Closed Incidents

Nr	Date	Description	Detected in	Fixed in	Game record	Move
1	990512	Assertion failure in double_attack_helper.	2.3.1	2.3.2	double_attack.sgf.gz	232
2	990512	Reading error. The attacker mistakenly thinks it can capture four stones.	2.3.2	2.3.16	gnugo-2.1.23gf-brad-199905121306.sgf.gz	212
3	990512	Bad shape. Pattern CC91 fills a halfeye instead of making a full eye.	2.3.2	2.3.91	seiwa-gnugo-2.1.23gf-199905121651.sgf.gz	110
4	990512	Semeai mistake. Playing into self atari. Also suboptimal moves earlier in the semeai.	2.3.2	2.3.91	greycat-gnugo-2.1.23gf-199905111941.sgf.gz	228
5	990512	Semeai problem. At the end of the game both groups in a semeai are considered dead, with the effect that the endgame is not played to finish. Also two other groups misclassified as dead.	2.3.2	2.3.91	fire-gnugo-2.1.23gf-199905121832.sgf.gz	270
6	990513	Reading error. The defender comes up with a bogus move. Although it's a double atari on the opponent, the stone itself is in atari.	2.3.3	2.3.16	gnugo-2.3.3-baduki.sgf.gz	219
7	990513	Reading error. Similar to the previous one except there is no double atari involved.	2.3.3	2.3.17	gnugo-2.3.3-baduki-2.sgf.gz	241
8	990513	Failure to connect an important ko. Not sure how best to fix this, but in early positions like this it's almost always correct to connect the ko.	2.3.3	2.3.52	gnugo-2.3.3-baduki-3.sgf.gz	15
9	990513	Reading error. Another self-atari from the defender.	2.3.3	2.3.17	gnugo-2.3.3-baduki-3.sgf.gz	231
10	990513	Missed semeai. From move 259 and forward very careless moves are played, filling liberties of a group that should be considered to be in a semeai. Not sure why it isn't classified as dead.	2.3.3	2.3.52	gnugo-2.3.3-baduki-3.sgf.gz	259
11	990513	Life and death. Several examples of need for better eye evaluation.	2.3.3	2.3.91	gnugo-2.3.3-baduki-3.sgf.gz	-
12	990514	Reading error. The attacker overrules a nakade pattern but only helps the opponent to live.	2.3.4	2.3.17	gnugo-2.3.4-baduki.sgf.gz	163
13	990514	Reading error. The defender suggests a bogus move. In the moves 132-152 there is also a hugely complex positions with 10 or so strings with very few liberties. This should be a very good game to debug the tactical reading.	2.3.4	2.3.18	gnugo-2.3.4-kinchan-199905141616.sgf.gz	132
14	990514	Reading error. The attacker thinks it can capture some stones but gives up after the opponent responds. There is also a lot of reading to do from move 208 and forward.	2.3.4	2.3.91	gnugo-2.3.4-kinchan-199905141616.sgf.gz	222
15	990516	Reading error. The attacker mistakenly thinks it can capture a stone. Although the move happens to be reasonable the premise isn't.	2.3.7	2.3.91	gnugo-2.3.7-baduki.sgf.gz	95
16	990516	Reading error. The defender plays a self atari that doesn't work at all.	2.3.7	2.3.52	gnugo-2.3.7-baduki.sgf.gz	161
17	990516	Semeai error. Failure to understand that a semeai was going on, resulting in the loss of the major part of the dragon involved.	2.3.7	2.7.12	gnugo-2.3.7-baduki.sgf.gz	275
18	990519	Shape blunder. The backfill pattern CD47 suggests with value 100 a move only accomplishing to fill an own eye. Might be caused by a reading error.	2.3.11	2.3.52	gnugo-2.3.9gf-baduki.sgf.gz	82
19	990519	Reading error. The attacker attacks at the wrong point, losing a semeai. There are also some strange moves suggested by the semeai player, the attacker, and the defender, both just before and after this move.	2.3.11	2.3.91	gnugo-2.3.9gf-baduki.sgf.gz	166
20	990520	Shape blunder. CD47 again comes up with a nonsense move. (This move was not suggested by the modified 2.3.9 that was playing the game, so I guess this has something to do with the reading changes for 2.3.11.)	2.3.11	2.3.52	gnugo-2.3.9gf-baduki-2.sgf.gz	121
21	990520	Almost blunder. Here CD47 comes up with the right move but for the wrong reason, overruling a bad move from the semeai player, which in turn overrules the correct move suggested by the defender for the right reason.	2.3.11	2.3.52	gnugo-2.3.9gf-baduki-2.sgf.gz	137
22	990520	Ko mistake. GNU starts a ko fight that hardly is playable.	2.3.11	2.3.13	gnugo-2.3.9gf-baduki-2.sgf.gz	183
23	990520	Blunder. Playing a move to destroy an eye for a presumably critical group, disregarding the fact that one of the surrounding strings was both weak and big.	2.3.13	2.3.52	gnugo-2.3.4-arndt-199905201730.sgf.gz	40
24	990520	Reading error. The attacker suggests a move that doesn't work. 2.3.4 came up with a different bogus move when the game was played.	2.3.13	2.3.52	gnugo-2.3.4-arndt-199905201730.sgf.gz	54
25	990523	Reading error. The defender suggests an ineffective move. Not really a blunder but the move only forces the opponent to secure an eye.	2.3.15	2.3.52	gnugo-2.3.14gf-gnugo-2.3.15.sgf.gz	69
26	990524	Assertion failure in the reading code, called from hane_backfill_helper.	2.3.12	2.3.52	celery-gnugo-2.3.12-199905212207.sgf.gz	220
27	990525	Reading error. The attacker incorrectly thinks it can capture three stones.	2.3.19	2.3.23	gnugo-2.3.19-baduki.sgf.gz	69
28	990525	Reading error. A true blunder, attacking from a self-atari.	2.3.19	2.3.23	gnugo-2.3.19-baduki.sgf.gz	89
29	990527	Reading error. The defender mistakenly thinks it can save three stones by playing another stone into atari.	2.3.21	2.3.23	straw-gnugo-2.3.21-199905271558.sgf.gz	166
30	990528	Semeai error. No idea what the semeai player tries to accomplish. Doesn't look much like a semeai in the first place.	2.3.22	2.3.52	gnugo-2.3.22-wartari-199905281830.sgf.gz	88
31	990528	Semeai error. The semeai player strikes again. Equally strange this time.	2.3.22	2.3.52	gnugo-2.3.22-wartari-199905281830.sgf.gz	92
32	990528	Strategical error. We are too eager to capture irrelevant stones. This move (value 74) is 2 points gote, which is way too small when we still are in the middle game.	2.3.22	2.3.52	gnugo-2.3.22-wartari-199905281830.sgf.gz	146
33	990530	Blunder. After an atari on one of GNU's cutting strings, the status of the dragons in the area were upset enough that both the defender and the eye finder came up with a bad move with higher value than the correct defense (which was of course also suggested by the defender, but with too small value).	2.3.23	2.3.91	gnugo-2.3.22gf-baduki.sgf.gz	309
34	990531	Pointless attempt to kill. Incorrectly detected half eye.	2.3.24	2.3.91	gnugo-2.3.24gf-baduki.sgf.gz	233
35	990531	Semeai error. The semeai player suggests an ineffective move. But the problem is as much with the life and death classification, because it shouldn't even be a semeai.	2.3.24	2.3.91	gnugo-2.3.24gf-baduki.sgf.gz	249
36	990531	Life and death error. The eye finder panics to solidify a halfeye, overriding a connection to a tail of genus one.	2.3.24	2.5.31	gnugo-2.3.24gf-baduki.sgf.gz	267
37	990614	Reading error. GNU fails to play the endgame move B8, presumably because it thinks it would be unsafe.	2.3.33	2.3.52	gnugo-2.3.25-unusual-199906021327.sgf.gz	226
38	990616	Shape error. The detect_trap_helper induces a substandard defense move.	2.3.36	2.3.91	gnugo-2.3.36-baduki.sgf.gz	127
39	990616	Shape error. The urgent connect helper misfires.	2.3.36	2.3.52	gnugo-2.3.36-baduki.sgf.gz	129
40	990616	Reading error. The attacker mistakenly thinks it can capture some stones. This error also shows up in some of the preceding moves.	2.3.36	2.3.52	gnugo-2.3.36-baduki.sgf.gz	141
41	990616	Strategical error. Instead of calmly cutting off 14 black stones for a nice profit, GNU instead tries to run away with a heavy group, which is later lost.	2.3.36	2.7.12	gnugo-2.3.36-baduki.sgf.gz	195
42	990617	Reading error. We really should be able to find the R14 move.	2.3.37	2.7.208	marcel-gnugo-2.3.37-199906171720.sgf.gz	58
43	990617	Reading error. Actually the same position as #42. Anyway the suggested move doesn't work as well as stated by the attacker.	2.3.37	2.7.208	marcel-gnugo-2.3.37-199906171720.sgf.gz	88
44	990617	Reading mistake. Not really an error since the suggested move does work, but it's fairly obscure and leaves a lot of unnecessary aji.	2.3.37	2.3.91	marcel-gnugo-2.3.37-199906171720.sgf.gz	170
45	990617	Life and death error. This was the last chance to save a large group.	2.3.37	2.3.91	marcel-gnugo-2.3.37-199906171720.sgf.gz	284
46	990617	Semeai blunder. We certainly shouldn't play inside the seki!	2.3.37	2.3.52	marcel-gnugo-2.3.37-199906171720.sgf.gz	286
47	990618	Defense mistake. The defense point is moved one time too many, ending up at the wrong place.	2.3.37	2.3.91	gnugo-2.3.37-baduki.sgf.gz	203
48	990618	Wrong attack. The attacker overrules a move that would have captured the same stones too, but also have saved some other stones.	2.3.37	2.3.91	gnugo-2.3.37-baduki.sgf.gz	235
49	990620	Reading error. The capture doesn't work as three cutting stones are in immediate danger.	2.3.37	2.3.91	gnugo-2.3.37-keithg-199906202041.sgf.gz	242
50	990620	Reading problem? The move takes almost forever (i.e. several minutes) to find, although the game is on a 9x9 board.	2.3.37	2.3.91	gnugo-2.3.37-Darroch-199906180607.sgf.gz	46
51	990622	Reading error. The played move doesn't work without backfilling and certainly doesn't capture anything. A number of the preceding positions are probably misread as well.	2.3.40	2.3.91	gnugo-2.3.40gf-baduki.sgf.gz	181
52	990623	Reading blackout. In the sequence from move 46 to move 70, GNU miserably fails to defend an increasingly heavy group of stones. Is this a reading depth problem?	2.3.41	2.5.37	gnugo-2.3.41-whowasi-199906231810.sgf.gz	48
53	990627	Reading problem. GNU takes almost forever generating the move. Just makeworms() spends a horrible amount of time doing reading, sometimes up to some 60 call levels deep.	2.3.45	2.3.46	greenman-gnugo-2.3.44-199906270110.sgf.gz	250
54	990704	Reading error. Two adjacent strings at A3 and A4 are both(!) classified as attackable but not defendable.	2.3.49	2.3.91	gnugo-2.3.49-baduki.sgf.gz	251
55	990709	Reading error. A tricky ladder is misread. Moving out with M5 doesn't work.	1.121	2.5.43	RomanAP-gnugo-1.121-199904161627.sgf.gz	98
56	990711	Reading error. Three adjacent strings are all classified as attackable but not defendable. As a result the dragon code amalgamates them all to friendly stones on the outside. Where there in fact is a semeai, everything is classified as alive.	2.3.51	2.3.91	kakaka-gnugo-2.3.51-199907102216.sgf.gz	148
57	990717	Reading error. The defender comes up with a sacrifice move at F16, supposedly defending B19, but in reality accomplishing nothing. The reason is probably fairly obscure since it took 17572 variations to find the move.	2.3.61gf	2.3.75	gnugo-2.3.61gf-baduki.sgf.gz	223
58	990725	Severe reading problem. find_defense() needs 6722903 variations at depth 14 to evaluate whether the string P7 can be defended. Needless to say, this takes a very long time.	2.3.65	2.3.91	Onzrock-gnugo-2.3.65-199907250206.sgf.gz	168
60	990729	Reading error. GNU fails to find the killing move at S13. Version 2.3.68 does not.	2.3.70	2.3.75	gnugo-2.3.70-baduki.sgf.gz	279
61	990816	Reading error. No attack is found on the O4 stones.	2.3.83	2.5.37	exoss-gnugo-2.3.83-199908161154.sgf.gz	268
62	990817	Life and death error. The eye finder comes up with a very strange move at L4.	2.3.83gf	2.3.91	gnugo-2.3.83gf-baduki.sgf.gz	179
63	990817	Reading error. find_defense() never considers the only effective move at T2, securing two eyes.	2.3.83gf	2.3.91	gnugo-2.3.83gf-baduki.sgf.gz	253
64	990823	Reading error. attacker() thinks it can capture the three white F7 stones by moving out the F9 stones with G9, but the latter cannot escape a ladder. Another two stones are subsequently added before the end of the ladder comes within reach of the reading depth limit.	2.3.87	2.5.43	regginos-gnugo-2.3.87-199908231340.sgf.gz	52
65	990824	General disaster. Capturing the wrong stone costs some 80 points.	2.3.87	2.3.93	gnugo-2.3.87-bconwil-199908231921.sgf.gz	211
66	990826	Life&death error. The eyespace of the upper black group is evaluated to a single eye. This is a good example of a position where our model for local eye space games breaks.	2.3.89	2.7.93	NicklasE-gnugo-2.3.89-199908261143.sgf.gz	52
67	990826	Reading error. Trying to capture the G3 stones, GNU fails to recognize a simple shortage of liberties tesuji.	2.3.89	2.7.143	gnugo-2.3.89-gogol.sgf.gz	20
68	990826	Game not finished. There are still moves to play, but to find them requires a two step backfilling.	2.3.89	2.3.93	gnugo-2.3.89-gogol.sgf.gz	48
69	990827	Life&death error. The status of the A8 group is misevaluated to dead (two half eyes) because the tactical reading thinks A9 is a safe move for black. But after three captures, a white stone resides there.	2.3.90	2.7.63	tsumego-gnugo-2.3.89-199908270320.sgf.gz	158
70	990827	Life&death error. The A8 group is classified as alive, but is in fact unsettled. Also the significance of the move at J7 is missed.	2.3.90	2.3.93	tsumego-gnugo-2.3.89-199908270320.sgf.gz	178
71	990827	Wrong move. It's much better to save many strings (and capture one) with C19 than only one with C13 (which is later lost anyway).	2.3.90	2.5.31	tsumego-gnugo-2.3.89-199908270320.sgf.gz	188
72	990827	Insufficient backfilling. The correct move is to start at T15. (This game record is obviously very artificial.)	2.3.90	2.3.93	fill_lib.sgf.gz	254
73	990829	Life&death error. make_domains() never realizes that there should be an eyespace around S9.	2.3.90	3.1.8	guenther-gnugo-2.3.90-199908291130.sgf.gz	210
74	990831	Life&death error. Pointlessly capturing dead stones. The problem is that the stones are classified as UNKNOWN due to the unreliability of the escape factor.	2.3.91	2.7.223	Storer-gnugo-2.3.91-199908311532.sgf.gz	236
75	990831	Tuning problem. Among three different cuts and connections, the worst one is chosen. This is currently fairly hard to do right, since the patterns have no way of knowing how much territory may get lost. There is also a problem related to the thresholding of the connection value.	2.3.91	2.7.113	Storer-gnugo-2.3.91-199908311532.sgf.gz	238
76	990901	Reading error. For some reason the tactical reading thinks the white stone at P18 can be saved by a move at O19.	2.3.91	2.3.92	fire-gnugo-2.1.23gf-199905121832.sgf.gz	270
77	990901	Reading error. The tactical reading thinks the black B2 stones can be captured by a white move at D2. This seems to be a hashing problem.	2.3.91	2.3.93	gnugo-2.3.7-baduki.sgf.gz	277
78	990901	Reading error. R16 is somehow supposed to defend the S15 stones, although it is a self-atari and a very ineffective sacrifice.	2.3.91	2.3.94	gnugo-2.3.91-h3rb-199909011138.sgf.gz	65
79	990901	Reading error. Tactical reading thinks white K19 captures L19, failing to see the defense at J17.	2.3.91	2.7.143	gnugo-2.3.9gf-baduki.sgf.gz	168
80	990901	Reading error. The attacker mistakenly thinks it can capture some stones. This error also shows up in some of the preceding moves. This is the same incident as number 40, which was once fixed.	2.3.91	2.7.95	gnugo-2.3.36-baduki.sgf.gz	141
81	990901	Life&death error. Dragon at N18 is incorrectly classified as alive, because the false eye at P12 is not discovered, because the P15 stones are classifed as dead, although they are living in seki. Conclusion: we need to reclassify seki stones as alive at an early stage.	2.3.91	3.1.8	marcel-gnugo-2.3.37-199906171720.sgf.gz	286
82	990901	Reading error. Black B13 is supposed to capture the B16 stones. In reality B13 must be played at A14.	2.3.91	2.7.95	Storer-gnugo-2.3.91-199909012159.sgf.gz	184
83	990901	Reading error. Black L6 is assumed to capture the K9 stones, which doesn't work. With "--hash 0" this is read correctly so this is most likely a hashing problem.	2.3.91	2.3.93	Storer-gnugo-2.3.91-199909012159.sgf.gz	218
84	990903	Reading error. Tactical reading believes D6 is sufficient to capture the A8 stones because it doesn't find the defense move at B5.	2.3.93	2.7.143	gnugo-2.3.92-NLove-199909030923.sgf.gz	56
85	990906	Life&death error. The move at A12 is supposed to kill but doesn't work. The correct move is at B14.	2.3.93	2.7.93	Onzrock-gnugo-2.3.93-199909060530.sgf.gz	104
86	990906	Reading error. The move at H18 is supposed to capture the H17 stones, but there is a saving move at F16, which is never considered.	2.3.93	2.7.143	Onzrock-gnugo-2.3.93-199909060530.sgf.gz	214
87	990906	Wrong connection. Except for the reading error, there is also a problem that the connection at H15 is valued higher than the connection (and cut) at F16. In part this is probably to blame on the semeai module.	2.3.93	2.7.15	Onzrock-gnugo-2.3.93-199909060530.sgf.gz	214
88	990906	Missed connection. R5 is unreasonable until the connection at P6 has been played. The main problem here is that the safety of the corner is overestimated.	2.3.93	2.7.244	Onzrock-gnugo-2.3.93-199909060530.sgf.gz	252
89	990906	Missed ko. If GNU knew how to play ko, the corner group could be saved by throwing in at T1.	2.3.93	2.7.15	Onzrock-gnugo-2.3.93-199909060530.sgf.gz	276
90	990907	Illegal move. fill_liberty() comes up with an illegal move. While this bug is easy to get around, the position is rather one that the semeai module should handle.	2.3.94	2.3.97	Storer-gnugo-2.3.94-199909071440.sgf.gz	254
91	990907	Life&death error. T2 is not the critical point of the eye shape, S2 is.	2.3.94	2.7.16	greycat-gnugo-2.3.94-199909070541.sgf.gz	68
92	990908	Reading error. P2 is assumed to capture the O2 stones, but doesn't. The superb move at Q1, on the other hand, is not even considered.	2.3.96	2.7.172	gnugo-2.3.95-grumpsie-199909081042.sgf.gz	250
93	990908	Reading error. E2 is assumed to capture the E1 stone, but when white answers with the connection at D3, there is no continuation.	2.3.96	2.7.172	myao-gnugo-2.3.95-199909080934.sgf.gz	42
94	990908	Ko mistake. D17 is unreasonable. The ko must be connected at F16 immediately. Later things become totally stupid with respect to kos, between moves 68 and 116 there are a total of 27 ko captures, but not a single connection.	2.3.96	2.7.244	gnugo-2.3.95-glea-199909081420.sgf.gz	18
95	990908	Reading error. The tactical reading neglects to play O12 because it doesn't realize that white can defend the O14 stones with O15. It seems this is caused by an incorrect implementation of the ko rule. Neither player is allowed to play on the ko location O15.	2.3.97	2.3.98	alfalfa-gnugo-2.3.95-199909081822.sgf.gz	242
96	990909	Life&death error. E2 is urgent.	2.3.97	2.7.208	gnugo-2.3.97-berni-199909091242.sgf.gz	71
97	990910	Life&death error. It's necessary to add a stone at S11 in order to live.	2.3.98	2.7.214	scf-gnugo-2.3.97-199909101233.sgf.gz	175
98	990910	Reading error. M10 is supposed to capture the P12 cutting stones, but it doesn't. They can be captured by a move at P13 though.	2.3.98	2.5.37	scf-gnugo-2.3.97-199909101233.sgf.gz	219
99	990910	Bad move. The endgame move at K5 sets up for a white double atari at L5. Black has no choice but playing at L5 himself.	2.3.98	2.7.16	berni-gnugo-2.3.98-199909101338.sgf.gz	91
100	990910	Status error. It can be discussed whether this is a life&death, a semeai, or a reading error, but at the end of the game every single stone in the lower left part of the board is considered dead. The correct answer is that only the white stones are dead.	2.3.98	2.7.244	venom.sgf.gz	60
101	990910	Illegal move. The fuseki module comes up with an illegal move at A19, which is already occupied (in this game the first six stones are scattered at random).	2.3.98	2.4	abortgo.sgf.gz	12
102	990913	Reading error. For some reason it is believed that the K19 and J18 stones have a defense at H19, but still an unrelated move such as A11 is considered an attack move.	2.4	2.5.37	Storer-gnugo-2.3.99-199909122059.sgf.gz	230
103	990913	Bad exchange. It's much larger to connect at M3 than to capture a single stone with O7.	2.4	2.7.16	gnugo-2.4-docredneck-199909132117.sgf.gz	181
104	990920	Wrong connection. Locally C13 is a much better connection than D13, but still better is to capture at E10, producing an extra liberty that is sufficient to secure the connection at the other end of the worm.	2.5.7	2.7.16	gnugo-2.5.7-bconwil-199909192253.sgf.gz	63
105	990920	Life&death error. A second eye for the J19 group can be obtained by playing J13 followed by G13. GNU Go needs to learn about forcing sequences.	2.5.7	2.7.213	gnugo-2.5.7-bconwil-199909192253.sgf.gz	169
106	990920	Wrong capture. It's more profitable to capture at S19 than at T15.	2.5.7	2.7.208	gnugo-2.5.7-bconwil-199909192253.sgf.gz	215
107	990920	Life&death error. Q18 has an escape factor of 1 although it has no way whatsoever to escape.	2.5.7	2.7.144	guestx-gnugo-2.5.7-199909192332.sgf.gz	216
108a	990921	Reading error. G16 is caught in a snapback. The reason for the move was to give the F19 stones four liberties, but that's very temporary.	2.5.8	2.5.37	gnugo-2.4-me2dt-199909180323.sgf.gz	192
108b	990921	Wrong connection. E16 only covers one cutting point. C18, D18, and D17 cover both.	2.5.8	2.7.16	Storer-gnugo-2.5.8-199909211446.sgf.gz	148
109	990921	Shape error. P11 is a pointless atari, and worst of all ending in gote. The urgent play is at P15.	2.5.8	2.7.144	Storer-gnugo-2.5.8-199909211446.sgf.gz	172
110	990921	Shape error. Better to capture at Q7 than to connect at Q8.	2.5.8	2.7.144	Storer-gnugo-2.5.8-199909211446.sgf.gz	176
111	990921	Shape error. This certainly isn't the right time to slide under P16 to P18. Much better to simply cut at P15.	2.5.8	3.2	Storer-gnugo-2.5.8-199909211446.sgf.gz	178
112	990921	Shape error. Better to connect at N4 than to panic and just secure an eye at L4.	2.5.8	2.5.31	Storer-gnugo-2.5.8-199909211446.sgf.gz	270
113	990921	Wrong connection. Better to connect at N18 than at L17.	2.5.8	2.7.208	Storer-gnugo-2.5.8-199909211446.sgf.gz	292
114	990922	Reading error. R5 doesn't work to capture S4.	2.5.9	2.7.16	f000x-gnugo-2.5.9-199909221017.sgf.gz	36
117	990922	Life&death error. Securing life with H5 is absolutely necessary before anything in the vicinity can be attacked.	2.5.9	2.7.16	f000x-gnugo-2.5.9-199909221017.sgf.gz	196
118	990923	Reading error. The tactical reading doesn't see that white can capture att J1 to get a ko. A move at F2 or F1 wins the semeai unconditionally.	2.5.10	3.1.8	JesseLih-gnugo-2.5.10-199909231318.sgf.gz	252
119	990923	Missed ko. Black can still play at F1 or F2 to get a huge ko.	2.5.10	2.7.172	JesseLih-gnugo-2.5.10-199909231318.sgf.gz	256
120	990924	Tactical mistake. It's way better to capture at J9, thereby connecting the F13 stones to safety, than to run away at C9, reaching four liberties and a certain death.	2.5.10	2.7.16	gnugo-2.5.10-demp-199909241535.sgf.gz	139
121	990928	Fuseki mistake. Jumping out to E9 is urgent. The move at R7 is merely big.	2.5.11	2.7.165	ManyFaces-gnugo-2.5.11-199909281239.sgf.gz	24
122	990928	Wrong connection. More efficient to connect with E17 than with D17.	2.5.11	2.7.16	ManyFaces-gnugo-2.5.11-199909281239.sgf.gz	28
123	990928	Pointless capture. Playing A8 to capture A9 is vastly overvalued. Securing life by capturing three stones with E9 is far far better. It may be still better to run with the H13 stones instead.	2.5.11	2.7.146	ManyFaces-gnugo-2.5.11-199909281239.sgf.gz	60
125	990928	Reading error. If we knew how to attack stones with four liberties, we could capture with P3 instead of trying to defend the cutting point at N4 by more stupid means.	2.5.11	2.7.180	ManyFaces-gnugo-2.5.11-199909281239.sgf.gz	74
126	990928	Wrong connection. The connection to the K10 stone is not all that important. If we want to play it the move should be at K8 rather than J9 to better develop the position.	2.5.11	2.7.16	ManyFaces-gnugo-2.5.11-199909281239.sgf.gz	80
127	990928	Wrong connection again. A cut at N4 doesn't matter very much. Locally M4 (or even P3) is the right move, but it's bigger to improve the moyo on the right or securing the lower left against an invasion at H3.	2.5.11	2.7.16	ManyFaces-gnugo-2.5.11-199909281239.sgf.gz	82
128	990928	Strange shape. S6 is a move of doubtful shape and timing.	2.5.11	2.7.16	ManyFaces-gnugo-2.5.11-199909281239.sgf.gz	92
129	990928	Strategical mistake. H3 is becoming urgent. O19 is also about time to play in sente.	2.5.11	2.7.16	ManyFaces-gnugo-2.5.11-199909281239.sgf.gz	102
130	990928	Strategical mistake. Playing H2 in order to connect is very big here.	2.5.11	2.7.16	ManyFaces-gnugo-2.5.11-199909281239.sgf.gz	108
131	990928	Life & death error. Maybe there still can be a small life without F2, but it's definitely urgent.	2.5.11	2.7.106 --mouse	ManyFaces-gnugo-2.5.11-199909281239.sgf.gz	110
132	990928	Life & death error. Dying in gote is terrible, which is all that C4 accomplishes. The big move here is to block at L7.	2.5.11	2.7.144	ManyFaces-gnugo-2.5.11-199909281239.sgf.gz	116
133	990928	Life & death error. The P12 stones are not recognized as dead, with the result that three more moves are spent on killing the dead stones even more. Meanwhile white devastates some 15 or so points of territory.	2.5.11	2.7.98	ManyFaces-gnugo-2.5.11-199909281239.sgf.gz	122
134	990928	Life & death error. Playing C2 means that the lower left corner dies a second time in gote.	2.5.11	2.7.16	ManyFaces-gnugo-2.5.11-199909281239.sgf.gz	130
135	990928	Useless connection. There really is no point in connecting at O11.	2.5.11	2.5.31	ManyFaces-gnugo-2.5.11-199909281239.sgf.gz	160
136	990928	Tactical mistake. The move at P16 backfires badly as it sets up a double threat move for white at R16.	2.5.11	3.1.8	ManyFaces-gnugo-2.5.11-199909281239.sgf.gz	166
137	990928	Wrong connection. S15 is perfectly okey to connect with and a few points better then S14.	2.5.11	2.7.144	ManyFaces-gnugo-2.5.11-199909281239.sgf.gz	172
138	990928	Pointless defense. The tactical reading finds that the move at C18 defends the D18 stones, but since it only does this by reaching four liberties with no tendency to eyes or escape, the move is wasted.	2.5.11	2.5.31	ManyFaces-gnugo-2.5.11-199909281239.sgf.gz	218
139	990929	Eye-making panic. There is no point making an eye at E4 when it suffices fully well to connect at G5.	2.5.12	2.5.31	iq-gnugo-2.5.11-199909291130.sgf.gz	46
140	990929	Life&death error. The eye space containing the A5 stones is counted for two secure eyes. The correct value is one eye.	2.5.12	2.5.13	iq-gnugo-2.5.11-199909291130.sgf.gz	50
141	990929	Wrong move. A move at H8 would capture all relevant enemy stones as well as saving all relevant own stones. The played move at G8, as a contrast, merely breaks the connection between two enemy stones which would both be captured by H8.	2.5.12	2.5.37	iq-gnugo-2.5.11-199909291130.sgf.gz	52
142	990930	Reading problems. Deciding the safety of the E7 string requires a large search tree and gives rather incorrect results. The string can in reality not be attacked at J2 and D6 isn't much of a defense move. The generated move at J4 is also mostly pointless. Strangely enough the move also took much longer to generate when it was first played than when later loaded from the sgf file.	2.5.13	2.5.37	gnugo-2.5.12-November-199909292108.sgf.gz	33
143	990930	Reading error. N7 doesn't work to capture the four-liberty string M7. The only move in the position is to play S5 (or T2).	2.5.13	2.5.37	gnugo-2.5.12-docredneck-199909292256.sgf.gz	261
144	990930	Reading error. F17 doesn't work to capture the J19 string. The move must be played at K19.	2.5.13	2.7.143	gnugo-2.5.12-depayne-199909301201.sgf.gz	98
145	990930	Missed connection. A connection at S9 is urgent. This seems to both be a missing connection pattern and an escape route problem. Clearly there is only one way to escape and that is urgent.	2.5.13	2.7.210	gnugo-2.5.12-depayne-199909301201.sgf.gz	118
146	990930	Reading error. S6 is useless for saving the S14 stones. Neither does T15 work. Only R6 does the job.	2.5.13	2.5.31	gnugo-2.5.12-depayne-199909301201.sgf.gz	136
147	990930	Wrong move. G14 is a fairly stupid move here. F16 or J15 work much better.	2.5.13	2.7.23	gnugo-2.5.12-depayne-199909301201.sgf.gz	174
148	990930	Strategical error. Neglecting to block at N6 is much more expensive than what can be earned from C15.	2.5.13	2.7.23	gnugo-2.5.12-depayne-199909301201.sgf.gz	210
149	990930	Strategical error. Blocking at M3 is bigger than capturing at K5.	2.5.13	2.7.23	gnugo-2.5.12-depayne-199909301201.sgf.gz	228
150	990930	Missed ko. Capturing at A15 would start a ko for the life of the upper left corner.	2.5.13	2.7.210	gnugo-2.5.12-depayne-199909301201.sgf.gz	242
151	990930	Life&death error. An atari at A17 would capture the upper left corner right out.	2.5.13	2.7.210	gnugo-2.5.12-depayne-199909301201.sgf.gz	244
152	991001	Strategical error. Better to capture at P12 than to connect at Q13.	2.5.13	2.7.23	gnugo-2.5.13-comma-199910011105.sgf.gz	99
153	991001	Life&death error. Urgent to play G2 for life. The move at P4 is complete nonsense.	2.5.13	2.7.224	gnugo-2.5.13-baduki.sgf.gz	199
154	991001	Life&death error. Playing at J3 doesn't destroy any eye. In fact it's worse than useless since it puts seven stones into atari.	2.5.13	2.7.98	gnugo-2.5.13-baduki.sgf.gz	205
155	991001	Assertion failure in moyo.c. Reproducible.	2.5.13	2.7.23	gnugo-2.5.13-baduki2.sgf.gz	243
156	991004	Reading error. A14 is not sufficient to defend the B15 stones. In fact there is no defense.	2.5.15	3.1.8	exoss-gnugo-2.5.15-199910040106.sgf.gz	186
157	991004	Wrong move. P5 is clearly better than P4.	2.5.15	2.7.210	exoss-gnugo-2.5.15-199910040106.sgf.gz	214
158	991004	Missed double threat. White K17 captures either L17 or G19 and must be defended against.	2.5.15	2.7.213	exoss-gnugo-2.5.15-199910040106.sgf.gz	232
159	991004	Life&death error. White cannot be allowed to live here. B4 should be the easiest way to ensure this.	2.5.15	2.7.93	exoss-gnugo-2.5.15-199910040106.sgf.gz	246
160	991004	Life&death error. B4 is still sufficient to kill.	2.5.15	2.7.93	exoss-gnugo-2.5.15-199910040106.sgf.gz	250
161	991004	Life&death error. D9 doesn't fill a purpose. B7 is the move to kill with.	2.5.15	2.7.98	NicklasE-gnugo-2.5.15-199910041156.sgf.gz	38
162	991004	Life&death error. The move must be played at C7.	2.5.15	2.7.98 --mouse	NicklasE-gnugo-2.5.15-199910041156.sgf.gz	42
163	991004	Reading error. The move at A1 is ineffective and rather unreasonable since we have to play at C2 in the next move anyway. The ko potential in the position might make it harder for the reading.	2.5.15	2.7.229	NicklasE-gnugo-2.5.15-199910041156.sgf.gz	56
164	991005	Reading problems. Very many positions are visited for high backfill_depth values (like 14).	2.5.16	3.1.8	NicklasE-gnugo-2.5.16-199910051334.sgf.gz	44
165	991005	Missed opportunity. Now is the perfect time to play the double threat of N11. Also a move like H11 should be considered.	2.5.16	3.2	andrej-gnugo-2.5.16-199910051550.sgf.gz	108
166a	991005	Strategical error. There is no hurry to capture two stones with N3. Much more important to connect at E4.	2.5.16	3.2	andrej-gnugo-2.5.16-199910051550.sgf.gz	142
167a	991005	Strategical error. It's stupid to play M2 just to get an eye locally, when K3 or L2 captures all opponent stones, saves all own stones and keeps everything connected.	2.5.16	2.5.31	andrej-gnugo-2.5.16-199910051550.sgf.gz	148
166b	991005	Strategical error. It's necessary to block and connect at C9.	2.5.16	2.7.144	andrej-gnugo-2.5.16-199910051550.sgf.gz	174
168	991005	Reading problem. The move takes forever to generate with backfill_depth 14.	2.5.16	2.7.210	andrej-gnugo-2.5.16-199910051550.sgf.gz	216
169	991005	Strategical error. It's fairly small and kind of heavy to move out with D6 at this point. Better to sacrifice the C6 stones and play somewhere big.	2.5.12	2.7.158	gnugo-2.5.12-brego-199910051832.sgf.gz	18
170	991005	Strategical error. Very heavy and not very big to try to run away with E6.	2.5.12	2.7.158	gnugo-2.5.12-brego-199910051832.sgf.gz	32
171	991005	Strategical error. The focus must be on attacking the invading stone at C14.	2.5.12	2.7.158	gnugo-2.5.12-brego-199910051832.sgf.gz	60
172	991005	Tuning error. B8 is not at all convincing here.	2.5.12	2.5.19	gnugo-2.5.12-brego-199910051832.sgf.gz	82
173	991005	Tactical error. To stop intrusion, this move must be played at M17.	2.5.12	2.7.230	gnugo-2.5.12-brego-199910051832.sgf.gz	110
174	991005	Tactical error. It's necessary to continue at L7.	2.5.12	2.7.158	gnugo-2.5.12-brego-199910051832.sgf.gz	132
176	991005	Life&death error. There's something strange going on with the classification of the B14 group in previous and following moves.	2.5.12	2.7.158	gnugo-2.5.12-brego-199910051832.sgf.gz	172
178	991005	Tactical error. The L2 stones are classified as dead but can escape with K2. White should stop this by playing K2 first.	2.5.12	3.1.12	gnugo-2.5.12-brego-199910051832.sgf.gz	210
179	991005	Tactical error. O2 must be played at once.	2.5.12	2.5.19	gnugo-2.5.12-brego-199910051832.sgf.gz	214
180	991005	Tactical error. S17 isn't the best way to kill. S16 or T17 is much cleaner.	2.5.12	3.1.9	gnugo-2.5.12-brego-199910051832.sgf.gz	224
181	991005	Reading error. M1 is completely useless to play.	2.5.12	2.5.19	gnugo-2.5.12-brego-199910051832.sgf.gz	230
182	991005	Life&death error. B10 is abolutely urgent.	2.5.12	2.5.19	gnugo-2.5.12-brego-199910051832.sgf.gz	250
183	991005	Strategical error. We really shouldn't waste moves on removing nakade stones. Especially not when it's a single stone.	2.5.12	2.7.208	gnugo-2.5.12-brego-199910051832.sgf.gz	300
184	991006	Reading error. P19 doesn't work to save L19.	2.5.17	2.7.110	gnugo-2.5.17-cybersun-199910061445.sgf.gz	170
185	991006	Reading error. For some reason it's believed that playing out a snapback with P16 will defend the O19 stones.	2.5.17	2.7.208	steve-gnugo-2.5.17-199910071053.sgf.gz	204
186	991007	Reading error. Black fails to find a way to attack E9, although F9 is completely obvious. The problem seems to be in readlad2.	2.5.18	2.7.158	readbug.sgf.gz	25
187	991008	Game not completed. This position is not covered by the assumptions in fill_liberty().	2.5.18	2.5.19	gnugo-2.5.18-bconwil-199910080057.sgf.gz	227
188	991008	Reading error. The suicidal move at O13 is in no way helping to defend the four liberty string at P19. This is quite possibly a horizon problem.	2.5.20	2.5.37	brent-gnugo-2.5.20-199910121523.sgf.gz	238
189	991013	Life&death error. Pointless to destroy an eye when there is sufficient space to extend along the edge for a second eye with better profit.	2.5.21	2.7.98	alfalfa-gnugo-2.5.21-199910131456.sgf.gz	180
190	991013	Strategical error. Better to secure second eye locally and simultaneously cutting opponent with D18 than to connect through a ko with M16.	2.5.21	3.5.3	alfalfa-gnugo-2.5.21-199910131456.sgf.gz	186
191	991013	Life&death error. Saving the lunch J11 by connecting at J10 has no strategical value what so ever. The only move is to kill the second eye with F12.	2.5.21	2.5.31	alfalfa-gnugo-2.5.21-199910131456.sgf.gz	198
192	991013	Failed connection. L3 is insufficient, only L1 works.	2.5.21	2.7.229	alfalfa-gnugo-2.5.21-199910131456.sgf.gz	208
193	991013	Reading error. For some reason Q6 is thought to defend the R5 stones.	2.5.21	2.7.110	alfalfa-gnugo-2.5.21-199910131456.sgf.gz	216
194	991013	Reading error. T8 is totally ineffective for defending S12 and in fact loses a few points.	2.5.21	2.7.110	alfalfa-gnugo-2.5.21-199910131456.sgf.gz	222
195	991013	Reading error. A6 fails to capture the A9 stones.	2.5.21	2.7.172	alfalfa-gnugo-2.5.21-199910131456.sgf.gz	252
196	991019	Reading error. The only move here is P3, but the tactical reading thinks that the K19 string can be defended by a move at N16 and therefore plays M12 to capture it. This looks very much like another fourlib depth horizon problem.	2.5.23	2.5.25	tank-gnugo-2.5.23-199910181732.sgf.gz	260
197	991020	Semeai error. The move at G15 should be sufficient to win the semeai, but simultaneously connecting at D19 is in every way superior.	2.5.24	2.7.210	steve-gnugo-2.5.24-199910201049.sgf.gz	122
198	991020	Life&death error. Capturing a lunch with H1 doesn't suffice to live. It's urgent to secure a second eye with H5.	2.5.24	2.7.105	NicklasE-gnugo-2.5.24-199910201542.sgf.gz	54
199	991024	Reading error. G2 doesn't work to capture any of the white stones unconditionally. The effective move is H2, which kills the entire corner.	2.5.27	2.7.110	sprotte-gnugo-2.5.27-199910242003.sgf.gz	42
200	991026	Small semeai error. Small_semeai() incorrectly removes the attacking points of the B3 and P19 strings.	2.5.27	2.5.31	gnugo-2.5.27-WeGo-199910261640.sgf.gz	226
201	991027	Life&death error. The eye space along the edge around P19 is badly misevaluated, confusing GNU Go enough to miss the obvious move at Q19.	2.5.28	2.7.98 --mouse	gnugo-2.5.28-baduki.sgf.gz	271
202	991027	Reading error. The tactical reading doesn't find the defense move at T13 for the worm at M19.	2.5.28	3.2	gnugo-2.5.28-baduki.sgf.gz	297
203	991028	Reading problem. Analyzing the tactical safety of the worm at B14 takes forever and finally, after two million nodes, ends with the incorrect result that the worm can be defended at A13.	2.5.28	2.5.31	gnugo-2.5.27-Obtuse-199910251009.sgf.gz	70
204	991028	Reading error. The tactical reading incorrectly believes that O11 can capture N11. There might be a ko problem in the position.	2.5.28	2.5.31	gnugo-2.5.28-Mary-199910281429.sgf.gz	190
205	991029	Ko reading error. The tactical status of H9 and H8 are both dependent on the ko starting with J8. It doesn't look like the reading notices this fact.	2.5.29	2.5.41	kotest.sgf.gz	34
206	991031	Ko reading error. The tactical reading fails to find that B17 can be captured by D15, conditioned on a ko.	2.5.30	2.5.35	mdjenk-gnugo-2.5.30-199910311741.sgf.gz	84
207	991114	Reading error. The move at O10 is assumed to defend P10, but fails.	2.5.37	2.5.41	Storer-gnugo-2.5.37-199911131535.sgf.gz	146
208	991122	Reading error. The tactical reading doesn't see the defense of S5 at O5, thus missing to play this huge move.	2.5.40	2.5.41	docredneck-gnugo-2.5.40-199911222005.sgf.gz	117
209	991123	Semeai error. White must immediately remove a liberty at S14 or T14 to stop black from making a local seki with R16.	2.5.41	2.7.229	gnugo-2.5.41-baduki.sgf.gz	193
210	991123	Semeai error. White could have made seki by playing R11 before P10.	2.5.41	3.2	gnugo-2.5.41-baduki.sgf.gz	259
211	991125	Wrong connection. J6 gives better eye shape than H6, which is important in this position.	2.5.42	3.1.8	NicklasE-gnugo-2.5.42-199911251237.sgf.gz	22
212	991125	Life&death error. H8 is urgent.	2.5.42	2.7.211	NicklasE-gnugo-2.5.42-199911251237.sgf.gz	26
213	991125	Life&death error. G9 is pointless.	2.5.42	2.7.208	NicklasE-gnugo-2.5.42-199911251237.sgf.gz	40
214	991125	Reading error. F5 is totally ineffective.	2.5.42	2.7.20	NicklasE-gnugo-2.5.42-199911251237.sgf.gz	50
215	991125	Pointless capture. Black doesn't benefit at all from capturing H8. The same goes for move 52.	2.5.42	2.7.208	NicklasE-gnugo-2.5.42-199911251237.sgf.gz	54
216	991125	Liberty filling error. D2 does not fill a liberty but loses a point by playing inside enemy territory.	2.5.42	2.5.45	NicklasE-gnugo-2.5.42-199911251237.sgf.gz	64
217	991126	Reading error. The tactical reading incorrectly thinks that the B19 string can be captured by a move at A17.	2.5.42	2.7.20	Storer-gnugo-2.5.42-199911261628.sgf.gz	216
218	991203	Life&death error. F1 is a killing move, but since find_lunch() doesn't consider diagonal neighbors, it fails to see the importance of saving the E1 stone and the eye shape analysis goes wrong.	2.5.45	2.7.158	mir-gnugo-2.5.45-199912031040.sgf.gz	136
219	991216	Caching error. The tactical reading finds several defenses for the truly captured Q11 group. If caching of read results is turned off, these defenses disappear.	2.7.3	2.7.20	demp-gnugo-2.7.2-199912150845.sgf.gz	144
220	991220	Caching error. Running --decidestring Q13 gives different results with and without caching.	2.7.6	2.7.20	gnugo-2.7.6-Loup-199912201052.sgf.gz	164
221	991230	Tactical blunder. B7 doesn't defend against the double atari at D7. The move should be played at C7.	2.7.10	2.7.210	double_threat.sgf.gz	35
222	991230	Tactical blunder. C9 sets up a double threat for white at C7. The correct move is for black to play C7.	2.7.10	2.7.210	double_threat.sgf.gz	39
223	000104	Tactical mistake. The move at S17 leads to a ko while the correct move at R19 kills the corner right out.	2.7.12	2.7.211	sfr-gnugo-2.7.12-200001031756.sgf.gz	228
224	000104	Reading error. The move at C14 does not defend E13. Only D13 defends.	2.7.12	2.7.63	sfr-gnugo-2.7.12-200001031756.sgf.gz	264
225	000104	Reading error. An unconditional defense of H18 is found at H17 but it should be contingent on ko.	2.7.13	2.7.208	gnugo-2.5.53-gnugo-2.7.13.sgf.gz	252
226	000104	Reading error. H17 leads to ko or seki but this move is not found. At least the tactical reading should find that H18 can be defended at H17 with ko.	2.7.13	2.7.208	gnugo-2.5.53-gnugo-2.7.13.sgf.gz	254
227	000106	Life&death mistake. J4 is thought to be an urgent move, although all black stones are already dead. Notice that no move is suggested for the left side of the board. The difference is that G3 can be tactically captured and that H2 can be tactically defended, which confuses the eye space analysis.	2.7.14	2.7.224	nakade2.sgf.gz	1
228	000112	Illegal move. The eye finder insists on doing an illegal ko capture at T18.	2.5.56	2.5.57	bconwil-gnugo-2.5.55-200001121850.sgf.gz	98
229	000114	Suicidal blunder. Playing at A3 loses a won semeai worth more than 80 points. It seems to be caused by mistakes in the tactical reading and make_worms/make_dragons together.	2.5.56	2.7.21	psa-gnugo-2.5.56-200001141224.sgf.gz	282
230	000126	Life&death blunder. Connecting at O16 gives two eyes. The proposed move at M1 is tiny.	2.7.23	2.7.43	gnugo-2.5.13-baduki2.sgf.gz	243
231	000216	Reading error. The tactical reading doesn't find the black defense at A1, which also means that the eye analysis doesn't get a chance to find A1, since whole the corner is considered white eyespace.	2.6	2.7.29	gnugo-2.6-arion-200002161821.sgf.gz	206
232	000317	Life&death mistake. The move at A6 lives but a 15 stone tail is lost. Playing D6 would live as well as save the tail.	2.7.50	2.7.98	gnugo-2.7.49-pochi-200003171922.sgf.gz	260
233	000321	Ineffective cut. The pattern matching finds a bogus cutting move at O15. The real cut is at O14.	2.7.52	3.3.11	seiwa-gnugo-2.7.50-200003211949.sgf.gz	186
234	000330	Life&death mistake. A18 is a critical point for white's corner group. GNU Go thinks the group is alive.	2.7.60	2.7.105 --mouse	takechan-gnugo-2.7.59-200003301102.sgf.gz	164
235	000331	Mouse code mistake. The mouse code thinks Q1 will save the S3 dragon. The true vital point is S2.	2.7.61 --mouse	2.7.98 --mouse	mdjenk-gnugo-2.7.61-200003310427.sgf.gz	110
236	000402	Mouse code disaster. The mouse code thinks that a move at P6 would attack the N8 dragon. But P6 puts nine own stones into atari and white lives easily by capturing those.	2.7.64 --mouse	2.7.98 --mouse	qwer-gnugo-2.7.64-200004021335.sgf.gz	116
237	000503	Mouse code mistake. The move at A3 is supposed to attack A4, which clearly is ineffective.	2.7.83 --mouse	2.7.105 --mouse	NicklasE-gnugo-2.7.83.sgf.gz	42
238	000524	Life&death mistake. Make_domains() fails to recognize the single point eyespace at E16.	2.7.99	2.7.105	genok-gnugo-2.7.98-200005232232.sgf.gz	252
239	000525	Reading and mouse mistake. The black string at A3 can't be defended unconditionally at B1. Only A5 works right out.	2.7.100	2.7.154	oriented-gnugo-2.7.98-200005242334.sgf.gz	58
240	000531	Mouse reading mistake. Attacking the L3 dragon at B1 is very ineffective since white lives easy and big by capturing at A5.	2.7.105 --mouse	2.7.208	epoupaert-gnugo-2.7.105-200005311831.sgf.gz	65
241	000531	Strategical mistake. It's urgent to block at D9.	2.7.105 --mouse	3.2	epoupaert-gnugo-2.7.105-200005311831.sgf.gz	69
244	000531	Life&death mistake. Defending at H3 leads at best to ko for the whole group. Better to e.g. play E2 and sacrifice the K3 stones.	2.7.105 --mouse	2.7.208	epoupaert-gnugo-2.7.105-200005311831.sgf.gz	101
245	000531	Life&death mistake. Still possible to get a ko by blocking at G1.	2.7.105 --mouse	2.7.211	epoupaert-gnugo-2.7.105-200005311831.sgf.gz	103
246	000604	Reading mistake. L9 doesn't suffice to capture M12.	2.7.111	2.7.112	amodeo-gnugo-2.7.111-200006041957.sgf.gz	74
247	000604	Reading mistake. M12 can't be attacked, in particular is O11 unsuccessful.	2.7.111	2.7.112	amodeo-gnugo-2.7.111-200006041957.sgf.gz	76
248	000615	Reading mistake. It's absolutely vital to play either G14, H14, or J14 to win a huge semeai. The played move at O12 is mostly ineffective.	2.7.115	2.7.158	berni-gnugo-2.7.115-200006092328.sgf.gz	166
249	000615	Mouse reading mistake. R18 is not an effective way to attack P18.	2.7.115 --mouse	2.7.158	berni-gnugo-2.7.115-200006092328.sgf.gz	216
250	000615	Life&death mistake. Urgent to play A2 in order to avoid ko.	2.7.115	2.7.229	berni-gnugo-2.7.115-200006092328.sgf.gz	228
251	000616	Mouse reading mistake. P17 is a very strange move. The correct move is of course to connect at N15.	2.7.117 --mouse	2.7.158	Storer-gnugo-2.7.117-200006161425.sgf.gz	258
252	000619	Reading error. The tactical reading incorrectly believes H3 can be attacked at J2.	2.7.120	2.7.172	heikkile-gnugo-2.7.119-200006192121.sgf.gz	134
253	000629	Mouse reading error. The mouse code doesn't find a defense for the R14 dragon, although R9 should be clearly sufficient.	2.7.123	2.7.124	gnugo-2.7.121-TSa-200006280343.sgf.gz	77
254	000701	Mouse reading error. The mouse code thinks black C16 defends D16, which obviously is unrealistic.	2.7.123	2.7.124	gnugo-2.7.123-MortenI-200006300026.sgf.gz	147
255	000701	Escape factor problem. The status of D16 is reported as unknown just because the escape factor turns out to be 2.	2.7.123	2.7.158	gnugo-2.7.123-MortenI-200006300026.sgf.gz	147
256	000703	Mouse reading mistake. J12 does not save the O9 dragon. The correct move is to connect and make a second eye with the capture at M11.	2.7.124	2.7.208	mso-gnugo-2.7.124-200007021749.sgf.gz	157
257	000703	Move valuation problem. It's better to live by defending at H7 than by blocking at J2.	2.7.124	2.7.158	mso-gnugo-2.7.124-200007021749.sgf.gz	215
258	000703	Mouse reading mistake. The C4 dragon can be defended, contrary to what the mouse reading says. At least the dragon manages to live in the game.	2.7.124	2.7.158	mso-gnugo-2.7.124-200007021728.sgf.gz	147
259	000703	Move valuation problem. Defending everything at M2 is superior to defending only something at N2.	2.7.124	2.7.158	mso-gnugo-2.7.124-200007021728.sgf.gz	251
260	000703	Mouse reading mistake. The mouse reading thinks that the B19 dragon can be attacked at D19 and defended at H19. In reality it is the other way round!	2.7.124	2.7.158	mso-gnugo-2.7.124-200007021728.sgf.gz	283
261	000705	Missed chance to escape. It's urgent to play N7 to escape with the center group.	2.7.127	2.7.208	Storer-gnugo-2.7.123-200007011548.sgf.gz	108
262	000708	Mouse reading problems. Deciding the mouse status of R3 requires some 3600 variations.	2.7.128	3.1.8	gnugo-2.7.128-rubus-200007071724.sgf.gz	73
263	000712	Mouse reading mistake. The mouse code states that the M4 dragon is dead as it stands but also that it can be defended at J2.	2.7.132	2.7.160	gnugo-2.7.132-Shaitan-200007121542.sgf.gz	32
264	000712	Mouse reading mistake. The mouse code thinks D13 can be defended at B13.	2.7.132	2.7.149	gnugo-2.7.132-Shaitan-200007121542.sgf.gz	116
265	000712	Mouse reading mistake. The mouse code fails to find a defense for S14. Jumping to R18 looks effective.	2.7.132	2.7.208	gnugo-2.7.132-Shaitan-200007121542.sgf.gz	302
266	000712	Mouse reading mistake. The mouse code still fails to find a defense for S14. Jumping to R18 again looks effective.	2.7.132	2.7.208	gnugo-2.7.132-Shaitan-200007121542.sgf.gz	304
267	000714	Mouse reading mistake. The mouse code thinks C7 can be attacked at E5, missing to take the cut at C4 into account.	2.7.133 --alternative_escape	2.7.229	gnugo-2.7.133-nixxx-200007141041.sgf.gz	46
268	000714	Mouse reading mistake. The mouse code thinks B6 can be defended at B3.	2.7.133	2.7.161	gnugo-2.7.133-nixxx-200007141041.sgf.gz	54
269	000714	Life and death mistake. Instead of P19 black can still kill the white corner.	2.7.133	2.7.210	Storer-gnugo-2.7.133-200007141436.sgf.gz	200
270	000714	Tactical blunder. F11 only saves some of the stones. J12 or H12 saves everything.	2.7.133	2.7.210	Storer-gnugo-2.7.133-200007141436.sgf.gz	236
272	000716	Reading mistake. M18 does not defend M17.	2.7.133	2.7.208	DerkJan-gnugo-2.7.133-200007151314.sgf.gz	24
273	000716	Semeai mistake. L18 is not sufficient to win the semeai between K17 and M18.	2.7.133	2.7.210	DerkJan-gnugo-2.7.133-200007151314.sgf.gz	28
274	000716	Missed connection. A move at P19 would connect L19 and R19.	2.7.133	2.7.134	DerkJan-gnugo-2.7.133-200007151314.sgf.gz	240
275	000716	Move misevaluation. J5 should be valued clearly less than P4.	2.7.133	2.7.222	DerkJan-gnugo-2.7.133-200007151314.sgf.gz	250
276	000716	Reading mistake. N2 can not be defended. Update, P5 gives ko.	2.7.133	2.7.210	DerkJan-gnugo-2.7.133-200007151314.sgf.gz	252
277	000716	Pointless move. F19 is a move inside opponent territory.	2.7.133	2.7.210	DerkJan-gnugo-2.7.133-200007151314.sgf.gz	266
278	000716	Semeai mistake. S17 is unreasonable.	2.7.133	2.7.208	Storer-gnugo-2.7.133-200007151450.sgf.gz	76
279	000716	Semeai mistake. L17 may work, but J18 leaves much less aji. Somewhere later in the sequence there definitely is some mistake done as black finally loses the semeai.	2.7.133	2.7.208	Storer-gnugo-2.7.133-200007151450.sgf.gz	146
280	000716	Reading mistake. F16 cannot be defended and clearly not at G10.	2.7.133	2.7.143	Storer-gnugo-2.7.133-200007151450.sgf.gz	160
281	000716	Wrong connection. M4 or N5 is much better than O5.	2.7.133	3.1.11	Storer-gnugo-2.7.133-200007151450.sgf.gz	212
282	000716	Wrong connection. M2 is the most efficient way to connect at P2.	2.7.133	3.1.8	Storer-gnugo-2.7.133-200007151450.sgf.gz	252
283	000716	Life&death mistake. Necessary to play B1, C2, or B3 to secure life with territory.	2.7.133	2.7.180	Storer-gnugo-2.7.133-200007151450.sgf.gz	316
284	000716	Life&death mistake. Urgent to play B3 and get a seki instead of ko.	2.7.133	3.1.11	Storer-gnugo-2.7.133-200007151450.sgf.gz	318
285	000716	Life&death blunder. B1 lives with territory, C1 dies.	2.7.133	2.7.180	Storer-gnugo-2.7.133-200007151450.sgf.gz	322
286	000716	Pointless capture. Capturing at A1 doesn't make life.	2.7.133	2.7.208	Storer-gnugo-2.7.133-200007151450.sgf.gz	324
287	000716	Life&death blunder. Playing H1 is suicidal. Correct way to live is D1.	2.7.133	2.7.158	gnugo-2.7.133-Shaitan-200007151622.sgf.gz	103
288	000716	Wrong connection. K7 is superior to J7.	2.7.133	2.7.158	gnugo-2.7.133-Shaitan-200007151622.sgf.gz	111
289	000716	Mouse reading mistake. The mouse code thinks white O12 can kill P12 but O10 defends. The easiest way to kill is O10 but N13 also seems to work while saving some stones as well.	2.7.133	2.7.180	gnugo-2.7.133-Shaitan-200007151622.sgf.gz	251
290	000729	Missed connection. R11 is unreasonable. Locally the connection at Q11 is the only move.	2.7.138	3.2	mso-gnugo-2.7.138-200007241933.sgf.gz	30
291	000729	Bad shape. Extending towards the edge with T12 fills no purpose. Both S13 and S11 are clearly superior.	2.7.140	2.7.158	qtr-gnugo-2.7.140-200007251353.sgf.gz	54
292	000729	Small move. Q2 is a small endgame gote move. K3 is the obvious big move in the area.	2.7.140	2.7.158	qtr-gnugo-2.7.140-200007251353.sgf.gz	70
293	000729	Bad shape. M5 is in all ways inferior to M6.	2.7.140	2.7.229	qtr-gnugo-2.7.140-200007251353.sgf.gz	86
294	000729	Strategical mistake. The tactical defense at K5 has no strategical viability.	2.7.140	2.7.158	qtr-gnugo-2.7.140-200007251353.sgf.gz	88
295	000729	Endgame mistake. Defending at F4 is clearly bigger than filling ko at B18.	2.7.140	2.7.147	qtr-gnugo-2.7.140-200007251353.sgf.gz	250
296	000729	Liberty filling mistake. T17 is entirely within opponent territory.	2.7.140	2.7.146	qtr-gnugo-2.7.140-200007251353.sgf.gz	274
297	000729	Mouse mistake. G5 must be defended immediately, preferrably at H2.	2.7.138	2.7.159	gnugo-2.7.138-Shaitan-200007222221.sgf.gz	95