Life In 19x19

Pachi 12.90 pre-release is out:

• Seki fixes
• Slightly better playout balance
• Better random number generator
• GoGui: RAVE & Criticality visualization

Have fun =)

pnprog wrote: Thu Apr 02, 2026 11:49 pm Since Katago comes without a Joseki dictionary, how do you use get the information from Katago? Do you just assume that any move in a corner is a joseki move, or something similar?

Pretty much, yes. Early on every move in the corner is a joseki move =)

Could you expand on what is exactly the KataGo CPU build? how is it different with the former Pachi? how is this different with KataGo? will there be two branches of Pachi in the future, or is this KataGo version the only version of Pachi going forward?


Hi,
By default Pachi uses KataGo as ...

Pachi 12.86 pre-release is out:

• katago cpu build
• added --modern-joseki option
• many fixes (rave, playout, mm, seki ...)
• updated joseki fixes database

Have fun =)
Let me know if you see anything broken.

Hi,

If you enjoy playing with pachipachi bot on kgs there's a little fundraising going on at the moment to buy new hardware.
(The hardware it was running on for the last 5 years just died !)

Many people find it useful and it'd be great to run it as a community supported project (until now i've ...

Pachi 12.84:

• Added --banner option (replaces "banner" uct param)
• Fixed options that got broken in 12.80 (--kgs, --kgs-chat, --noundo, --name work again)
• Fixed memory corruptions found by address sanitizer
• Minor josekifix and mm fixes
• Updated joseki fixes database (2023-08-21)

Pachi 12.82:

• Retuned dcnn and pattern priors
• Use progressive dcnn prior
• Normalize pattern probabilities
• Set number of threads for dcnn evaluation

Retuned prior parameters, slightly stronger than 12.80 and should deal with dcnn blind spots better.
Have fun =)

Pachi 12.80 is out:

• Filter more dcnn blunders
• Better MM atari patterns
• Updated joseki fixes database (2023-05-12) + code
• UCT search speedup

Haven't tested extensively yet, if you see something unusual please let me know (create an issue on github).
Enjoy =)

Thanks And =)

Pachi 12.60 is out:

• fixed remaining seki issues
• fixed tree gc issues
• removed --accurate-scoring, unnecessary now
• --smart-pass: more user-friendly pass behavior

See homepage for more info.

Pachi 12.50 is out:

• fixed analyze issues in Sabaki
• automatically grow tree memory as needed
• fixed kgsgtp-3.5.20+ crashes (kgs-chat disabled by default now)

See homepage for more info.

Pachi 12.40 is out:

• speedup, about 20% faster
• multiple dcnns support: detlef44, detlef54, darkforest
• lz-genmove_analyze support (get winrates in sabaki etc)
• lz-analyze: handle frequency
• set engine options over gtp
• dcnn pondering fix

See homepage for more info.

Pachi 12.30 is out:

• lizzie support !
• dcnn pondering
• japanese rules: territory scoring in playouts
• full gtp undo

See homepage for more info.

update: Nov 06, Pachi 12.20 is out.

• dcnn: handle smaller board sizes
• don't play slack when taking handicap
• mm patterns fixes, better prediction rate
• use joseki engine when playing without dcnn

See homepage for more info.

Hi again !

How come they don't share the same winning percentage?

I was under the impression that Pachi was performing a sort of mini max algorithm, so the win rate of the complete "path" would be deduced from the deepest leaf (following a "mini max" propagation from deepest leaf to first leaf ...