This match has inspired yet another great commentary, this one by Rob van Zeijst. In his announcement on Facebook, he prefaced the article and commentary with: "To all Go Associations and Go Lovers: If you want to publish this article in your own language for free, it is possible, but send me an email for the conditions: rob.yourteacher[at]gmail.com or send a message to Rob van Zeijst on Facebook. Thank you"
http://www.kiseido.com/printss/guliten1.html
Commentary on the first game by Rob van Zeijst
-
daal
- Oza
- Posts: 2508
- Joined: Wed Apr 21, 2010 1:30 am
- GD Posts: 0
- Has thanked: 1304 times
- Been thanked: 1128 times
-
macelee
- Lives in sente
- Posts: 928
- Joined: Mon Dec 31, 2012 1:46 pm
- Rank: 5 dan
- GD Posts: 0
- KGS: macelee
- Location: UK
- Has thanked: 72 times
- Been thanked: 480 times
- Contact:
Re: Commentary on the first game by Rob van Zeijst
I like the way the commentary is presented. Any idea how that web page was generated?
-
Knotwilg
- Oza
- Posts: 2432
- Joined: Fri Jan 14, 2011 6:53 am
- Rank: KGS 2d OGS 1d Fox 4d
- GD Posts: 0
- KGS: Artevelde
- OGS: Knotwilg
- Online playing schedule: UTC 18:00 - 22:00
- Location: Ghent, Belgium
- Has thanked: 360 times
- Been thanked: 1021 times
- Contact:
Re: Commentary on the first game by Rob van Zeijst
The best argument against odd numbered game series is that, if a 6-5 win in an 11 game series suffices to declare a winner, then 1 game would be sufficient too. Here's my proof:
Let X be the better player and Y the other player. Let p be the probability that X wins a game against Y (p>0,5).
The game series has a double purpose: find out who is X, but also get a better idea about p.
Our current estimate would be Lee Sedol being X and p is determined by the 18-17 score in his favor.
At 5-5, we would still estimate Lee Sedol being stronger but p is now determined by a 23-22 score in his favor.
If Lee Sedol wins 6-5, he clinches the title while the probability that they are actually of equal strength, or even Gu is slightly stronger, is higher than if Lee Sedol would win 1-0.
So it would be fairer to make Lee decisively stronger after 1-0 than after 6-5.
With 6-4 it's all different. Now the probability that Gu is better, given 24-21, is lower than given 19-17.
Similarly, it would be unfair to give Gu the title at a 6-5 win (being an overall 23-23 tie), than at a 1-0 win (18-18 tie), because chances they're equal are even increased after the 11 game series.
An even jubango only yields a winner if there are at least 2 games difference. The difference between the probabilities to indicate a wrong winner in the 6-4 case and the 6-5 case is quite high.
Let X be the better player and Y the other player. Let p be the probability that X wins a game against Y (p>0,5).
The game series has a double purpose: find out who is X, but also get a better idea about p.
Our current estimate would be Lee Sedol being X and p is determined by the 18-17 score in his favor.
At 5-5, we would still estimate Lee Sedol being stronger but p is now determined by a 23-22 score in his favor.
If Lee Sedol wins 6-5, he clinches the title while the probability that they are actually of equal strength, or even Gu is slightly stronger, is higher than if Lee Sedol would win 1-0.
So it would be fairer to make Lee decisively stronger after 1-0 than after 6-5.
With 6-4 it's all different. Now the probability that Gu is better, given 24-21, is lower than given 19-17.
Similarly, it would be unfair to give Gu the title at a 6-5 win (being an overall 23-23 tie), than at a 1-0 win (18-18 tie), because chances they're equal are even increased after the 11 game series.
An even jubango only yields a winner if there are at least 2 games difference. The difference between the probabilities to indicate a wrong winner in the 6-4 case and the 6-5 case is quite high.