR has apart from those potentially blocked by China. I think India generally has an edge over China, but it's not decisive due to the difficulty of securing Lanzhou quickly (even when China is distracted). But.... that's why I'd love to see the stats!

Mercy's post copied below:

"""I looked at the results of 34 full press Draw-Size Scoring [World Diplomacy IX] games and 17 full and rulebook press Sum-Of-Squares Scoring [World Diplomacy IX] games and I listed how well the different countries did.

Here are the scores. A high score is good, a low score is bad, and a score of 1 is average.

Draw-Size Scoring Scores:

1. Ghana 1.69
2. India 1.54
3. Quebec 1.49
4. Western-Canada 1.22
5. Oz 1.13
6. Argentina 1.12
7. Europe 1.12
8. Kenya 1.08
9. USA 1.04
10. Near-East 0.98
11. South-Africa 0.92
12. Frozen 0.91
13. Russia 0.71
14. Pacific-Russia 0.61
15. Libya 0.58
16. Brazil 0.48
17. China 0.37

Sum-Of Squares Scoring Scores:

1. Ghana 3.49
2. Russia 2.21
3. India 1.52
4. Western-Canada 1.45
5. Kenya 1.39
6. Quebec 1.02
7. Brazil 0.91
8. Argentina 0.89
9. Frozen 0.74
10. Libya 0.68
11. Oz 0.59
12. Near-East 0.52
13. Europe 0.48
14. Pacific-Russia 0.49
15. USA 0.34
16. China 0.21
17. South-Africa 0.06

The Draw-Size Scoring games were all* the full press games that started before gameID=180270 up to gameID=158048. In every game I noted the percentage of the bet players got back; so for instance, being part in a n-way draw gives a percentage of n/17. The score of every country is the average percentage the country got back in the 34 games. I used the following tweeks:
- *Some games had a lot of CD's at the start and were drawn quickly after that. I didn't include these games.
- In both games that somebody won, the win was due to NMRs/CD's. I used a percentage of 17/2 for these winners and split the rest evenly between the players I thought would have been part of the draw if there were no NMRs in the endgame.

For Sum-Of-Squares Scoring, I looked at the first 17 full and rulebook press games on this site. For Sum-Of Sqares, I also looked at the percentage of bets players got back. I didn't need to use any 'tweaking'. The results of Sum-Of Squares Scoring are less reliable than the other results because of less games used and because the scoring system results in a higher variance in points won per game per country."""

Also, I don't get why Argentina is called stronger than Brazil. It simply is not. If Argentina and Brazil battle 1v1, Argentina has the advantage since he can block a build from Brazil.

Ah. Thanks for the search!

In my second post above this one, I meant to write: "I don't get why *Brazil is called stronger than *Argentina".

This is ignoring the primary determinant of success in World: the country that wins is the one with the most neighbors going into CD.

In my opinion, China is the worst country in World Dip IX to the point where I think a balancing fix is necessary for the game to be fair. The only game I can ever recall drawing China and not getting eliminated is a game where Pacific Russia had multiple NMRs in the first couple of turns -- which makes sense because the imbalance against China is that Pacific Russia takes little risk in all-out attacking China at the beginning. (China also offers this incredible, juicy target at the beginning for India as well, and India rarely faces early game trouble from any power.)

This would be a really tough balancing fix because I think Pacific Russia is also a pretty crappy power. The usual course of events is that Pacific Russia attacks China together with India, only to later be pushed out of China by India and attacked from another direction by Russia or Western Canada.

As for my tier list, maybe I'm wrong about Brazil vs. Argentina, according to these empirics. I've just seen so many games where Brazil rapidly dominated South America and faced no pressure from any power other than Argentina (who might be attacked by South America or Frozen). Maybe my experience has been unrepresentative.

Libya is a bad power like China. Libya functions as a lightly-populated neutral area for African powers and Near East to fight over.

That said, I guess I could persuade myself that Libya is slightly better than China because Libya at least has a central position to expand into several continents if somehow not eliminated early on. For example, the only win I ever got in World Dip IX was as Libya in a Gunboat game, and nobody CD'd except the very distant power Western Canada, whose replacement almost ended up fighting me to a draw: http://webdiplomacy.net/board.php?gameID=185901

In that game, through incredible fortune and insane guesses I managed to kick apart my neighbors and, once that happened, I expanded really easily and invaded nearly all the continents on the board.

China contrasts very unfavorably. China not only has the difficulty of being an immediate target that Libya has (such that China functions as some neutrals for them to fight over), but China also doesn't have a good expansion path (unlike Libya).

@Ogion: Just play with minimum RR requirements and CD's will be a lot less of a problem.

@swordsman: I agree with your analysis. Speaking about balancing fixes, one year ago I worked on a remake of the World Diplomacy IX variant with the intention to get it implemented on vDiplomacy; I mainly made countries have more effective neighbours in order to create more balance and make the game more exciting. But it was quite a large project and I didn't finish it, but maybe I will pick it up again in the future.