Handicap For Smaller Board Sizes/ Discussion

Path: <= Handicap =>

Charles Tim was a Cambridge player for many years. This is a 'smoothed' version of the system, adopting a Japanese practice of reverse komi. Actually I think this makes it easier for White.

Tim Hunt There is more of the same on my web site: [ext] http://www.timhunt.me.uk/go/handicaps/

Vesa We had recently a round-robin 9x9 tournament with 14 players from 15 kyu to 5 dan. The results indicate that the handicap is not easy for White. The results are available in [ext] Helsinki Go Club page

The top right vs. bottom left is 36.5 - 54.5. This might not be statistically good enough sample but anyway some kind of response to the "easier for white" argument.

Charles The page [ext] http://jakobsen.dyndns.dk/edo/pages_english/handicaptables.htm has the full handicap table.

The politics is usually that the stronger players argue about the system until White has an easier time. 'Correct' handicap systems will mean the events are won by undergraded players. After a while the stronger players normally have understood more about what you have to do, to win.

KichiGuy: In my first attempt to figure out how to adjust Komi by board size, I just naively tried adjusting komi by difference in board area to see what it would come up with for a 9x9 board:

(9 * 9) / (19 * 19) = 0.224 (rounded)

Take 6.5 for the 19x19 and multiply by 0.224:

6.5 * 0.224 = 1.458 or about 1.5

This kind of sounds reasonable to me but I just started learning Go and don't know anything about the math.

On another note, doing this area adjustment for handicap stones (2 / 0.224 = 8.929) makes it seem like 2 handicap stones on a 9x9 board is similar to 9 stones on a 19x19 board. Again though, I don't know if adjusting these things by proportional difference in area is mathematically sound.

This conflicts with Tim Hunt's recommendations on komi, but seems roughly similar to the table he provided for handicap stones.

Can anyone comment on this?


ferdi:On DGS, I have two accounts only for 7x7 and 9x9 games (all games as 2 dan, nearly all at proper handicap):
ferdi9: won 99, lost 104, jigo 3
ferdi7: won 87, lost 79, jigo 4
So, at least for me, the DGS system is fine for 7x7 and 9x9 :)


(moved from main page)

I've been thinking about komi settings for smaller boards. The logic behind modifying the number of handicap stones given seems to be that one stone is worth more on a smaller board. By the same reasoning, shouldn't komi be higher for smaller boards, because the one initial black stone will also be worth more?

Evan: Your presumption is wrong, the idea behind modifying the number of handicap stones is because there is less space available for White to attempt to negate Black's initial advantage. So relative to the whole board the handicap stones are bigger (= are more valuable), but there (their) actual(l)value is roughly the same. All this said, I do believe the komi should be 1 or 2 points more for 13x13, the hypothesis behind this is that because the sides are all closer together the inital moves are so big as they influence all 4 corners at the same time.

Tderz: I agree with the position of the first person, but do these two positions above differ at all?

I feel that both express the same:

  • On smaller boards, the value of a move (or tempo/sente) expressed in percentage of total points available on the board increases (19*19: komi is 6.5/361 = 1.8% of total points available)
  • Therefore the komi on a 13*13 board should be bigger than the directly derived 6.5 * (13*13/19*19) = 6.5*47% = 3.04 > 3 points.

3.5 komi would be 2%, 4.5 komi would be 2.7% of the total 169 points available.
The 5.5 komi from the table above equals a 3.3%

These are just hypothesises. Eventually one adapts the handicap+/-komi empirically. There will be persons stronger on small boards, as there are preferences for long or Blitz games.


Path: <= Handicap =>
Handicap For Smaller Board Sizes/ Discussion last edited by axd on March 30, 2005 - 00:37
RecentChanges · StartingPoints · About
Edit page ·Search · Related · Page info · Latest diff
[Welcome to Sensei's Library!]
RecentChanges
StartingPoints
About
RandomPage
Search position
Page history
Latest page diff
Partner sites:
Go Teaching Ladder
Goproblems.com
Login / Prefs
Tools
Sensei's Library