Analysis of a lot of BGA results

[laggercat]

I've posted up a couple of threads on Boardgamegeek on 7 wonders (2nd), looking at the statistics on BGA.

https://boardgamegeek.com/thread/269712 ... unbalanced

https://boardgamegeek.com/thread/267386 ... -stats-bga

I have looked at around 11,000 results.

The brief summary is that Hali B is the strongest wonder. It does not matter how many players are in the game or their BGA ratings. Olympia B is the worst, although its a bit closer and arguably not the worst for some player counts.

For four player games between players all with ratings of 300+ (average rating of 380, mostly played on Arena and many involving Wonderful Plays, Iguana, RedMeeple, giog21, birdbirdtree, Pragyatf, ZeroGravity etc), Hali B is comfortably the best and Olympia B is comfortably the worst in every statistic. In these games, Hali B wins 41 per cent of games it is in and Olympia B wins 10 per cent of games it is in. Its an enormous difference.

[doctorchris]

Thank you. Very helpful analysis. Confirms the gut feelings a lot of us have had.

[laggercat]

Now nearly 15,000 results entered - you can see the updated data in the third last post of the thread:

https://boardgamegeek.com/thread/269712 ... unbalanced

[Wonderful Plays]

First let me thank you again for doing excellent job for the community.

I have 2 takeaways from these stats:
1. Hali B is too good - we knew this in previous edition - we expected it in new edition, we got multiple confirmations before, here there is another big one - it is too good.
2. Other than that the numbers are pretty balanced

If we exclude Hali B (and Olympia B whenever it is possible as Olympia A is prefered among top players for a reason):
A. From 3p wonders average place is between 1.9 and 2.1 (in an ideal world everything would be 2.0) - actually amazing.
B. From 4p wonders average place is between 2.3 and 2.8 (in an ideal world everything would be 2.5) - good
C. From 5p wonders average place is between 2.6 and 3.3 (in an ideal world everything would be 3.0) - good (here exclusion does nothing bc Rhodes)
D. From 6p wonders average place is between 3.3 and 3.9 (in an ideal world everything would be 3.5) - good (couldnt exclude olympia)
E. From 7p wonders average place is between 3.3 and 4.8 (in an ideal world everything would be 4) - bad but here without Olympia B it would be 4.3 and 7 players is considered too random among top players.

The simplest solution is just adding an option for ban for the B side of Hali.

[laggercat]

Fun fact:
In all the games I've analyzed, there is only two where a player has finished last in a game where their rating was higher than the sum of the ratings of all the other players in the game (before the game).

253 Markvg
110 YHShaw
436 Pragyatf
https://boardgamearena.com/table?table=184692358

100 compote de pomme
131 antho051509
125 DejaVu_NewYork
385 Realism
https://boardgamearena.com/table?table=192947703

The closest for a five player game is this (me - oh, the shame)
115 One1Five5
178 Marouschka
100 ZIHNAN
188 a28123015
334 laggercat
https://boardgamearena.com/table?table=186878884

[laggercat]

More fun facts.

The five fastest players I have found (median minutes to play a game) are (2.50 = 2 mins and 30 seconds):

2.30 Dvlta
2.30 Killer danus
2.33 ina-tensei
2.53 LauMaia
2.57 giog21

The five slowest players are:

4.77 Le Sconce
4.72 smoth5431
4.35 Jeremy55
4.33 Chaniju
4.25 Biweck

The median time for a 400+ player is 3.07 minutes

The fastest time to play any game (out of those I have found) are these two, where the players (ina-tensei and Dvlta) each took 1 minute and 36 seconds. Both won the game.

https://boardgamearena.com/table?table=194661179
https://boardgamearena.com/table?table=188313859

[laggercat]

First let me thank you again for doing excellent job for the community.

I have 2 takeaways from these stats:
1. Hali B is too good - we knew this in previous edition - we expected it in new edition, we got multiple confirmations before, here there is another big one - it is too good.
2. Other than that the numbers are pretty balanced

If we exclude Hali B (and Olympia B whenever it is possible as Olympia A is prefered among top players for a reason):
A. From 3p wonders average place is between 1.9 and 2.1 (in an ideal world everything would be 2.0) - actually amazing.
B. From 4p wonders average place is between 2.3 and 2.8 (in an ideal world everything would be 2.5) - good
C. From 5p wonders average place is between 2.6 and 3.3 (in an ideal world everything would be 3.0) - good (here exclusion does nothing bc Rhodes)
D. From 6p wonders average place is between 3.3 and 3.9 (in an ideal world everything would be 3.5) - good (couldnt exclude olympia)
E. From 7p wonders average place is between 3.3 and 4.8 (in an ideal world everything would be 4) - bad but here without Olympia B it would be 4.3 and 7 players is considered too random among top players.

The simplest solution is just adding an option for ban for the B side of Hali.

Here are the results for 4 player games where EVERY player in the game has a BGA ratings of at least 250, and where Hali B was NOT in the game.

rank win% last%
2.5 25.8 24.7 Olympia A
2.6 13.7 25.5 Alex A
2.5 23.9 21.0 Giza B
2.4 33.3 22.8 Babylon B
3.0 10.3 44.5 Olympia B
2.2 32.9 14.6 Rhodes B
2.5 26.0 22.1 Ephesus B
2.7 19.7 33.1 Alex B

Olympia B is still very weak, but Olympia A looks fine. Rhodes is strong.

Now for games where there is no Hali B or Olympia B (there is not enough data to show Alex A):

rank win% last%
2.5 25.8 24.7 Olympia A
2.6 18.6 22.5 Giza B
2.4 32.4 26.7 Babylon B
2.2 34.4 18.4 Rhodes B
2.5 24.2 25.8 Ephesus B
2.8 15.0 35.5 Alex B

The main concern here is that Alex B is a bit weak - though not as bad as Olympia B.

Three player games:

rank win% last%
2.1 28.0 36.0 Olympia A
2.0 28.6 26.5 Giza B
1.9 42.0 31.3 Babylon B
1.9 40.4 30.3 Rhodes B
1.9 33.3 22.2 Ephesus B
2.1 33.3 47.1 Alex B

This generally looks ok.

[Wonderful Plays]

That was what I was going to ask about, but forgot in my post! Thanks so much, those stats are more meaningful that my ramblings in the previous post.

I would say Rhodos and Olympia are two wonders that change when you become better. The better the players the worse Rhodos is and better Olympia is.

There is a lot of points that can be seen from the stats that confirm some of my (and other top players) observations:
- Rhodos B is strongest at 5P (due to combination of fairly low red in the pool for 1st 2 ages and the ability to build stages thanks to double browns)
- Alexandria B is fairly weak, but with higher player counts it gets better (more reliable than some other wonders who have troubles with resource spreads in the higher player counts)
- Giza is best at 4

I think the main problematic one is Halikarnassos and it is not only the power level, but the feeling that it is very hard to play against it. You cannot even throw, because Hali will just access the discard. I plan to make some NHB (no hali b) tournaments and see how it goes.

[laggercat]

I would say Rhodos and Olympia are two wonders that change when you become better. The better the players the worse Rhodos is and better Olympia is.

I've put up a post on BGG on this, at the end of the thread: https://boardgamegeek.com/thread/269712 ... unbalanced

Comparing results between players rated <200 and rated >300, there is only a big difference in relativities for Rhodes A. There's lot much difference for the other wonders.

[laggercat]

More fun facts:

Highest and lowest scores for each placing:

First number is the highest score I've found for that placing and the second number is the lowest score I've found for that placing

3 player

1st 81 46
2nd 69 38
3rd 65 31

4 player

1st 84 45
2nd 69 38
3rd 63 36
4th 59 8

This is the game where ina-tensei scored 8:
https://boardgamearena.com/table?table=193309583

Not counting this, the lowest 4th placed score is 25

5 player

1st 86 48
2nd 70 46
3rd 63 40
4th 61 31
5th 57 26

6 player

1st 88 51
2nd 69 49
3rd 66 46
4th 61 42
5th 58 34
6th 55 27

7 player

1st 89 51
2nd 71 48
3rd 67 46
4th 62 43
5th 59 37
6th 57 33
7th 51 19

Here is the game with a score of 89 (Hali B, of course): https://boardgamearena.com/table?table=186834213