1 Pachi can refer to: a simple modular framework for programs playing
2 the game of Go/Weiqi/Baduk, and a reasonably strong engine built
9 To build Pachi, simply type:
13 The resulting binary program `pachi` is a GTP client; connect to it
14 with your favorite Go program interface (e.g. gogui or qgo), or use
15 kgsGtp to connect it to KGS. (DO NOT make the GTP interface accessible
16 directly to untrusted users since the parser is not secure - see the
17 HACKING file for details.)
19 The pachi program can take many parameters, as well as the particular
20 engine being used; the defaults should be fine for initial usage,
21 see below for some more tips.
23 In case you hit compilation issues (e.g. when building on MacOS/X)
24 or want to change the build configuration, check the user configurable
25 section at the top of the Makefile.
31 The default engine plays by Chinese rules and should be about
32 3d KGS strength on 9x9. On 19x19, it might be about KGS 2k, assuming
33 reasonable hardware, e.g. two-core Athlon64 machine. On a higher-end
34 (e.g. six-way Intel i7) machine, it can hold a solid KGS 1d rank.
35 When using a large cluster (64 machines, 20 cores each), it maintains
36 KGS 3d and has won a 7-stone handicap game against Zhou Junxun 9p.
38 By default, Pachi currently uses the UCT engine that combines
39 Monte Carlo approach with tree search; UCB1AMAF tree policy using
40 the RAVE method is used for tree search, while the Moggy playout
41 policy using 3x3 patterns and various tactical checks is used for
42 the semi-random Monte Carlo playouts.
44 At the same time, we keep trying a wide variety of other approaches
45 and enhancements. Pachi is an active research platform and quite a few
46 improvements have been already achieved. We rigorously play-test new
47 features and enable them by default only when they give a universal
51 By default, Pachi will run on a single CPU core, taking up to 1.4GiB
52 of memory, not pondering and taking completely arbitrary amount of time
53 per turn. You can adjust these parameters by passing it extra command
56 ./pachi -t _1200 threads=8,max_tree_size=3072,pondering
58 This will make Pachi play with time settings 20:00 S.D. (unless it
59 gets told otherwise over GTP), with 8 threads, taking up to 3GiB
60 of memory (+ several tens MiB as a constant overhead) and thinking
61 during the opponent's turn as well.
63 Pachi can use an opening book in a Fuego-compatible format - you can
64 obtain one at http://gnugo.baduk.org/fuegoob.htm and use it in Pachi
65 with the -f parameter:
67 ./pachi -f book.dat ...
69 You may wish to append some custom Pachi opening book lines to book.dat;
70 take them from the book.dat.extra file. If using the default Fuego book,
71 you may want to remove the lines listed in book.dat.bad.
73 For now, there is no comprehensive documentation of options, but you
74 can get a pretty good idea by looking at the uct_state_init() function
75 in uct/uct.c - you will find the list of UCT engine options there, each
76 with a description. At any rate, usually the three options above are
77 the only ones you really want to tweak.
80 Except UCT, Pachi supports a simple idiotbot-like engine and an example
81 treeless MonteCarlo-player. The MonteCarlo simulation ("playout")
82 policies are also pluggable, by default we use the one that makes use of
83 heavy domain knowledge.
85 Other special engines are also provided:
86 * a "distributed" engine for cluster play; the description at the top of
87 distributed/distributed.c should provide all the guidance
88 * a simple "replay" engine that will simply play moves according
89 to the playout policy suggestions
90 * few other purely for development usage
93 Pachi can be used as a test opponent for development of other go-playing
94 programs. For example, to get the "plainest UCT" player, use:
96 ./pachi -t =5000 policy=ucb1,playout=light,prior=eqex=0,dynkomi=none
98 This will fix the number of playouts per move to 5000, switch the node
99 selection policy from ucb1amaf to ucb1 (i.e. disable RAVE), switch the
100 playouts from heuristic-heavy moggy to uniformly random light, stop
101 prioring the node values heuristically, and turn off dynamic komi.
103 You can of course selectively re-enable various features or tweak this
104 further. But please note that using Pachi in this mode is not tested
105 extensively, so check its performance in whatever version you test
106 before you use it as a reference.
108 Note that even in this "basic UCT" mode, Pachi optimizes tree search
109 by considering board symmetries at the beginning. Currently, there's no
110 easy option to turn that off. The easiest way is to tweak board.c so
111 that board_symmetry_update() has goto break_symmetry at the beginning
112 and board_clear has board->symmetry.type = SYM_NONE.
118 Pachi can also help you analyze your games by being able to provide
119 its opinion on various positions. The user interface is very rudimentary,
120 but the ability is certainly there.
122 There are currently several Pachi interfaces provided for this purpose.
127 Pachi can evaluate all moves within a given game and show how
128 the winrates for both players evolved - i.e. who was winning at which
129 game stage. This is implemented using the `tools/sgf-analyse.pl` script.
130 See the comment on top of the script about its usage.
135 Pachi can evaluate all available moves in a given situation
136 and for each give a value between 0 and 1 representing perceived
137 likelihood of winning the game if one would play that move. I.e. it can
138 suggest which moves would be good and bad in a single given situation.
140 To achieve the latter, note the number of move at the situation you
141 want to evaluate and run the `tools/sgf-ratemove.sh` script.
142 See the comment on top of the script about its usage.
148 The aim of the software framework is to make it easy to plug your
149 engine to the common infrastructure and implement your ideas while
150 minimalizing the overhead of implementing the GTP, speed-optimized
151 board implementation, etc. Also, there are premade random playout
152 and UCT tree engines, so that you can directly tweak only particular
153 policies. The infrastructure is pretty fast and it should be quite
154 easy for you (or us) to extend it to provide more facilities for
157 See the HACKING file for a more detailed developer's view of Pachi.
159 Also, if you are interested about Pachi's architecture, algorithms
160 etc., consider taking a look at Petr Baudis' Master's Thesis:
162 http://pasky.or.cz/go/prace.pdf
168 Pachi is distributed under the GPLv2 licence (see the COPYING file for
169 details and full text of the licence); you are welcome to tweak it as
170 you wish (contributing back upstream is welcome) and distribute
171 it freely, but only together with the source code. You are welcome
172 to make private modifications to the code (e.g. try new algorithms and
173 approaches), use them internally or even to have your bot play on the
174 internet and enter competitions, but as soon as you want to release it
175 to the public, you need to release the source code as well.
177 One exception is the Autotest framework, which is licenced under the
178 terms of the MIT licence (close to public domain) - you are free to
179 use it any way you wish.