Victor Kulibaba
Russia
flag msg tools

Hello, my dear BSW players - current or future...

Probably this subject was covered before here, then please give me a link in reply, though I'd like to share my thoughts on BSW raiting system and how it relates to real player strength.

Consider any 2 player game started between Player1 with 20% wins (in this game) and Player2 with 60% wins. What would be real probability to win for each of them? Unfortunately, you can't just say that Player2 is 3 times stronger than Player1 because 60% = 20% x 3 and thus his chance to win is also 3 times higher and equal 75% (in relation to 25% of Player1). Actually, this question has no real answer because of lack of information, and there is no prediction of game result possible. Real probablility to win for Player2 in this example may vary from ~0% to ~100% due to additional factors involved, an I will cover some of them below.

Factor 1 - Average opponent strength

Consider the case that before the game mentioned above happened Player1 only played with very strong players and still got 20%, and Player2 only played with complete noobs (being in fact one of them) and randomly got 60%. In this case Player1 will more likely win against Player2.

Nice conclusion 1
When you see someone with many games played and unbelievable win rates like 90% and above, it just directly says that he mostly played with much weaker players, but say nothing about his strength against you. You can only know your chance after several games with him, manually calculating win rate in this serie.

Nice conclusion 2
Consider group of professional players staying on top of the game, with approximately equal strengths, playing with each other to have fun, learn and progress even more, not bothering to spend time on easy beatings of noobs. Suprisingly, average win rate for all of them in 2 player game will not go much above 50%. For 3 player games it will be even less - around 33%.

Factor 2 - Win rate inertia

This factor is affecting even more or less constant player groups. Consider someone has just learned rules of new game and started playing on BSW right away with strong opponents (to learn and improve skills fast). First 100 games he mostly lost to experienced players, but still got some lessons and won 3 of them, so his win rate at this point was 3%. From this point, continuing to learn in next 100 games he managed to win 20 of them (by that showing actual 20% win rate), but his total win rate became only (3 + 20) / 200 = 11.5% due to inertia. If he played last 100% from another account - win rate for that account would have been 20%. Going further, consider that after 200 games our player has got enough experience and become a strong player in his group with 60% winning chance, so in next 100 games he won 60 (being equal with good players and winning against novices). But in total his win rate is now only (3 + 20 + 60) / 300 = 27.6% and so on...

In fact win rate inertia is so strong, that sometimes it's almost impossible to increase you win rate even having multiple consequent 100% wins. In my experience I was on ~69.03% win rate in Notre Dame (2 player game) with 465 games played and 321 won. How much games do you think I needed to win in a row (without single loss) to pass 70% win rate? Let's solve:

(321 + X) / (465 + X) >= 0.7
0.3*Х >= 465*0.7 - 321
X >= 4.5 / 0.3
X >= 15

15 games in a row!! (showing real current 100% win rate) just to step 1% further in total profile win rate. That's a tough task.

Factor 3 - Current player condition

Even strong players can sometimes be drunk, or be in a bad psychological state for playing, or trying new crazy (hardly winning) strategy just for fun, or allow their cat to play from their accounts and so on... Vise versa, average players can sometimes be inspired an play better than usual. These all can strongly affect resulting chance to win, but as you see is not realated to total win rates of players at all.

Factor 4 - Allies (works for games with 3+ players)

Consider 1 strong and 2 weak players are playing together. In most games on BSW 1 weak player can help another weak player to win (by intention or randomly by lack of knowledge) so that strong player has not much to do against it.

There are many other factors, though above are probably the strongest and I will stop here.

To conclude, I wish you good luck in your games on BSW and urge not to judge players strengths and your chances to win in particular game by comparing win rates of players :-)

Cheers,
Poher
3 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Front Page | Welcome | Contact | Privacy Policy | Terms of Service | Advertise | Support BGG | Feeds RSS
Geekdo, BoardGameGeek, the Geekdo logo, and the BoardGameGeek logo are trademarks of BoardGameGeek, LLC.