Author: Vincent Diepeveen
Date: 19:46:18 12/17/99
Go up one level in this thread
On December 16, 1999 at 21:17:46, Dann Corbit wrote:
>A similar thread brings up an interesting question, "What is AI?"
Lucky for AI everyone has his own definition.
>An old test was supposed to be that if we are playing a remote opponent we can't
>tell if it is a human or a machine. I think that can probably be achieved now
>(especially if we throw in a bit of randomness).
This is dangerous to say. In principle i agree with you.
However when we try to with a program which is written in mind to
play chess as strong as possible, then recognition is very easy.
On the other hand if the program doesn't need to play chess as well
as possible (for example at 1300 level), then it's i personally think
very easy to make something that's not recognizable from an opponent
except if you allow communication.
Let's take the easiest example. Suppose you start playing a game. You play
1.e4, and hear nothing from your opponent. He or it simply forfeits.
Is it a human or is it a computer?
- it can be a computer which got hung
- it can be a human who had to go to the toilet or had to go somewhere
else for an appointment. He might have gone sick, or just wanted to
make fun of you.
Suppose you play someone with black, and your unknown opponent plays:
Nc3-b1-c3-b1-c3-b1-c3-b1, till he gets mated by you or till there is
repetition. in case of capture he simply moves ra1-b1-a1-b1, or the
other knight.
Is it a computer or is it a human who's just doing something for fun?
In all these cases we talk however about a kind of silly form of chess.
As soon as we get to the category 'strong playing chessprograms' and
'strong playing humans'. Suppose we know someone of 2200+, it's not known
how much exactly; so it can be both a 2200 player but also a 2820 player,
plays you. It can be either computer or human. Will you detect whether
it's a computer or a human?
My answer then directly is YES. It's VERY easy to detect whether it's
a human or a computer. Many tests are possible. the funniest test is
described by Bruce Moreland at this thread. Against some commercial
programs that test would already fail however.
A few of those funny and short tests are however already sufficient.
Most convincing thing by far however is time division. Give me a
time division and i'll tell you whether it's a computer or a human.
Even if the operator goes to the toilet regurarly i still can tell
it to you. I need of course the moves too. Just printout from a dude
a number of games with time division, post them together with the
games itselve. I'll pick the computer and the human.
By far not all games it's possible to say just based upon the moves who
it is. Recently someone accused another one of making usage of a computer,
as all moves but 2 were matched by a certain program. We talk about a
strong player here > 2500 blitz at ICC.
I go have a look at the game. All moves looked very obvious to me.
Very hard to say whether it would have been a computer or a human
at first sight.
The only thing i know is that i would have played that game with black
exactly the same.
To protect account names i have removed them. Here is the game.
the unattended computer is white. The question is: is black 100% human
or not?
1. d4 (0:05) d5 (0:07)
2. Nf3 (0:00) Nf6 (0:06)
3. c4 (0:00) e6 (0:03)
4. Bg5 (0:00) Be7 (0:04)
5. e3 (0:00) O-O (0:03)
6. Nbd2 (0:00) Nbd7 (0:03)
7. Rc1 (0:00) c5 (0:05)
8. cxd5 (0:41) Nxd5 (0:43)
9. Bxe7 (0:04) Qxe7 (0:05)
10. e4 (0:31) Nb4 (0:14)
11. a3 (0:38) Nc6 (0:09)
12. dxc5 (0:35) Nxc5 (0:08)
13. Qc2 (1:06) Nd7 (0:26)
14. Bb5 (0:00) Nde5 (0:16)
15. O-O (0:11) Rd8 (0:26)
16. Bxc6 (0:21) Nxc6 (0:11)
17. e5 (0:13) Rd5 (0:20)
18. Ne4 (0:00) Bd7 (0:25)
19. Nc3 (0:01) Nxe5 (0:08)
20. Nxd5 (0:17) Nxf3+ (0:06)
21. Kh1 (0:01) exd5 (0:23)
22. gxf3 (0:01) Qf6 (0:22)
23. Qb3 (0:00) Bc6 (0:14)
24. Rfd1 (1:11) d4 (0:23)
25. Rd3 (0:19) Re8 (0:15)
26. Qd1 (0:26) Qf5 (0:09)
27. Rxc6 (0:21) bxc6 (0:06)
28. Rxd4 (0:20) h6 (0:12)
29. Kg2 (0:34) Qf6 (0:07)
30. Qd3 (0:20) c5 (0:56)
31. Ra4 (0:14) Qxb2 (0:30)
32. Rxa7 (0:18) Qe5 (0:09)
33. Qd7 (0:17) Qg5+ (0:17)
34. Kf1 (0:00) Rf8 (0:29)
35. a4 (0:22) Qh5 (0:01)
36. Kg2 (0:19) Qg6+ (0:06)
37. Qg4 (0:06) Qb6 (0:23)
38. Re7 (0:21) Rd8 (0:27)
39. Qf5 (0:16) Qg6+ (0:13)
40. Qxg6 (0:01) fxg6 (0:02)
41. Ra7 (0:11) Rc8 (0:31)
42. Rb7 (0:05) c4 (0:11)
43. Rb2 (0:04) c3 (0:29)
44. Rc2 (0:00) Kf7 (0:13)
45. Kf1 (0:21) Rc4 (0:10)
46. a5 (0:19) Rc5 (0:15)
47. Kg2 (0:17) Kf6 (0:11)
48. a6 (0:13) Ke5 (0:07)
49. a7 (0:13) Ra5 (0:08)
50. Rxc3 (0:01) Rxa7 (0:01)
51. h4 (0:10) Rb7 (0:08)
52. Kg3 (0:12) Rb6 (0:20)
53. f4+ (0:12) Kf5 (0:01)
54. f3 (0:10) Kf6 (0:21)
55. Rc8 (0:01)
{Game drawn by mutual agreement} 1/2-1/2
Please AI researchers give your judgement on this game.
I'm very interested to hear PRO's and Contra's on this game.
I sure didn't see on first sight who's being right or wrong.
Remember, a reasonable player can already take back a piece very
quickly. If you're using a computer at icc you just force recapture.
This player hell sure is a decent player.
I'm very interested in opinions on this game. and i'm sure not
willingly to give other games of the blackplayer at standard level,
as that would make task too easy...
>Then what tends to happen is that we say, "That's not really artificial
>intelligence. After all, it's just a machine, so it _can't_ be." We simply
>move the target and we are safe from the encroachment of the machine into "our"
>domain.
>If (for instance) I was playing some opponent using Winboard and I only knew it
>was one of:
>"Kasparov"
>"Deep Blue"
>
>I would have no way of guessing which was which, since either one would pound my
>stuffings out effortlessly.
The stronger the computer/human question is the easier it's to figure it out.
Something very TOUGH to figure out however is a human who's using a computer
for blunder checking but futher playing himselve.
if the choice is however a programs move at ALL MOVES, or a human who has
made ALL MOVES himselve then it's easy to figure out.
Here the classical time division of 2 unattended computers playing
each other.
Cruijff - Crafty (smoves crafty %07)
1. d4 (0:02) d5 (0:01)
2. c4 (0:03) e6 (0:00)
3. Nc3 (0:02) Nf6 (0:00)
4. cxd5 (0:03) exd5 (0:00)
5. Bg5 (0:01) Be7 (0:00)
6. e3 (0:01) O-O (0:00)
7. Bd3 (0:02) Nbd7 (0:00)
8. Nf3 (0:02) c6 (0:00)
9. h3 (0:02) Ne4 (0:11)
10. Bxe7 (0:03) Qxe7 (0:10)
11. Qc2 (0:01) Ndf6 (0:11)
12. O-O (0:06) Re8 (0:02)
13. Ne5 (0:03) Nxc3 (0:04)
14. bxc3 (0:02) c5 (0:06)
15. a4 (0:05) c4 (0:08)
16. Be2 (0:13) Nd7 (0:03)
17. Nxd7 (0:03) Bxd7 (0:05)
18. Rfb1 (0:01) Rab8 (0:08)
19. a5 (0:02) Qg5 (0:08)
20. Kh2 (0:09) Bf5 (0:08)
21. Qb2 (0:02) Bxb1 (0:08)
22. Rxb1 (0:01) b6 (0:07)
23. Bf3 (0:07) Qf5 (0:17)
24. Qb5 (0:07) a6 (0:08)
25. Qb4 (0:10) b5 (0:07)
26. Rb2 (0:02) Rbd8 (0:04)
27. Qc5 (0:06) Qd3 (0:07)
28. Qb4 (0:08) Qf1 (0:07)
29. Rd2 (0:14) Qa1 (0:07)
30. Rd1 (0:06) Qa2 (0:07)
31. Kg1 (0:03) Re6 (0:03)
32. Kf1 (0:11) Qc2 (0:06)
33. Ke1 (0:03) f6 (0:07)
34. Rd2 (0:05) Qg6 (0:06)
35. Kd1 (0:02) Red6 (0:06)
36. Kc1 (0:02) Rc6 (0:07)
37. h4 (0:07) Qf7 (0:06)
38. Kb2 (0:05) f5 (0:06)
39. g3 (0:03) Qe6 (0:02)
40. Rd1 (0:03) Rdd6 (0:06)
41. Rh1 (0:02) Qd7 (0:02)
42. Ka2 (0:09) Rh6 (0:06)
43. Ka3 (0:08) Rc8 (0:06)
44. Ka2 (0:02) Rc7 (0:04)
45. Ka3 (0:01) Qe6 (0:02)
46. Ka2 (0:02) Rc6 (0:00)
47. Ka3 (0:01) Rc8 (0:02)
48. Ka2 (0:01) Kh8 (0:02)
49. Ka3 (0:01) Rf6 (0:02)
50. Ka2 (0:01) Rd8 (0:02)
51. Ka3 (0:01) Rg6 (0:02)
52. Ka2 (0:01) Qd7 (0:02)
53. Ka3 (0:01) Rc8 (0:02)
54. Ka2 (0:01) Rh6 (0:02)
55. Ka3 (0:01) Rd8 (0:02)
56. Ka2 (0:01) Re8 (0:02)
57. Ka3 (0:01) Kg8 (0:02)
58. Ka2 (0:01) Rc6 (0:03)
59. Ka3 (0:00) Ree6 (0:02)
60. Ka2 (0:01) Rf6 (0:01)
61. Ka3 (0:01) Qf7 (0:02)
62. Ka2 (0:01) Qe6 (0:02)
63. Ka3 (0:01) Rc8 (0:02)
64. Ka2 (0:01) Qd7 (0:02)
65. Ka3 (0:01) Rcc6 (0:02)
66. Ka2 (0:01) Rg6 (0:01)
67. Ka3 (0:01) Rc8 (0:02)
68. Ka2 (0:01) Rc7 (0:02)
69. Ka3 (0:01) Rd6 (0:03)
70. Ka2 (0:00) Rf6 (0:02)
71. Ka3 (0:00) Qd8 (0:02)
72. Ka2 (0:00) Rcc6 (0:02)
73. Ka3 (0:00) Rh6 (0:02)
74. Ka2 (0:00) Rhd6 (0:02)
75. Ka3 (0:01) Qe8 (0:02)
76. Ka2 (0:01) Qd7 (0:02)
77. Ka3 (0:01) Qf7 (0:02)
78. Ka2 (0:01) Rd8 (0:03)
79. Ka3 (0:01) Qd7 (0:02)
80. Ka2 (0:01) Rdc8 (0:02)
81. Ka3 (0:02) Kh8 (0:02)
82. Ka2 (0:01) Rf6 (0:02)
83. Ka3 (0:01) Kg8 (0:02)
84. Ka2 (0:17) Re6 (0:02)
85. Ka3 (0:01) Kh8 (0:03)
86. Ka2 (0:01) h6 (0:03)
87. h5 (0:03) Rf6 (0:00)
88. Ka3 (0:03) Kg8 (0:02)
89. Rh4 (0:03) Qf7 (0:02)
90. Ka2 (0:02) Rcc6 (0:02)
91. Ka3 (0:01) Qd7 (0:03)
92. Ka2 (0:01) Qd8 (0:03)
93. Ka3 (0:03) Rf8 (0:02)
94. Ka2 (0:02) Rc7 (0:02)
95. Rh1 (0:04) Rf6 (0:02)
96. Ka3 (0:01) Qd7 (0:02)
97. Ka2 (0:01) Rc8 (0:01)
98. Ka3 (0:01) Rfc6 (0:03)
99. Ka2 (0:01) Re8 (0:03)
100. Ka3 (0:00) Qe6 (0:02)
101. Ka2 (0:00) Qf7 (0:03)
102. Ka3 (0:01) Rd8 (0:02)
103. Ka2 (0:00) Rc7 (0:03)
104. Ka3 (0:00) Qe6 (0:02)
105. Ka2 (0:01) Rcd7 (0:03)
106. Ka3 (0:00) Rc8 (0:03)
107. Ka2 (0:00) Rcc7 (0:02)
108. Ka3 (0:00) Rc6 (0:02)
109. Ka2 (0:00) Rdd6 (0:02)
110. Ka3 (0:00) Rd8 (0:02)
111. Ka2 (0:00) Rcc8 (0:02)
112. Ka3 (0:00) Re8 (0:03)
113. Ka2 (0:00) Rf8 (0:03)
114. Ka3 (0:00) Rc6 (0:03)
115. Ka2 (0:00) Rf6 (0:02)
116. Ka3 (0:00) Qd7 (0:02)
117. Ka2 (0:00) Qd8 (0:02)
118. Ka3 (0:00) Rfd6 (0:02)
119. Ka2 (0:00) Qf6 (0:02)
120. Ka3 (0:00) Rd7 (0:03)
121. Ka2 (0:00) Qd8 (0:02)
122. Ka3 (0:00) Rf7 (0:02)
123. Ka2 (0:01) Rcc7 (0:02)
124. Ka3 (0:00) Rc8 (0:02)
125. Ka2 (0:00) Rf8 (0:04)
126. Ka3 (0:00) Rc6 (0:02)
127. Ka2 (0:01) Qd7 (0:01)
128. Ka3 (0:01) Qf7 (0:04)
129. Ka2 (0:00) Rb8 (0:01)
130. Ka3 (0:01) Ra8 (0:02)
131. Ka2 (0:01) Rac8 (0:02)
132. Ka3 (0:00) Qe6 (0:01)
133. Ka2 (0:00) Kh8 (0:01)
134. Ka3 (0:01) Kh7 (0:03)
135. Ka2 (0:01) Rg8 (0:02)
136. Ka3 (0:01) g6 (0:01)
137. hxg6+ (0:03) Rxg6 (0:00)
138. Ka2 (0:07) Qd7 (0:03)
139. Ka3 (0:01) Rgf6 (0:03)
140. Ka2 (0:01) Kg7 (0:03)
141. Ka3 (0:01) Kg8 (0:03)
142. Ka2 (0:00) Qd6 (0:02)
143. Ka3 (0:00) Qd8 (0:03)
144. Ka2 (0:01) Rg6 (0:01)
145. Ka3 (0:02) Rg7 (0:03)
146. Ka2 (0:03) Rg5 (0:00)
147. Ka3 (0:02) Kf7 (0:03)
148. Ka2 (0:01) Kg7 (0:03)
149. Ka3 (0:01) Kh7 (0:03)
150. Ka2 (0:01) Kh8 (0:02)
151. Ka3 (0:01) Kg8 (0:03)
152. Ka2 (0:01) Qd7 (0:03)
153. Ka3 (0:02) Rg7 (0:03)
154. Ka2 (0:01) Rf7 (0:04)
155. Ka3 (0:01) Rd6 (0:03)
156. Ka2 (0:01) Rff6 (0:03)
157. Ka3 (0:00) Qc7 (0:02)
158. Ka2 (0:01) Rd8 (0:02)
159. Ka3 (0:01) Qd7 (0:01)
160. Ka2 (0:01) Qd6 (0:01)
161. Ka3 (0:04) Qe6 (0:03)
162. Ka2 (0:01) Kf7 (0:03)
163. Ka3 (0:01) Qd7 (0:04)
164. Ka2 (0:01) Qd6 (0:01)
165. Qxd6 (0:09) Rdxd6 (0:03)
166. Ka3 (0:00) Ke6 (0:03)
167. Kb4 (0:01) Rd8 (0:03)
168. Rh4 (0:03) Kd6 (0:03)
169. Rh5 (0:03) Rg8 (0:03)
170. Rh4 (0:02) Rg7 (0:01)
171. Rh1 (0:01) Rgg6 (0:03)
172. Rh4 (0:01) Rg5 (0:01)
173. Rh1 (0:01) Re6 (0:03)
174. Bd1 (0:08) Rg7 (0:03)
175. Bf3 (0:03) Kc6 (0:05)
176. Rh5 (0:02) Rf6 (0:03)
177. Rh1 (0:04) Kd6 (0:03)
178. Rh5 (0:11) Re7 (0:03)
179. Rh1 (0:02) Rb7 (0:03)
180. Rh5 (0:01) Rd7 (0:01)
181. Rh1 (0:02) Kc6 (0:03)
182. Rh5 (0:01) Rd8 (0:03)
183. Rh1 (0:00) Rg8 (0:03)
184. Rh5 (0:01) Rg5 (0:03)
185. Rh1 (0:01) Re6 (0:03)
186. Rh4 (0:09) Reg6 (0:05)
187. Rh1 (0:01) Rd6 (0:01)
188. Rh4 (0:02) Rgg6 (0:00)
189. Rh5 (0:02) Rgf6 (0:03)
190. Rh1 (0:01) Rde6 (0:03)
191. Rh5 (0:01) Re8 (0:01)
192. Rh1 (0:09) Re7 (0:03)
193. Rh5 (0:01) Rd7 (0:01)
194. Rh1 (0:01) Rc7 (0:03)
195. Rh5 (0:01) Ra7 (0:03)
196. Rh1 (0:01) Kd6 (0:03)
197. Rh5 (0:01) Ra8 (0:03)
198. Rh1 (0:02) Rc8 (0:03)
199. Rh5 (0:01) Rh8 (0:01)
200. Bd1 (0:10) Kc6 (0:03)
201. Bf3 (0:02) Rc8 (0:00)
202. Rh1 (0:03) Ra8 (0:00)
203. Rh5 (0:01) Raf8 (0:03)
204. Rh1 (0:01) Kd6 (0:03)
205. Rh5 (0:01) R8f7 (0:03)
206. Rh1 (0:01) Kc6 (0:06)
207. Rh5 (0:02) Re7 (0:01)
208. Rh1 (0:02) Rb7 (0:03)
209. Rh5 (0:01) Rb8 (0:01)
210. Rh1 (0:01) Rg6 (0:03)
211. Rh5 (0:02) Rf8 (0:01)
212. Rh1 (0:03) Rh8 (0:03)
213. Rh5 (0:07) Rg5 (0:03)
214. Rh1 (0:02) h5 (0:03)
215. Rh4 (0:01) Kd6 (0:03)
216. Rf4 (0:04) Rh7 (0:03)
217. Rh4 (0:01) Rg6 (0:03)
218. Rf4 (0:10) Rf6 (0:00)
219. Rh4 (0:01) Rh8 (0:04)
220. Rh1 (0:40) Rfh6 (0:10)
221. Rh4 (0:03) Rg8 (0:06)
222. Bd1 (0:06) Rg7 (0:03)
223. Bf3 (0:01) Rgh7 (0:03)
224. Bd1 (0:02) Rg6 (0:03)
225. Bf3 (0:01) Rg8 (0:03)
226. Bd1 (0:01) Rgh8 (0:01)
227. Bf3 (0:01) Rb8 (0:01)
228. Bd1 (0:07) Rf8 (0:00)
229. Bf3 (0:01) Rhh8 (0:03)
230. Bd1 (0:03) Rh6 (0:03)
231. Bf3 (0:01) Rf7 (0:02)
232. Bd1 (0:01) Rfh7 (0:03)
233. Bf3 (0:01) Ra7 (0:01)
234. Bd1 (0:01) Rh8 (0:03)
235. Bf3 (0:01) Rg7 (0:04)
236. Bd1 (0:01) Rg5 (0:03)
237. Bf3 (0:02) Rg6 (0:03)
238. Bd1 (0:01) Rhh6 (0:03)
239. Bf3 (0:01) Rg5 (0:01)
240. Bd1 (0:01) Kc6 (0:03)
241. Bf3 (0:01) Rh8 (0:03)
242. Bd1 (0:01) Ra8 (0:05)
243. Bf3 (0:01) Kd6 (0:03)
244. Bd1 (0:05) Rag8 (0:03)
245. Bf3 (0:01) Kc6 (0:02)
246. Bd1 (0:02) Rh8 (0:03)
247. Bf3 (0:01) Rd8 (0:01)
248. Bd1 (0:02) Kd6 (0:03)
249. Bf3 (0:01) Rb8 (0:02)
250. Bd1 (0:01) Ra8 (0:03)
251. Bf3 (0:02) Rf8 (0:03)
252. Bd1 (0:01) Rc8 (0:03)
253. Bf3 (0:01) Kc6 (0:03)
254. Bd1 (0:01) Rf8 (0:02)
255. Bf3 (0:02) Re8 (0:04)
256. Rf4 (0:12) Re7 (0:04)
257. Rh4 (0:02) Kd6 (0:00)
258. Rf4 (0:03) Rd7 (0:03)
259. Rh4 (0:01) Rh7 (0:01)
260. Bd1 (0:02) Rgg7 (0:01)
261. Bf3 (0:02) Re7 (0:05)
262. Bd1 (0:03) f4 (0:03)
263. gxf4 (0:13) Reg7 (0:00)
264. Bc2 (1:17) Rh8 (0:00)
265. f5 (0:36) Rg2 (0:10)
266. Rf4 (0:31) h4 (0:09)
267. f6 (0:02) h3 (0:06)
268. f7 (0:02) Ke7 (0:07)
{White resigns} 0-1
>So the question stands, "What is AI?" and along with it, the related question,
>"Are chess programs intelligent?"
Not a single chessprogrammer has willingly tried to make a program play
strong using randomness in his program on purpose (only some bugs might
cause some randomness). Chess Programs are therefore
just following a set of predefined rules.
Some experiments with random moves are practical big BS to consider.
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.