uct/uct.c

   1 #include <assert.h>
   2 #include <math.h>
   3 #include <stdio.h>
   4 #include <stdlib.h>
   5 #include <string.h>
   6 #include <time.h>
   7
   8 #define DEBUG
   9
  10 #include "debug.h"
  11 #include "board.h"
  12 #include "gtp.h"
  13 #include "move.h"
  14 #include "mq.h"
  15 #include "playout.h"
  16 #include "playout/elo.h"
  17 #include "playout/moggy.h"
  18 #include "playout/light.h"
  19 #include "tactics.h"
  20 #include "timeinfo.h"
  21 #include "uct/dynkomi.h"
  22 #include "uct/internal.h"
  23 #include "uct/prior.h"
  24 #include "uct/search.h"
  25 #include "uct/slave.h"
  26 #include "uct/tree.h"
  27 #include "uct/uct.h"
  28 #include "uct/walk.h"
  29
  30 struct uct_policy *policy_ucb1_init(struct uct *u, char *arg);
  31 struct uct_policy *policy_ucb1amaf_init(struct uct *u, char *arg);
  32 static void uct_pondering_start(struct uct *u, struct board *b0, struct tree *t, enum stone color);
  33
  34 /* Maximal simulation length. */
  35 #define MC_GAMELEN      MAX_GAMELEN
  36
  37
  38 static void
  39 setup_state(struct uct *u, struct board *b, enum stone color)
  40 {
  41         u->t = tree_init(b, color, u->fast_alloc ? u->max_tree_size : 0,
  42                          u->local_tree_aging, u->stats_hbits);
  43         if (u->force_seed)
  44                 fast_srandom(u->force_seed);
  45         if (UDEBUGL(0))
  46                 fprintf(stderr, "Fresh board with random seed %lu\n", fast_getseed());
  47         //board_print(b, stderr);
  48         if (!u->no_book && b->moves == 0) {
  49                 assert(color == S_BLACK);
  50                 tree_load(u->t, b);
  51         }
  52 }
  53
  54 static void
  55 reset_state(struct uct *u)
  56 {
  57         assert(u->t);
  58         tree_done(u->t); u->t = NULL;
  59 }
  60
  61 static void
  62 setup_dynkomi(struct uct *u, struct board *b, enum stone to_play)
  63 {
  64         if (u->t->use_extra_komi && !u->pondering && u->dynkomi->permove)
  65                 u->t->extra_komi = u->dynkomi->permove(u->dynkomi, b, u->t);
  66 }
  67
  68 void
  69 uct_prepare_move(struct uct *u, struct board *b, enum stone color)
  70 {
  71         if (u->t) {
  72                 /* Verify that we have sane state. */
  73                 assert(b->es == u);
  74                 assert(u->t && b->moves);
  75                 if (color != stone_other(u->t->root_color)) {
  76                         fprintf(stderr, "Fatal: Non-alternating play detected %d %d\n",
  77                                 color, u->t->root_color);
  78                         exit(1);
  79                 }
  80                 uct_htable_reset(u->t);
  81
  82         } else {
  83                 /* We need fresh state. */
  84                 b->es = u;
  85                 setup_state(u, b, color);
  86         }
  87
  88         u->ownermap.playouts = 0;
  89         memset(u->ownermap.map, 0, board_size2(b) * sizeof(u->ownermap.map[0]));
  90         u->played_own = u->played_all = 0;
  91 }
  92
  93 static void
  94 dead_group_list(struct uct *u, struct board *b, struct move_queue *mq)
  95 {
  96         struct group_judgement gj;
  97         gj.thres = GJ_THRES;
  98         gj.gs = alloca(board_size2(b) * sizeof(gj.gs[0]));
  99         board_ownermap_judge_group(b, &u->ownermap, &gj);
 100         groups_of_status(b, &gj, GS_DEAD, mq);
 101 }
 102
 103 bool
 104 uct_pass_is_safe(struct uct *u, struct board *b, enum stone color, bool pass_all_alive)
 105 {
 106         if (u->ownermap.playouts < GJ_MINGAMES)
 107                 return false;
 108
 109         struct move_queue mq = { .moves = 0 };
 110         dead_group_list(u, b, &mq);
 111         if (pass_all_alive && mq.moves > 0)
 112                 return false; // We need to remove some dead groups first.
 113         return pass_is_safe(b, color, &mq);
 114 }
 115
 116 static char *
 117 uct_printhook_ownermap(struct board *board, coord_t c, char *s, char *end)
 118 {
 119         struct uct *u = board->es;
 120         assert(u);
 121         const char chr[] = ":XO,"; // dame, black, white, unclear
 122         const char chm[] = ":xo,";
 123         char ch = chr[board_ownermap_judge_point(&u->ownermap, c, GJ_THRES)];
 124         if (ch == ',') { // less precise estimate then?
 125                 ch = chm[board_ownermap_judge_point(&u->ownermap, c, 0.67)];
 126         }
 127         s += snprintf(s, end - s, "%c ", ch);
 128         return s;
 129 }
 130
 131 static char *
 132 uct_notify_play(struct engine *e, struct board *b, struct move *m)
 133 {
 134         struct uct *u = e->data;
 135         if (!u->t) {
 136                 /* No state, create one - this is probably game beginning
 137                  * and we need to load the opening book right now. */
 138                 uct_prepare_move(u, b, m->color);
 139                 assert(u->t);
 140         }
 141
 142         /* Stop pondering, required by tree_promote_at() */
 143         uct_pondering_stop(u);
 144         if (UDEBUGL(2) && u->slave)
 145                 tree_dump(u->t, u->dumpthres);
 146
 147         if (is_resign(m->coord)) {
 148                 /* Reset state. */
 149                 reset_state(u);
 150                 return NULL;
 151         }
 152
 153         /* Promote node of the appropriate move to the tree root. */
 154         assert(u->t->root);
 155         if (!tree_promote_at(u->t, b, m->coord)) {
 156                 if (UDEBUGL(0))
 157                         fprintf(stderr, "Warning: Cannot promote move node! Several play commands in row?\n");
 158                 reset_state(u);
 159                 return NULL;
 160         }
 161
 162         /* If we are a slave in a distributed engine, start pondering once
 163          * we know which move we actually played. See uct_genmove() about
 164          * the check for pass. */
 165         if (u->pondering_opt && u->slave && m->color == u->my_color && !is_pass(m->coord))
 166                 uct_pondering_start(u, b, u->t, stone_other(m->color));
 167
 168         return NULL;
 169 }
 170
 171 static char *
 172 uct_chat(struct engine *e, struct board *b, char *cmd)
 173 {
 174         struct uct *u = e->data;
 175         static char reply[1024];
 176
 177         cmd += strspn(cmd, " \n\t");
 178         if (!strncasecmp(cmd, "winrate", 7)) {
 179                 if (!u->t)
 180                         return "no game context (yet?)";
 181                 enum stone color = u->t->root_color;
 182                 struct tree_node *n = u->t->root;
 183                 snprintf(reply, 1024, "In %d playouts at %d threads, %s %s can win with %.2f%% probability",
 184                          n->u.playouts, u->threads, stone2str(color), coord2sstr(n->coord, b),
 185                          tree_node_get_value(u->t, -1, n->u.value) * 100);
 186                 if (u->t->use_extra_komi && abs(u->t->extra_komi) >= 0.5) {
 187                         sprintf(reply + strlen(reply), ", while self-imposing extra komi %.1f",
 188                                 u->t->extra_komi);
 189                 }
 190                 strcat(reply, ".");
 191                 return reply;
 192         }
 193         return NULL;
 194 }
 195
 196 static void
 197 uct_dead_group_list(struct engine *e, struct board *b, struct move_queue *mq)
 198 {
 199         struct uct *u = e->data;
 200
 201         /* This means the game is probably over, no use pondering on. */
 202         uct_pondering_stop(u);
 203
 204         if (u->pass_all_alive)
 205                 return; // no dead groups
 206
 207         bool mock_state = false;
 208
 209         if (!u->t) {
 210                 /* No state, but we cannot just back out - we might
 211                  * have passed earlier, only assuming some stones are
 212                  * dead, and then re-connected, only to lose counting
 213                  * when all stones are assumed alive. */
 214                 uct_prepare_move(u, b, S_BLACK); assert(u->t);
 215                 mock_state = true;
 216         }
 217         /* Make sure the ownermap is well-seeded. */
 218         while (u->ownermap.playouts < GJ_MINGAMES)
 219                 uct_playout(u, b, S_BLACK, u->t);
 220         /* Show the ownermap: */
 221         if (DEBUGL(2))
 222                 board_print_custom(b, stderr, uct_printhook_ownermap);
 223
 224         dead_group_list(u, b, mq);
 225
 226         if (mock_state) {
 227                 /* Clean up the mock state in case we will receive
 228                  * a genmove; we could get a non-alternating-move
 229                  * error from uct_prepare_move() in that case otherwise. */
 230                 reset_state(u);
 231         }
 232 }
 233
 234 static void
 235 playout_policy_done(struct playout_policy *p)
 236 {
 237         if (p->done) p->done(p);
 238         if (p->data) free(p->data);
 239         free(p);
 240 }
 241
 242 static void
 243 uct_done(struct engine *e)
 244 {
 245         /* This is called on engine reset, especially when clear_board
 246          * is received and new game should begin. */
 247         struct uct *u = e->data;
 248         uct_pondering_stop(u);
 249         if (u->t) reset_state(u);
 250         free(u->ownermap.map);
 251
 252         free(u->policy);
 253         free(u->random_policy);
 254         playout_policy_done(u->playout);
 255         uct_prior_done(u->prior);
 256 }
 257
 258
 259
 260 /* Run time-limited MCTS search on foreground. */
 261 static int
 262 uct_search(struct uct *u, struct board *b, struct time_info *ti, enum stone color, struct tree *t)
 263 {
 264         struct uct_search_state s;
 265         uct_search_start(u, b, color, t, ti, &s);
 266         if (UDEBUGL(2) && s.base_playouts > 0)
 267                 fprintf(stderr, "<pre-simulated %d games>\n", s.base_playouts);
 268
 269         /* The search tree is ctx->t. This is currently == . It is important
 270          * to reference ctx->t directly since the
 271          * thread manager will swap the tree pointer asynchronously. */
 272
 273         /* Now, just periodically poll the search tree. */
 274         while (1) {
 275                 time_sleep(TREE_BUSYWAIT_INTERVAL);
 276                 /* TREE_BUSYWAIT_INTERVAL should never be less than desired time, or the
 277                  * time control is broken. But if it happens to be less, we still search
 278                  * at least 100ms otherwise the move is completely random. */
 279
 280                 int i = uct_search_games(&s);
 281                 /* Print notifications etc. */
 282                 uct_search_progress(u, b, color, t, ti, &s, i);
 283                 /* Check if we should stop the search. */
 284                 if (uct_search_check_stop(u, b, color, t, ti, &s, i))
 285                         break;
 286         }
 287
 288         struct uct_thread_ctx *ctx = uct_search_stop();
 289         if (UDEBUGL(2)) tree_dump(t, u->dumpthres);
 290         if (UDEBUGL(2))
 291                 fprintf(stderr, "(avg score %f/%d value %f/%d)\n",
 292                         u->dynkomi->score.value, u->dynkomi->score.playouts,
 293                         u->dynkomi->value.value, u->dynkomi->value.playouts);
 294         if (UDEBUGL(0))
 295                 uct_progress_status(u, t, color, ctx->games);
 296
 297         u->played_own += ctx->games;
 298         return ctx->games;
 299 }
 300
 301 /* Start pondering background with @color to play. */
 302 static void
 303 uct_pondering_start(struct uct *u, struct board *b0, struct tree *t, enum stone color)
 304 {
 305         if (UDEBUGL(1))
 306                 fprintf(stderr, "Starting to ponder with color %s\n", stone2str(stone_other(color)));
 307         u->pondering = true;
 308
 309         /* We need a local board copy to ponder upon. */
 310         struct board *b = malloc2(sizeof(*b)); board_copy(b, b0);
 311
 312         /* *b0 did not have the genmove'd move played yet. */
 313         struct move m = { t->root->coord, t->root_color };
 314         int res = board_play(b, &m);
 315         assert(res >= 0);
 316         setup_dynkomi(u, b, stone_other(m.color));
 317
 318         /* Start MCTS manager thread "headless". */
 319         static struct uct_search_state s;
 320         uct_search_start(u, b, color, t, NULL, &s);
 321 }
 322
 323 /* uct_search_stop() frontend for the pondering (non-genmove) mode, and
 324  * to stop the background search for a slave in the distributed engine. */
 325 void
 326 uct_pondering_stop(struct uct *u)
 327 {
 328         if (!thread_manager_running)
 329                 return;
 330
 331         /* Stop the thread manager. */
 332         struct uct_thread_ctx *ctx = uct_search_stop();
 333         if (UDEBUGL(1)) {
 334                 if (u->pondering) fprintf(stderr, "(pondering) ");
 335                 uct_progress_status(u, ctx->t, ctx->color, ctx->games);
 336         }
 337         if (u->pondering) {
 338                 free(ctx->b);
 339                 u->pondering = false;
 340         }
 341 }
 342
 343
 344 void
 345 uct_genmove_setup(struct uct *u, struct board *b, enum stone color)
 346 {
 347         if (b->superko_violation) {
 348                 fprintf(stderr, "!!! WARNING: SUPERKO VIOLATION OCCURED BEFORE THIS MOVE\n");
 349                 fprintf(stderr, "Maybe you play with situational instead of positional superko?\n");
 350                 fprintf(stderr, "I'm going to ignore the violation, but note that I may miss\n");
 351                 fprintf(stderr, "some moves valid under this ruleset because of this.\n");
 352                 b->superko_violation = false;
 353         }
 354
 355         uct_prepare_move(u, b, color);
 356
 357         assert(u->t);
 358         u->my_color = color;
 359
 360         /* How to decide whether to use dynkomi in this game? Since we use
 361          * pondering, it's not simple "who-to-play" matter. Decide based on
 362          * the last genmove issued. */
 363         u->t->use_extra_komi = !!(u->dynkomi_mask & color);
 364         setup_dynkomi(u, b, color);
 365
 366         if (b->rules == RULES_JAPANESE)
 367                 u->territory_scoring = true;
 368
 369         /* Make pessimistic assumption about komi for Japanese rules to
 370          * avoid losing by 0.5 when winning by 0.5 with Chinese rules.
 371          * The rules usually give the same winner if the integer part of komi
 372          * is odd so we adjust the komi only if it is even (for a board of
 373          * odd size). We are not trying  to get an exact evaluation for rare
 374          * cases of seki. For details see http://home.snafu.de/jasiek/parity.html */
 375         if (u->territory_scoring && (((int)floor(b->komi) + board_size(b)) & 1)) {
 376                 b->komi += (color == S_BLACK ? 1.0 : -1.0);
 377                 if (UDEBUGL(0))
 378                         fprintf(stderr, "Setting komi to %.1f assuming Japanese rules\n",
 379                                 b->komi);
 380         }
 381 }
 382
 383 static coord_t *
 384 uct_genmove(struct engine *e, struct board *b, struct time_info *ti, enum stone color, bool pass_all_alive)
 385 {
 386         double start_time = time_now();
 387         struct uct *u = e->data;
 388         uct_pondering_stop(u);
 389         uct_genmove_setup(u, b, color);
 390
 391         /* Start the Monte Carlo Tree Search! */
 392         int base_playouts = u->t->root->u.playouts;
 393         int played_games = uct_search(u, b, ti, color, u->t);
 394
 395         coord_t best_coord;
 396         struct tree_node *best;
 397         best = uct_search_result(u, b, color, pass_all_alive, played_games, base_playouts, &best_coord);
 398
 399         if (UDEBUGL(2)) {
 400                 double time = time_now() - start_time + 0.000001; /* avoid divide by zero */
 401                 fprintf(stderr, "genmove in %0.2fs (%d games/s, %d games/s/thread)\n",
 402                         time, (int)(played_games/time), (int)(played_games/time/u->threads));
 403         }
 404
 405         if (!best) {
 406                 /* Pass or resign. */
 407                 reset_state(u);
 408                 return coord_copy(best_coord);
 409         }
 410         tree_promote_node(u->t, &best);
 411
 412         /* After a pass, pondering is harmful for two reasons:
 413          * (i) We might keep pondering even when the game is over.
 414          * Of course this is the case for opponent resign as well.
 415          * (ii) More importantly, the ownermap will get skewed since
 416          * the UCT will start cutting off any playouts. */
 417         if (u->pondering_opt && !is_pass(best->coord)) {
 418                 uct_pondering_start(u, b, u->t, stone_other(color));
 419         }
 420         return coord_copy(best_coord);
 421 }
 422
 423
 424 bool
 425 uct_genbook(struct engine *e, struct board *b, struct time_info *ti, enum stone color)
 426 {
 427         struct uct *u = e->data;
 428         if (!u->t) uct_prepare_move(u, b, color);
 429         assert(u->t);
 430
 431         if (ti->dim == TD_GAMES) {
 432                 /* Don't count in games that already went into the book. */
 433                 ti->len.games += u->t->root->u.playouts;
 434         }
 435         uct_search(u, b, ti, color, u->t);
 436
 437         assert(ti->dim == TD_GAMES);
 438         tree_save(u->t, b, ti->len.games / 100);
 439
 440         return true;
 441 }
 442
 443 void
 444 uct_dumpbook(struct engine *e, struct board *b, enum stone color)
 445 {
 446         struct uct *u = e->data;
 447         struct tree *t = tree_init(b, color, u->fast_alloc ? u->max_tree_size : 0, u->local_tree_aging, 0);
 448         tree_load(t, b);
 449         tree_dump(t, 0);
 450         tree_done(t);
 451 }
 452
 453
 454 struct uct *
 455 uct_state_init(char *arg, struct board *b)
 456 {
 457         struct uct *u = calloc2(1, sizeof(struct uct));
 458         bool using_elo = false;
 459
 460         u->debug_level = debug_level;
 461         u->gamelen = MC_GAMELEN;
 462         u->mercymin = 0;
 463         u->expand_p = 2;
 464         u->dumpthres = 1000;
 465         u->playout_amaf = true;
 466         u->playout_amaf_nakade = false;
 467         u->amaf_prior = false;
 468         u->max_tree_size = 3072ULL * 1048576;
 469
 470         u->dynkomi_mask = S_BLACK;
 471
 472         u->threads = 1;
 473         u->thread_model = TM_TREEVL;
 474         u->virtual_loss = true;
 475
 476         u->fuseki_end = 20; // max time at 361*20% = 72 moves (our 36th move, still 99 to play)
 477         u->yose_start = 40; // (100-40-25)*361/100/2 = 63 moves still to play by us then
 478         u->bestr_ratio = 0.02;
 479         // 2.5 is clearly too much, but seems to compensate well for overly stern time allocations.
 480         // TODO: Further tuning and experiments with better time allocation schemes.
 481         u->best2_ratio = 2.5;
 482
 483         u->val_scale = 0.04; u->val_points = 40;
 484         /* TODO: Adjust this by number of simulations - it's more important how
 485          * many times per move we do the adjustment. */
 486         u->dynkomi_interval = 500;
 487
 488         u->tenuki_d = 4;
 489         u->local_tree_aging = 2;
 490
 491         if (arg) {
 492                 char *optspec, *next = arg;
 493                 while (*next) {
 494                         optspec = next;
 495                         next += strcspn(next, ",");
 496                         if (*next) { *next++ = 0; } else { *next = 0; }
 497
 498                         char *optname = optspec;
 499                         char *optval = strchr(optspec, '=');
 500                         if (optval) *optval++ = 0;
 501
 502                         if (!strcasecmp(optname, "debug")) {
 503                                 if (optval)
 504                                         u->debug_level = atoi(optval);
 505                                 else
 506                                         u->debug_level++;
 507                         } else if (!strcasecmp(optname, "mercy") && optval) {
 508                                 /* Minimal difference of black/white captures
 509                                  * to stop playout - "Mercy Rule". Speeds up
 510                                  * hopeless playouts at the expense of some
 511                                  * accuracy. */
 512                                 u->mercymin = atoi(optval);
 513                         } else if (!strcasecmp(optname, "gamelen") && optval) {
 514                                 u->gamelen = atoi(optval);
 515                         } else if (!strcasecmp(optname, "expand_p") && optval) {
 516                                 u->expand_p = atoi(optval);
 517                         } else if (!strcasecmp(optname, "dumpthres") && optval) {
 518                                 u->dumpthres = atoi(optval);
 519                         } else if (!strcasecmp(optname, "best2_ratio") && optval) {
 520                                 /* If set, prolong simulating while
 521                                  * first_best/second_best playouts ratio
 522                                  * is less than best2_ratio. */
 523                                 u->best2_ratio = atof(optval);
 524                         } else if (!strcasecmp(optname, "bestr_ratio") && optval) {
 525                                 /* If set, prolong simulating while
 526                                  * best,best_best_child values delta
 527                                  * is more than bestr_ratio. */
 528                                 u->bestr_ratio = atof(optval);
 529                         } else if (!strcasecmp(optname, "playout_amaf")) {
 530                                 /* Whether to include random playout moves in
 531                                  * AMAF as well. (Otherwise, only tree moves
 532                                  * are included in AMAF. Of course makes sense
 533                                  * only in connection with an AMAF policy.) */
 534                                 /* with-without: 55.5% (+-4.1) */
 535                                 if (optval && *optval == '0')
 536                                         u->playout_amaf = false;
 537                                 else
 538                                         u->playout_amaf = true;
 539                         } else if (!strcasecmp(optname, "playout_amaf_nakade")) {
 540                                 /* Whether to include nakade moves from playouts
 541                                  * in the AMAF statistics; this tends to nullify
 542                                  * the playout_amaf effect by adding too much
 543                                  * noise. */
 544                                 if (optval && *optval == '0')
 545                                         u->playout_amaf_nakade = false;
 546                                 else
 547                                         u->playout_amaf_nakade = true;
 548                         } else if (!strcasecmp(optname, "playout_amaf_cutoff") && optval) {
 549                                 /* Keep only first N% of playout stage AMAF
 550                                  * information. */
 551                                 u->playout_amaf_cutoff = atoi(optval);
 552                         } else if ((!strcasecmp(optname, "policy") || !strcasecmp(optname, "random_policy")) && optval) {
 553                                 char *policyarg = strchr(optval, ':');
 554                                 struct uct_policy **p = !strcasecmp(optname, "policy") ? &u->policy : &u->random_policy;
 555                                 if (policyarg)
 556                                         *policyarg++ = 0;
 557                                 if (!strcasecmp(optval, "ucb1")) {
 558                                         *p = policy_ucb1_init(u, policyarg);
 559                                 } else if (!strcasecmp(optval, "ucb1amaf")) {
 560                                         *p = policy_ucb1amaf_init(u, policyarg);
 561                                 } else {
 562                                         fprintf(stderr, "UCT: Invalid tree policy %s\n", optval);
 563                                         exit(1);
 564                                 }
 565                         } else if (!strcasecmp(optname, "playout") && optval) {
 566                                 char *playoutarg = strchr(optval, ':');
 567                                 if (playoutarg)
 568                                         *playoutarg++ = 0;
 569                                 if (!strcasecmp(optval, "moggy")) {
 570                                         u->playout = playout_moggy_init(playoutarg, b);
 571                                 } else if (!strcasecmp(optval, "light")) {
 572                                         u->playout = playout_light_init(playoutarg, b);
 573                                 } else if (!strcasecmp(optval, "elo")) {
 574                                         u->playout = playout_elo_init(playoutarg, b);
 575                                         using_elo = true;
 576                                 } else {
 577                                         fprintf(stderr, "UCT: Invalid playout policy %s\n", optval);
 578                                         exit(1);
 579                                 }
 580                         } else if (!strcasecmp(optname, "prior") && optval) {
 581                                 u->prior = uct_prior_init(optval, b);
 582                         } else if (!strcasecmp(optname, "amaf_prior") && optval) {
 583                                 u->amaf_prior = atoi(optval);
 584                         } else if (!strcasecmp(optname, "threads") && optval) {
 585                                 /* By default, Pachi will run with only single
 586                                  * tree search thread! */
 587                                 u->threads = atoi(optval);
 588                         } else if (!strcasecmp(optname, "thread_model") && optval) {
 589                                 if (!strcasecmp(optval, "tree")) {
 590                                         /* Tree parallelization - all threads
 591                                          * grind on the same tree. */
 592                                         u->thread_model = TM_TREE;
 593                                         u->virtual_loss = false;
 594                                 } else if (!strcasecmp(optval, "treevl")) {
 595                                         /* Tree parallelization, but also
 596                                          * with virtual losses - this discou-
 597                                          * rages most threads choosing the
 598                                          * same tree branches to read. */
 599                                         u->thread_model = TM_TREEVL;
 600                                         u->virtual_loss = true;
 601                                 } else {
 602                                         fprintf(stderr, "UCT: Invalid thread model %s\n", optval);
 603                                         exit(1);
 604                                 }
 605                         } else if (!strcasecmp(optname, "pondering")) {
 606                                 /* Keep searching even during opponent's turn. */
 607                                 u->pondering_opt = !optval || atoi(optval);
 608                         } else if (!strcasecmp(optname, "fuseki_end") && optval) {
 609                                 /* At the very beginning it's not worth thinking
 610                                  * too long because the playout evaluations are
 611                                  * very noisy. So gradually increase the thinking
 612                                  * time up to maximum when fuseki_end percent
 613                                  * of the board has been played.
 614                                  * This only applies if we are not in byoyomi. */
 615                                 u->fuseki_end = atoi(optval);
 616                         } else if (!strcasecmp(optname, "yose_start") && optval) {
 617                                 /* When yose_start percent of the board has been
 618                                  * played, or if we are in byoyomi, stop spending
 619                                  * more time and spread the remaining time
 620                                  * uniformly.
 621                                  * Between fuseki_end and yose_start, we spend
 622                                  * a constant proportion of the remaining time
 623                                  * on each move. (yose_start should actually
 624                                  * be much earlier than when real yose start,
 625                                  * but "yose" is a good short name to convey
 626                                  * the idea.) */
 627                                 u->yose_start = atoi(optval);
 628                         } else if (!strcasecmp(optname, "force_seed") && optval) {
 629                                 u->force_seed = atoi(optval);
 630                         } else if (!strcasecmp(optname, "no_book")) {
 631                                 u->no_book = true;
 632                         } else if (!strcasecmp(optname, "dynkomi") && optval) {
 633                                 /* Dynamic komi approach; there are multiple
 634                                  * ways to adjust komi dynamically throughout
 635                                  * play. We currently support two: */
 636                                 char *dynkomiarg = strchr(optval, ':');
 637                                 if (dynkomiarg)
 638                                         *dynkomiarg++ = 0;
 639                                 if (!strcasecmp(optval, "none")) {
 640                                         u->dynkomi = uct_dynkomi_init_none(u, dynkomiarg, b);
 641                                 } else if (!strcasecmp(optval, "linear")) {
 642                                         u->dynkomi = uct_dynkomi_init_linear(u, dynkomiarg, b);
 643                                 } else if (!strcasecmp(optval, "adaptive")) {
 644                                         u->dynkomi = uct_dynkomi_init_adaptive(u, dynkomiarg, b);
 645                                 } else {
 646                                         fprintf(stderr, "UCT: Invalid dynkomi mode %s\n", optval);
 647                                         exit(1);
 648                                 }
 649                         } else if (!strcasecmp(optname, "dynkomi_mask") && optval) {
 650                                 /* Bitmask of colors the player must be
 651                                  * for dynkomi be applied; you may want
 652                                  * to use dynkomi_mask=3 to allow dynkomi
 653                                  * even in games where Pachi is white. */
 654                                 u->dynkomi_mask = atoi(optval);
 655                         } else if (!strcasecmp(optname, "dynkomi_interval") && optval) {
 656                                 /* If non-zero, re-adjust dynamic komi
 657                                  * throughout a single genmove reading,
 658                                  * roughly every N simulations. */
 659                                 /* XXX: Does not work with tree
 660                                  * parallelization. */
 661                                 u->dynkomi_interval = atoi(optval);
 662                         } else if (!strcasecmp(optname, "val_scale") && optval) {
 663                                 /* How much of the game result value should be
 664                                  * influenced by win size. Zero means it isn't. */
 665                                 u->val_scale = atof(optval);
 666                         } else if (!strcasecmp(optname, "val_points") && optval) {
 667                                 /* Maximum size of win to be scaled into game
 668                                  * result value. Zero means boardsize^2. */
 669                                 u->val_points = atoi(optval) * 2; // result values are doubled
 670                         } else if (!strcasecmp(optname, "val_extra")) {
 671                                 /* If false, the score coefficient will be simply
 672                                  * added to the value, instead of scaling the result
 673                                  * coefficient because of it. */
 674                                 u->val_extra = !optval || atoi(optval);
 675                         } else if (!strcasecmp(optname, "local_tree") && optval) {
 676                                 /* Whether to bias exploration by local tree values
 677                                  * (must be supported by the used policy).
 678                                  * 0: Don't.
 679                                  * 1: Do, value = result.
 680                                  * Try to temper the result:
 681                                  * 2: Do, value = 0.5+(result-expected)/2.
 682                                  * 3: Do, value = 0.5+bzz((result-expected)^2).
 683                                  * 4: Do, value = 0.5+sqrt(result-expected)/2. */
 684                                 u->local_tree = atoi(optval);
 685                         } else if (!strcasecmp(optname, "tenuki_d") && optval) {
 686                                 /* Tenuki distance at which to break the local tree. */
 687                                 u->tenuki_d = atoi(optval);
 688                                 if (u->tenuki_d > TREE_NODE_D_MAX + 1) {
 689                                         fprintf(stderr, "uct: tenuki_d must not be larger than TREE_NODE_D_MAX+1 %d\n", TREE_NODE_D_MAX + 1);
 690                                         exit(1);
 691                                 }
 692                         } else if (!strcasecmp(optname, "local_tree_aging") && optval) {
 693                                 /* How much to reduce local tree values between moves. */
 694                                 u->local_tree_aging = atof(optval);
 695                         } else if (!strcasecmp(optname, "local_tree_allseq")) {
 696                                 /* By default, only complete sequences are stored
 697                                  * in the local tree. If this is on, also
 698                                  * subsequences starting at each move are stored. */
 699                                 u->local_tree_allseq = !optval || atoi(optval);
 700                         } else if (!strcasecmp(optname, "local_tree_playout")) {
 701                                 /* Whether to adjust ELO playout probability
 702                                  * distributions according to matched localtree
 703                                  * information. */
 704                                 u->local_tree_playout = !optval || atoi(optval);
 705                         } else if (!strcasecmp(optname, "local_tree_pseqroot")) {
 706                                 /* By default, when we have no sequence move
 707                                  * to suggest in-playout, we give up. If this
 708                                  * is on, we make probability distribution from
 709                                  * sequences first moves instead. */
 710                                 u->local_tree_pseqroot = !optval || atoi(optval);
 711                         } else if (!strcasecmp(optname, "pass_all_alive")) {
 712                                 /* Whether to consider passing only after all
 713                                  * dead groups were removed from the board;
 714                                  * this is like all genmoves are in fact
 715                                  * kgs-genmove_cleanup. */
 716                                 u->pass_all_alive = !optval || atoi(optval);
 717                         } else if (!strcasecmp(optname, "territory_scoring")) {
 718                                 /* Use territory scoring (default is area scoring).
 719                                  * An explicit kgs-rules command overrides this. */
 720                                 u->territory_scoring = !optval || atoi(optval);
 721                         } else if (!strcasecmp(optname, "random_policy_chance") && optval) {
 722                                 /* If specified (N), with probability 1/N, random_policy policy
 723                                  * descend is used instead of main policy descend; useful
 724                                  * if specified policy (e.g. UCB1AMAF) can make unduly biased
 725                                  * choices sometimes, you can fall back to e.g.
 726                                  * random_policy=UCB1. */
 727                                 u->random_policy_chance = atoi(optval);
 728                         } else if (!strcasecmp(optname, "max_tree_size") && optval) {
 729                                 /* Maximum amount of memory [MiB] consumed by the move tree.
 730                                  * For fast_alloc it includes the temp tree used for pruning.
 731                                  * Default is 3072 (3 GiB). */
 732                                 u->max_tree_size = atol(optval) * 1048576;
 733                         } else if (!strcasecmp(optname, "fast_alloc")) {
 734                                 u->fast_alloc = !optval || atoi(optval);
 735                         } else if (!strcasecmp(optname, "slave")) {
 736                                 /* Act as slave for the distributed engine. */
 737                                 u->slave = !optval || atoi(optval);
 738                         } else if (!strcasecmp(optname, "shared_nodes") && optval) {
 739                                 /* Share at most shared_nodes between master and slave at each genmoves.
 740                                  * Must use the same value in master and slaves. */
 741                                 u->shared_nodes = atoi(optval);
 742                         } else if (!strcasecmp(optname, "shared_levels") && optval) {
 743                                 /* Share only nodes of level <= shared_levels. */
 744                                 u->shared_levels = atoi(optval);
 745                         } else if (!strcasecmp(optname, "stats_hbits") && optval) {
 746                                 /* Set hash table size to 2^stats_hbits for the shared stats. */
 747                                 u->stats_hbits = atoi(optval);
 748                         } else if (!strcasecmp(optname, "banner") && optval) {
 749                                 /* Additional banner string. This must come as the
 750                                  * last engine parameter. */
 751                                 if (*next) *--next = ',';
 752                                 u->banner = strdup(optval);
 753                                 break;
 754                         } else {
 755                                 fprintf(stderr, "uct: Invalid engine argument %s or missing value\n", optname);
 756                                 exit(1);
 757                         }
 758                 }
 759         }
 760
 761         u->resign_ratio = 0.2; /* Resign when most games are lost. */
 762         u->loss_threshold = 0.85; /* Stop reading if after at least 2000 playouts this is best value. */
 763         if (!u->policy)
 764                 u->policy = policy_ucb1amaf_init(u, NULL);
 765
 766         if (!!u->random_policy_chance ^ !!u->random_policy) {
 767                 fprintf(stderr, "uct: Only one of random_policy and random_policy_chance is set\n");
 768                 exit(1);
 769         }
 770
 771         if (!u->local_tree) {
 772                 /* No ltree aging. */
 773                 u->local_tree_aging = 1.0f;
 774         }
 775         if (!using_elo)
 776                 u->local_tree_playout = false;
 777
 778         if (u->fast_alloc)
 779                 u->max_tree_size = (100ULL * u->max_tree_size) / (100 + MIN_FREE_MEM_PERCENT);
 780
 781         if (!u->prior)
 782                 u->prior = uct_prior_init(NULL, b);
 783
 784         if (!u->playout)
 785                 u->playout = playout_moggy_init(NULL, b);
 786         u->playout->debug_level = u->debug_level;
 787
 788         u->ownermap.map = malloc2(board_size2(b) * sizeof(u->ownermap.map[0]));
 789
 790         if (u->slave) {
 791                 if (!u->stats_hbits) u->stats_hbits = DEFAULT_STATS_HBITS;
 792                 if (!u->shared_nodes) u->shared_nodes = DEFAULT_SHARED_NODES;
 793                 if (!u->shared_levels) u->shared_levels = 1;
 794                 assert(u->shared_levels * board_bits2(b) <= 8 * (int)sizeof(path_t));
 795         }
 796
 797         if (!u->dynkomi)
 798                 u->dynkomi = uct_dynkomi_init_linear(u, NULL, b);
 799
 800         /* Some things remain uninitialized for now - the opening book
 801          * is not loaded and the tree not set up. */
 802         /* This will be initialized in setup_state() at the first move
 803          * received/requested. This is because right now we are not aware
 804          * about any komi or handicap setup and such. */
 805
 806         return u;
 807 }
 808
 809 struct engine *
 810 engine_uct_init(char *arg, struct board *b)
 811 {
 812         struct uct *u = uct_state_init(arg, b);
 813         struct engine *e = calloc2(1, sizeof(struct engine));
 814         e->name = "UCT Engine";
 815         e->printhook = uct_printhook_ownermap;
 816         e->notify_play = uct_notify_play;
 817         e->chat = uct_chat;
 818         e->genmove = uct_genmove;
 819         e->genmoves = uct_genmoves;
 820         e->dead_group_list = uct_dead_group_list;
 821         e->done = uct_done;
 822         e->data = u;
 823         if (u->slave)
 824                 e->notify = uct_notify;
 825
 826         const char banner[] = "I'm playing UCT. When I'm losing, I will resign, "
 827                 "if I think I win, I play until you pass. "
 828                 "Anyone can send me 'winrate' in private chat to get my assessment of the position.";
 829         if (!u->banner) u->banner = "";
 830         e->comment = malloc2(sizeof(banner) + strlen(u->banner) + 1);
 831         sprintf(e->comment, "%s %s", banner, u->banner);
 832
 833         return e;
 834 }