Eeny, meeny, miny, moe, go is not spelled goe

rec.games.go: Checklist Go

2.4.2002 7:45

Xenafan wrote:

1. absolutely forced moves (anything which is obviously urgent)
2. thick plays (if there's a weakness)
3. influence
4. oba (for the purposes of this list, an oba is defined as any move  which is mutually big)
5. large territorial plays (this can include invasions and keshi)
6. others (e.g., gote-no-sente, kikashi)

I think you should use higher priority for kikashi.

I suggest following procedure after you have decided your move: Now you have more information about the future of the game than your opponent. Think if there are some plays in other parts of the board that your opponent would answer your moves differently if he knew your next move. Then play all "non-harmful" sente-enough kikashi plays of this type before your actual move.

Also I would suggest that always try to visualize a couple of moves ahead after decision and before placing the stone to avoid rash mistakes we amateurs do so often.

Use your opponent's time for life&death considerations, calculate territories etc.

Also in the late endgame forget the rule reverse-sente is twice the value of gote. It is not accurate enough if you just have a couple of minutes for thinking. Quite simple and as far as I understand better algorithm for simple endgame is to put gote plays in descending order g1,(g2),g3,(g4),g5, .. and calculate the difference of sum of odd and even indexed lists of gote plays and compare this to the size of biggest reverse-sente. Now you can see that the value of gote is something between 0-1 times the actual value of the move (and reverse-sente is its actual "gote" size). But more importantly: Most mistakes in the endgame of amateurs at my level (2k*) are misreadings of size of the move, not the actual order of moves, I think.

If you want to apply this algorithm to the beginning or the middlegame, maybe you could just pick the couple of biggest gote looking moves, evaluate their approximate size and do the subtraction to these moves (like above) and approximate the end of the gote-move list as 0.5*biggest remaining gote (0.5 is the average size of the gote play - My guess is that this approximation of gote size will be much more accurate in "slow" territory contest games than in fighting/mojo games). It is also nice to see how the concept of miai comes from this algorithm.

 (Good Timin'-Jimmy Jones-original song-1960)