WEBVTT - Game On: Computers vs. Humans 0:00:00.280 --> 0:00:02.840 Brought to you by the reinvented two thousand twelve camera. 0:00:03.160 --> 0:00:08.920 It's ready. Are you get in touch with technology? With 0:00:09.039 --> 0:00:17.760 tech Stuff from how stuff works dot com. Hello again, everyone, 0:00:17.800 --> 0:00:20.079 welcome to tech stuff. My name is Chris Pollette, and 0:00:20.079 --> 0:00:22.760 I am an editor at how stuff works dot com. 0:00:23.160 --> 0:00:26.880 Seeming across from me, as usual, is senior writer Jonathan Strickland. Yes, 0:00:26.960 --> 0:00:29.480 I'm owed old enough to remember when the m c 0:00:29.640 --> 0:00:33.480 P was just a chess program. There you go, that's 0:00:33.479 --> 0:00:35.920 pretty good. It's a good quote. Huh yeah, yeah, we's 0:00:36.159 --> 0:00:40.120 now I'm hungry for Pepperidge Farm. You can't get that 0:00:40.280 --> 0:00:44.400 from hreh Um. We're going to talk today about some 0:00:44.479 --> 0:00:51.479 pretty famous matchups between human beings and computers, uh, in 0:00:51.479 --> 0:00:54.960 in various games. And this is one of those things 0:00:55.000 --> 0:00:59.040 that people have been interested in. Well, really it goes 0:00:59.080 --> 0:01:03.520 back even before computers, because it really kind of dips 0:01:03.520 --> 0:01:08.959 into the whole man versus machine, um the idea, right, 0:01:09.000 --> 0:01:11.360 I mean this goes all the way back to John 0:01:11.440 --> 0:01:17.600 Henry versus the the steam steam engine. Yeah, and uh, 0:01:17.840 --> 0:01:20.160 people have wondered like when we're when we reach a 0:01:20.160 --> 0:01:25.800 point where computers would become better than humans at various 0:01:25.840 --> 0:01:29.759 games and we've reached the point with several games where 0:01:29.959 --> 0:01:34.039 we can pretty much definitively say computers are better at 0:01:34.040 --> 0:01:37.240 it than we are, you know, And a lot of 0:01:37.280 --> 0:01:41.120 the podcast where we've mentioned artificial intelligence or AI in 0:01:41.160 --> 0:01:47.680 the past, UM, we have talked about gaming and in 0:01:47.720 --> 0:01:52.600 those situations we were talking more about computer opponents when 0:01:52.600 --> 0:01:55.560 you're playing against something in a computer game or a 0:01:55.640 --> 0:01:58.800 console game, and that's a little different than what we're 0:01:58.840 --> 0:02:03.320 talking about here. But because and UH, I think actually 0:02:03.360 --> 0:02:07.680 I was I was reading an article UM by Jonathan Schaefer, 0:02:08.000 --> 0:02:12.120 Vadim Blikko, and Michael Burrow in the I E E. 0:02:12.520 --> 0:02:17.960 Spectrum and UH, it was talking about computer opponents in 0:02:18.200 --> 0:02:22.440 UH in software, specifically the game Fear f E A R, 0:02:22.520 --> 0:02:25.400 which is the first Encounter Assault Recon and how that 0:02:25.480 --> 0:02:29.360 has that's known as for its excellent AI. But UM, 0:02:29.400 --> 0:02:34.480 basically these games we're talking about today are are classic 0:02:34.560 --> 0:02:39.959 board games, classic tests of UH, you know, one player's 0:02:40.120 --> 0:02:44.640 skill and strategy and thinking against another human being. And 0:02:44.680 --> 0:02:47.320 we wanted to know, you know, whether that's possible now 0:02:47.880 --> 0:02:49.440 and sort of depends on the game and it sort 0:02:49.440 --> 0:02:52.480 of depends on the opponent. But that's really according to 0:02:52.560 --> 0:02:56.200 the authors in this in the Spectrum article. Um, it 0:02:56.280 --> 0:02:58.760 really started with with board games, and that's one of 0:02:58.800 --> 0:03:01.840 the things that has gone into a development of AI 0:03:01.919 --> 0:03:05.440 in computer games of all kinds, was you know, these 0:03:05.480 --> 0:03:10.640 attempts in the past to develop a formidable opponent, somebody 0:03:10.680 --> 0:03:12.519 that or something that you can play against where you 0:03:12.520 --> 0:03:14.200 don't have to go you don't really like to play 0:03:14.200 --> 0:03:16.919 a game at chess, but you know, Steve's work in 0:03:16.960 --> 0:03:19.440 a night and and I don't have anybody else that 0:03:19.520 --> 0:03:20.919 I can think of, and I really want to play 0:03:20.960 --> 0:03:23.160 a game of chess. We want something. You don't want to, uh, 0:03:23.280 --> 0:03:26.360 something that's going to annihilate you instantly, nor do you 0:03:26.400 --> 0:03:31.760 want something that's going to be a complete pushover. So yeah, exactly, 0:03:31.880 --> 0:03:35.400 he's like playing me and chess. I only know how 0:03:35.400 --> 0:03:40.640 to lose it chess. Um. So yeah, I mean it's 0:03:40.680 --> 0:03:44.800 it's an interesting question and sort of a philosophical question. 0:03:44.840 --> 0:03:49.280 It's not just programming. Yeah, I mean, there's a lot 0:03:49.320 --> 0:03:52.120 of these games have have elements to them that go 0:03:52.240 --> 0:03:56.520 beyond just a the physical number of moves that are 0:03:56.600 --> 0:04:00.800 possible at any given point. Right, some games that's pretty 0:04:00.880 --> 0:04:05.520 much all the game is. For example, Connect four. Oh yeah, 0:04:05.520 --> 0:04:09.800 Connect four. We're gonna talk a little bit about solving games. Now, 0:04:09.960 --> 0:04:13.680 to solve a game means that you have determined what 0:04:13.760 --> 0:04:16.400 all the possible moves are, and you know what the 0:04:16.440 --> 0:04:20.960 best possible move in any given situation is. Um. If 0:04:20.960 --> 0:04:23.760 you have perfectly solved a game, then in theory, you 0:04:23.800 --> 0:04:26.920 would be able to play a game, even pick up 0:04:26.920 --> 0:04:29.440 a game that had already been started by someone else, 0:04:29.760 --> 0:04:32.960 and play it from that point forward, playing the best 0:04:33.000 --> 0:04:36.040 possible moves, even if that person who had previously been 0:04:36.040 --> 0:04:38.880 playing made a lot of mistakes. Not many games have 0:04:38.960 --> 0:04:42.160 been solved to that level, but that a lot of 0:04:42.240 --> 0:04:44.719 games have been solved to the point where two players 0:04:44.839 --> 0:04:47.880 starting off with a fresh board and going from there. 0:04:48.400 --> 0:04:51.920 Uh is there is there a perfect way to play 0:04:52.000 --> 0:04:56.360 where you will either what will essentially where you'll never lose. Yeah, 0:04:56.880 --> 0:04:59.800 you're speaking for personal experience. I think Reversey is one 0:04:59.800 --> 0:05:03.720 of the games, UM, if you're not familiar, it's the 0:05:04.160 --> 0:05:07.640 game with the black and white pieces that you try 0:05:07.680 --> 0:05:11.200 to basically surround your opponent to the point where you 0:05:11.200 --> 0:05:14.640 flip all the pieces to your color and or or 0:05:14.680 --> 0:05:17.640 as many as you can and outnumber your opponents. Now 0:05:17.960 --> 0:05:21.080 I've played, uh, computer opponents in which on the easy 0:05:21.120 --> 0:05:23.800 mode it was a cakewalk, and I'd switch to the 0:05:23.839 --> 0:05:26.599 next level, and it already it seemed as though the 0:05:26.640 --> 0:05:29.679 machine had already been programmed with the best possible counter 0:05:29.760 --> 0:05:32.240 move to whatever it was, to the point where if 0:05:32.279 --> 0:05:35.159 I played on easy, I would almost always win unless 0:05:35.160 --> 0:05:39.719 I did something really stupid, and on even medium, it 0:05:39.800 --> 0:05:43.280 would know exactly how to counter every move I make, 0:05:43.360 --> 0:05:45.760 to the point where I could not win no matter 0:05:45.839 --> 0:05:49.400 how well I would play the game. And uh, I 0:05:49.760 --> 0:05:52.159 think there are certain games for which you could say, okay, 0:05:52.160 --> 0:05:55.080 well these are the moves. In this case, it's the 0:05:55.080 --> 0:05:58.200 best possible move. And you know, when you're playing a person, 0:05:58.800 --> 0:06:01.479 you go, well, I you know, oh, look he made 0:06:01.480 --> 0:06:04.240 a mistake. And you know that. You know, Jonathan in 0:06:04.279 --> 0:06:07.200 this case is sitting across from me. He made a mistake. Well, 0:06:07.240 --> 0:06:09.400 he may not have been thinking about it. The computer 0:06:09.480 --> 0:06:11.320 is not going to make a mistake. It's already been 0:06:11.360 --> 0:06:15.120 programmed with that in mind. Yeah. The mistakes computers make 0:06:15.480 --> 0:06:20.560 are based upon not expecting behavior from a human, right, right, 0:06:20.600 --> 0:06:22.760 I mean that's that's the way computers make mistakes. They 0:06:22.760 --> 0:06:25.680 don't make mistakes in the same way humans do. So 0:06:25.760 --> 0:06:29.560 for example, getting back to the Connect for uh example, 0:06:30.240 --> 0:06:34.240 that game has been solved to the point where, assuming 0:06:34.279 --> 0:06:37.640 perfect play on both sides, yes, the first player will 0:06:37.640 --> 0:06:40.240 always win. Yes, there is no way for the first 0:06:40.240 --> 0:06:44.520 player to lose if both players are playing perfectly. Uh. 0:06:44.680 --> 0:06:47.920 Perfect play means you go for that center spot, folks, 0:06:47.960 --> 0:06:50.719 Just just an f y I if you need to, uh, 0:06:50.800 --> 0:06:52.880 if you need to play Connect four and you're the 0:06:52.920 --> 0:06:56.080 first player, get grabbed that center spot on the bottom 0:06:56.160 --> 0:06:58.480 row and uh and then just play perfectly from there 0:06:58.480 --> 0:07:01.599 and you'll always win. Easier so than done. Really, but 0:07:02.120 --> 0:07:04.400 that that was one of the games that was solved 0:07:04.440 --> 0:07:06.200 fairly early on. But there are other games have been 0:07:06.200 --> 0:07:09.680 solved as well, and some games are not truly solvable. 0:07:10.520 --> 0:07:14.240 In fact, I wrote an article about computer versus human 0:07:14.360 --> 0:07:17.680 game matchups and the first one I mentioned was b 0:07:18.040 --> 0:07:22.800 KG nine point eight. That would be the computer versus 0:07:23.080 --> 0:07:27.360 Luigi Villa or Villa so uh so Luigi here he 0:07:27.480 --> 0:07:31.200 was a backgammon player. Yes, as am I, I nowhere 0:07:31.240 --> 0:07:36.840 near as good as Luigi. So so Luigi played a 0:07:36.960 --> 0:07:41.040 match against b KG nine point eight in June nineteen 0:07:41.240 --> 0:07:47.160 seventy nine, and the game was got a fair amount 0:07:47.160 --> 0:07:52.200 of press because the computer program beat the the player. However, 0:07:52.360 --> 0:07:56.160 backgammon has an element in it that is unlike games 0:07:56.200 --> 0:08:01.200 like connect for or checkers or chess. Yes, that would 0:08:01.200 --> 0:08:04.880 be dice. There's an element of chance in backgam and 0:08:04.960 --> 0:08:09.440 so depending upon the dice rolls, you may be luckier 0:08:09.440 --> 0:08:12.040 than your opponent, and that luck might be enough for 0:08:12.080 --> 0:08:15.800 you to beat someone who is technically more skilled at 0:08:15.840 --> 0:08:17.440 the game than you are, someone who is who is 0:08:17.480 --> 0:08:20.720 better at formulating strategies and picking the right moves and 0:08:20.760 --> 0:08:24.720 the best possible outcomes. Uh, you might win just because 0:08:24.800 --> 0:08:26.760 you get on a lucky streak, and that seems to 0:08:26.800 --> 0:08:29.960 be what happened with this computer program. Yes, should we 0:08:30.000 --> 0:08:35.440 mention our earlier podcast on random number generators, because if 0:08:35.440 --> 0:08:38.000 you were playing a computer and you were not using 0:08:38.280 --> 0:08:42.480 an actual die or in this case dice, then uh, 0:08:42.920 --> 0:08:45.840 and you're relying on the computer to generate random numbers, Well, 0:08:45.920 --> 0:08:50.400 in general you can't because they're they're using an algorithm 0:08:50.760 --> 0:08:56.360 and the computer number generators will generally go in a sequence. Now, 0:08:56.480 --> 0:08:58.360 this kind of goes back to what Chris was saying 0:08:58.360 --> 0:09:00.920 earlier about playing a video game in a video game 0:09:00.920 --> 0:09:04.360 world where you're playing against video game opponents. Technically in 0:09:04.440 --> 0:09:07.920 that kind of environment. Uh, the AI is totally different 0:09:08.000 --> 0:09:11.319 from a board game because it controls the entire environment. 0:09:11.800 --> 0:09:15.040 So you could create a video game that compensates for 0:09:15.120 --> 0:09:19.680 really strong players by pitting the environment against the players. Yes, 0:09:19.840 --> 0:09:21.920 you know. I mean it's it's completely within the realm 0:09:21.920 --> 0:09:25.440 of possibility. In the situations we're talking about. It's supposed 0:09:25.480 --> 0:09:29.480 to be a fair game between our as fair game 0:09:29.520 --> 0:09:31.640 as you can get between a computer and a human. 0:09:32.200 --> 0:09:33.920 There's not supposed to be any element that's going to 0:09:34.000 --> 0:09:36.440 be within the computer's control that could that could weigh 0:09:36.480 --> 0:09:39.560 it in favor of the computer, right, because it's supposed 0:09:39.559 --> 0:09:43.040 to be a test of intelligence relative intelligence, like as 0:09:43.080 --> 0:09:47.640 we mentioned with Alan Touring, Yes, yes, the touring test um. Now, 0:09:47.640 --> 0:09:50.240 in that case, the touring test was a tested determined 0:09:50.280 --> 0:09:53.320 whether or not you could tell the difference between a 0:09:53.520 --> 0:09:57.800 human respondent and a computer responded to a series of questions. 0:09:58.160 --> 0:10:00.760 In this case, we're talking about what or a computer 0:10:00.840 --> 0:10:03.480 is able to formulate a winning strategy against the human. 0:10:04.320 --> 0:10:08.360 The next one on my list was Chinook versus Marion 0:10:08.480 --> 0:10:13.840 Tinsley with Checkers. Yes, this was Checkers. Now Tinsley, Uh, 0:10:14.600 --> 0:10:20.800 was an accomplished checkers player, amazing champion. He had won 0:10:20.880 --> 0:10:26.520 the championship from nineteen fifty five to nineteen two, and 0:10:26.559 --> 0:10:30.959 he had only lost five games between nineteen fifty and 0:10:33.160 --> 0:10:36.160 that seems reasonably decent. I think I've only played five 0:10:36.200 --> 0:10:41.200 games since n probably lost all five of them. Well, 0:10:41.600 --> 0:10:45.320 this is another game, though, if unless I misunderstand the 0:10:45.480 --> 0:10:49.720 article that you so aptly wrote, um uh, in which 0:10:49.760 --> 0:10:53.200 there is a way to play perfectly. Yeah, this was 0:10:53.360 --> 0:10:56.120 that was not discovered till two thousand seven, or not 0:10:56.120 --> 0:10:59.840 not truly proven until two thousand seven. So when Tinsley 0:11:00.040 --> 0:11:04.480 played against chinook Um, the game had yet to be solved, 0:11:04.520 --> 0:11:10.120 and sadly, Tensley actually passed away before he could. Before 0:11:10.160 --> 0:11:12.480 they could definitively say whether or not the computer was 0:11:12.520 --> 0:11:16.320 a superior player, they would play thirties something games in 0:11:16.320 --> 0:11:20.400 a row and end and draw a draw every time. Um. 0:11:20.480 --> 0:11:23.080 But that suggests that I'm sorry, go ahead, I'm just 0:11:23.080 --> 0:11:27.680 gonna say that suggests that that Tense was able to 0:11:27.679 --> 0:11:31.120 play a perfect game. Yes, because in two thousand and seven, 0:11:31.640 --> 0:11:36.080 the team that that created chinook did prove or demonstrate 0:11:36.200 --> 0:11:38.240 that they had solved the game, and that if you 0:11:38.280 --> 0:11:41.200 play perfectly on either side. Let's say both sides are 0:11:41.200 --> 0:11:43.520 playing perfectly, it will the game will always end in 0:11:43.520 --> 0:11:46.840 a draw. So checkers is different from Connect four and 0:11:46.880 --> 0:11:49.360 that if you if both sides are playing perfectly, there 0:11:49.520 --> 0:11:51.280 is no winner, It's going to be a draw, whereas 0:11:51.280 --> 0:11:54.280 with Connect four, whoever goes first wins. There's actually I 0:11:54.320 --> 0:11:56.599 remember reading about a game I didn't jot down it 0:11:56.720 --> 0:11:59.720 shot down my notes unfortunately, where it was the second 0:11:59.720 --> 0:12:02.720 play if the second player plays perfectly, If both players 0:12:02.760 --> 0:12:05.600 are playing perfectly, the second player will always win, which 0:12:05.600 --> 0:12:08.240 is which is interesting because you know, it just shows 0:12:08.280 --> 0:12:12.440 that it all depends upon the actual style of the game. Now, 0:12:12.440 --> 0:12:14.160 the next one that's on my list is probably the 0:12:14.160 --> 0:12:17.760 biggest one. Yes, yeah, the one that everyone knows or 0:12:17.840 --> 0:12:22.280 has heard about, right, IBM's famous deep blue Yes against 0:12:22.360 --> 0:12:26.839 Gary Kasparov, who was the world champion chess player at 0:12:26.840 --> 0:12:30.360 the time. And uh they actually had two matches to 0:12:30.559 --> 0:12:34.000 two series of games. The first was in n and 0:12:34.040 --> 0:12:39.160 in that series Kasparov actually came out the victor. He 0:12:39.320 --> 0:12:47.160 managed to thank you, came out the winner, you know, wiener. 0:12:47.880 --> 0:12:51.280 Uh So Kasparov one in the first series of games 0:12:51.280 --> 0:12:56.320 in and then IBM redoubled their efforts, and as Kasparov 0:12:56.440 --> 0:13:00.480 says in one of his articles, um they redoubled it's uh, 0:13:00.520 --> 0:13:03.600 they doubled its its processing power. Uh. The article I 0:13:03.640 --> 0:13:05.720 was referring to it is called the Chess Master and 0:13:05.920 --> 0:13:11.280 the Computer, which was a uh, actually pretty interesting article 0:13:11.920 --> 0:13:15.720 at any rate. In that game, Kasparov won the first 0:13:16.160 --> 0:13:19.640 game of the match in the series, lost the second one. 0:13:20.360 --> 0:13:22.960 The next three were draws, and the final game deep 0:13:23.000 --> 0:13:26.920 blue one. So in that case, deep Blue one the series. 0:13:27.679 --> 0:13:31.760 And this was this made news worldwide because chess was 0:13:31.800 --> 0:13:35.360 one of those games that that people thought this game 0:13:35.400 --> 0:13:39.239 is so complex and it relies so much upon intuition 0:13:39.480 --> 0:13:43.720 and and strategy that goes beyond just numbers, that it 0:13:43.760 --> 0:13:46.880 was going to take ages before a computer could beat 0:13:46.960 --> 0:13:51.880 a true chess champion. And go ahead, I was gonna say, 0:13:51.880 --> 0:13:54.400 And this isn't just any chess champion. I mean he 0:13:54.440 --> 0:13:59.280 had the highest rating, yes, offered by the the world 0:13:59.280 --> 0:14:03.280 body that covered chess. Yeah, there are various ways to 0:14:03.400 --> 0:14:07.080 rate chess players, but yes, Kasparov held the highest rating 0:14:07.240 --> 0:14:10.040 of all human players at this point. There are computer 0:14:10.120 --> 0:14:13.160 players that have higher ratings, in fact, ratings higher than 0:14:13.200 --> 0:14:19.600 what people thought were possible. UM. But UH. The interesting 0:14:19.640 --> 0:14:23.080 thing here about this, this UH matchup was that Kasparov 0:14:23.160 --> 0:14:26.400 said he thought that he had not prepared properly for 0:14:26.440 --> 0:14:28.160 the games, and that he wished he could have had 0:14:28.200 --> 0:14:35.720 another rematch. But ibm H shut down that that project 0:14:35.760 --> 0:14:37.800 because they had pretty much proven their point, you know 0:14:37.840 --> 0:14:39.720 that they were They set out to create a computer 0:14:39.760 --> 0:14:42.160 program that could beat the world champion in chess, and 0:14:42.160 --> 0:14:44.680 it and it worked, So why continue that? I mean, 0:14:44.720 --> 0:14:48.480 there's no other application really for that. So they shut 0:14:48.520 --> 0:14:51.000 down that program and and Kesperov never had a chance 0:14:51.040 --> 0:14:54.720 to to try and and UH and and prove that 0:14:54.760 --> 0:14:59.240 he could beat this this computer. Now, since then, Kasparov 0:14:59.280 --> 0:15:04.120 has played other computer programs that are on UM extremely 0:15:04.160 --> 0:15:07.400 powerful machines, and in some cases he has played to 0:15:07.400 --> 0:15:10.360 a draw or one, and in other cases he has lost. 0:15:10.600 --> 0:15:13.000 And he has said that we have essentially reached the 0:15:13.040 --> 0:15:18.320 era where a a powerful computer running the right software, 0:15:18.680 --> 0:15:22.080 or even desktop computers running the right software, can beat 0:15:22.400 --> 0:15:26.480 a a grand master chess player. UM. And part of 0:15:26.520 --> 0:15:31.400 that is because these these computer programs often have an 0:15:31.560 --> 0:15:37.560 entire database of opening moves available to to UH to 0:15:37.680 --> 0:15:40.680 look at before making any kind of a move against 0:15:40.720 --> 0:15:43.800 an opponent. And the other is that chess is a 0:15:43.880 --> 0:15:46.280 really complicated game. You can't really map out all the 0:15:46.320 --> 0:15:49.800 possible moves easily because there's so many different pieces and 0:15:49.840 --> 0:15:51.560 they all behave differently, and there are a lot of 0:15:51.560 --> 0:15:56.160 different options at any given time. But what chess computers 0:15:56.240 --> 0:16:00.480 do very well is they can plot out all the 0:16:00.480 --> 0:16:03.280 possible moves once you get down to a certain number 0:16:03.280 --> 0:16:07.320 of chess pieces towards the end game. And so a 0:16:07.320 --> 0:16:10.200 lot of these chess programs are very very good at 0:16:10.320 --> 0:16:13.520 plotting out the best moves when they're only say seven 0:16:13.800 --> 0:16:17.040 or eight pieces on the board. Once you get more 0:16:17.080 --> 0:16:22.160 than that, the the different variables are so vast that 0:16:22.160 --> 0:16:25.280 it's a lot harder to account for all of them. 0:16:25.560 --> 0:16:31.680 That being said, it's gonna be pretty much it'll it'll 0:16:31.720 --> 0:16:34.600 take a lot of luck, really to beat a powerful 0:16:34.680 --> 0:16:38.360 chess program at this point, it can happen. Um. Casparov 0:16:38.440 --> 0:16:42.479 has shown that by playing in a kind of unconventional 0:16:42.480 --> 0:16:45.120 way you can fool a machine. He he actually did 0:16:45.160 --> 0:16:49.320 fool Deep Blue into sacrificing a piece that it shouldn't have, 0:16:49.520 --> 0:16:51.760 and that was how he managed to win one of 0:16:51.800 --> 0:16:56.479 the games in the round. So computers do make mistakes, 0:16:57.160 --> 0:17:01.600 but um, it's it, you know, it's it's hard to 0:17:02.120 --> 0:17:04.720 it's hard to predict when that's gonna happen. They don't 0:17:04.760 --> 0:17:08.240 behave the same way humans do. So um and Kasprov 0:17:08.280 --> 0:17:12.879 actually has said that now champion chess players are starting 0:17:12.920 --> 0:17:17.200 to adopt more computer like approaches to playing and has 0:17:17.240 --> 0:17:22.840 even uh talked about he's he's kind of a champion 0:17:22.880 --> 0:17:27.280 to a concept called advanced chess, which involves a player 0:17:28.000 --> 0:17:32.919 consulting a computer during play. So it's so it's a 0:17:32.920 --> 0:17:36.960 combination of human intuition and the ability of a computer 0:17:37.119 --> 0:17:40.800 to have that entire database of every game that's ever 0:17:40.840 --> 0:17:45.080 been played, essentially, and to rely on that and see like, oh, well, 0:17:45.080 --> 0:17:47.920 what what is this series of opening moves? Is that? 0:17:48.200 --> 0:17:50.840 Is that something that's been played before? And if so, 0:17:50.960 --> 0:17:53.560 what's the best thing for me to do? And you know, 0:17:53.760 --> 0:17:55.919 there are points where the human takes over and starts 0:17:55.960 --> 0:17:58.159 to make moves, or perhaps the human has a move 0:17:58.240 --> 0:18:00.240 in mind and programs it into the computer, or to 0:18:00.280 --> 0:18:04.320 see what possible counter moves could happen. Um, And it's 0:18:04.400 --> 0:18:06.200 it's kind of it in a way. You might think, well, 0:18:06.240 --> 0:18:08.560 that's sort of cheating, but in another way, you might think, Hey, 0:18:08.600 --> 0:18:11.920 this is an example of computers and human intelligence merging 0:18:11.960 --> 0:18:15.680 in a way. Yeah, yeah, I see what you're saying. Yeah, 0:18:15.720 --> 0:18:17.399 that's one of the things that I have a problem 0:18:17.400 --> 0:18:22.480 with with certain games is the lack of encyclopedic knowledge 0:18:22.840 --> 0:18:27.439 of everything you could possibly do. And in reading your article, 0:18:28.080 --> 0:18:31.240 I actually was thinking about this from a personal standpoint 0:18:31.240 --> 0:18:36.080 before I got to the next example. But yeah, I um, 0:18:36.119 --> 0:18:38.840 I'm a fan of the Scrabble brand cross word game. 0:18:39.600 --> 0:18:43.840 I love their trademark. We have to use that trademark, um. 0:18:43.920 --> 0:18:45.439 Although I love the game, so it doesn't bother me 0:18:45.480 --> 0:18:49.479 so much as quirky. UM. So the thing is, if 0:18:49.480 --> 0:18:53.440 you go to a lot of the uh, the clubs 0:18:53.600 --> 0:18:58.280 that really get into the game, UM, you will find 0:18:58.400 --> 0:19:01.399 that a lot of them offer lit and lists and 0:19:01.520 --> 0:19:05.080 lists of words, all kinds of words, lists of words 0:19:05.080 --> 0:19:08.159 that you can make with different particular letter combinations that 0:19:08.240 --> 0:19:11.720 you might have on your tile rack. And I keep thinking, Man, 0:19:11.800 --> 0:19:14.520 if I had access to all these words without having 0:19:14.560 --> 0:19:17.639 to study these lists, I might, you know, really be 0:19:17.720 --> 0:19:20.480 able to wamp up on some people playing scrabble. Now, 0:19:20.560 --> 0:19:23.160 there are people who spend lots and lots and lots 0:19:23.200 --> 0:19:25.399 of time studying the word lists, so they will be 0:19:25.480 --> 0:19:31.800 armed with all these different lexicographical weapons. Um. Yeah, I 0:19:31.800 --> 0:19:33.760 don't know if lexicography is a word. I'll have to 0:19:33.800 --> 0:19:40.440 look it up. I'm not sure that would work in scrabble. Um. Okay, However, Um, 0:19:40.480 --> 0:19:44.640 it seems that, uh that in your next example, that 0:19:45.280 --> 0:19:49.800 there is a computer able to play at that level, 0:19:49.840 --> 0:19:52.160 and that seems like it would have a severe pose, 0:19:52.200 --> 0:19:56.239 a severe disadvantage for the human opponent. Yeah. Quackle is 0:19:56.320 --> 0:20:01.040 the the program um that you're referring to calls a 0:20:01.040 --> 0:20:04.320 computer program that um that was able to beat a 0:20:04.440 --> 0:20:08.760 champion named David Boys in a series of games of 0:20:08.840 --> 0:20:12.560 scrabble and Quackle. Again. Scrabble is one of those games 0:20:12.600 --> 0:20:15.960 that also comes down in part to luck, because it 0:20:16.000 --> 0:20:19.040 all depends on what tiles you have in your you know, 0:20:19.280 --> 0:20:23.119 you have at your disposal. Uh. And Quackle was able 0:20:23.160 --> 0:20:26.600 to beat David Boys fair and square. Right, It was 0:20:26.640 --> 0:20:29.040 able to fair and square in the sense that it 0:20:29.080 --> 0:20:32.800 wasn't it wasn't making guesswork of the game. It was 0:20:33.320 --> 0:20:36.080 building words based upon what it had available and what 0:20:36.119 --> 0:20:39.000 was on the board already. Um. And yes, it did 0:20:39.040 --> 0:20:41.919 have an enormous database of words. So that gave it 0:20:42.080 --> 0:20:45.440 an advantage in that, you know, human beings, some of 0:20:45.520 --> 0:20:48.760 us are really good at remembering you know, thousands of 0:20:48.800 --> 0:20:53.000 different word letter combinations that make legitimate words. Others we 0:20:53.080 --> 0:20:55.200 rack our brains like I've got a V, A, B, 0:20:55.520 --> 0:20:57.960 and F and in a queue, and oh what can 0:20:58.000 --> 0:21:03.160 I make? You know? So that one was playing pretty 0:21:03.240 --> 0:21:05.359 much by the rules. What was interesting to me was 0:21:06.160 --> 0:21:10.760 I ran across a story about a graduate student named 0:21:10.760 --> 0:21:14.800 Mark Richards who came up with a program, a scrabble 0:21:14.840 --> 0:21:17.080 program that goes a step further. It doesn't just have 0:21:17.280 --> 0:21:20.800 a massive database of words that can play uh. And 0:21:21.359 --> 0:21:25.800 another advantage that scrabble UH computer programs have is that 0:21:26.080 --> 0:21:29.080 let's say that you have two lines of like two 0:21:29.119 --> 0:21:32.000 words that are fairly close to one another on the 0:21:32.040 --> 0:21:35.280 scrabble board, um, and but you know they're separated. There's 0:21:35.320 --> 0:21:39.800 maybe like five blank tiles between the two words. These 0:21:39.800 --> 0:21:43.119 computer programs are much better at looking at those those 0:21:43.200 --> 0:21:47.520 configurations and determining words that can span those gaps that 0:21:47.560 --> 0:21:51.760 are than humans are. And so you'll see words played 0:21:51.800 --> 0:21:54.439 by computers that most humans never would have thought of. 0:21:54.520 --> 0:21:57.760 Because to to have taken that into account, that big 0:21:57.800 --> 0:22:00.880 gap into account is just kind of beyond what most 0:22:00.920 --> 0:22:03.560 of us do when we play. There are good players 0:22:03.560 --> 0:22:05.720 out there who are really good at this, but most 0:22:05.760 --> 0:22:07.199 of us, you know, well, we sit there and we 0:22:07.240 --> 0:22:09.600 concentrate on one letter at a time, or if words 0:22:09.600 --> 0:22:11.080 are close enough, we might be able to get two 0:22:11.160 --> 0:22:14.080 letters that are already on the board incorporated into whatever 0:22:14.119 --> 0:22:16.720 we're about to lay down. Well, what Mark Richards did 0:22:16.760 --> 0:22:19.480 was he on a step further than that. He created 0:22:19.520 --> 0:22:25.080 a program that could guess what letters the opponent had 0:22:25.119 --> 0:22:29.159 at his or her disposal bye bye. You know, there 0:22:29.200 --> 0:22:32.200 are only so many letters that are in a scrabble game, 0:22:32.520 --> 0:22:36.119 so it's counting tiles, counting cards. By counting cards in Vegas, 0:22:36.400 --> 0:22:39.320 it would count tiles. So you would start playing the game, 0:22:39.359 --> 0:22:41.359 and at the beginning of the game, the computer really 0:22:41.400 --> 0:22:44.680 can't tell what you have because it's only the only 0:22:44.760 --> 0:22:46.840 information it has at the very beginning of the game 0:22:46.960 --> 0:22:50.520 is which tiles are in it's uh, it's vault right. 0:22:50.640 --> 0:22:52.960 Anything beyond that it doesn't know. So I mean it 0:22:53.000 --> 0:22:56.119 would know like, okay, well one of the q you 0:22:56.840 --> 0:23:00.760 tiles is in uh in my hand, which means that 0:23:00.800 --> 0:23:06.560 there is one less one fewer out in the actual game, right, Yeah, 0:23:07.080 --> 0:23:09.119 you can't really say, well, what are the odds that 0:23:09.160 --> 0:23:12.159 he's holding that? You know, it's it's astronomical. You know, 0:23:12.200 --> 0:23:14.760 there's no qu and scrabble, right is not qu I 0:23:14.800 --> 0:23:17.159 thought of what it's just Q some of the some 0:23:17.200 --> 0:23:21.400 of the others that are similar, because almost every instance 0:23:21.480 --> 0:23:23.600 that Q appears in the English language is followed by 0:23:23.640 --> 0:23:26.640 the letter you. Um, that's right. So at any rate, 0:23:26.720 --> 0:23:30.360 the so Q, I'm just gonna say, somebody's gonna write 0:23:30.359 --> 0:23:32.359 in so well, I don't play a lot of scrabble, clearly, 0:23:32.720 --> 0:23:36.360 but at any rate, so the computer what it can 0:23:36.400 --> 0:23:38.000 do is as the game goes on, it can start 0:23:38.040 --> 0:23:42.760 predicting with better and better accuracy, which tiles you probably 0:23:42.760 --> 0:23:45.400 hold in your in your hand. So what it what 0:23:45.440 --> 0:23:48.040 it does, It will start playing words and playing parts 0:23:48.040 --> 0:23:51.240 of the board that will block off the best options 0:23:51.400 --> 0:23:55.359 you would have and when it becomes your turn, so 0:23:55.400 --> 0:23:58.639 it's blocking you from the combinations that would get you 0:23:58.640 --> 0:24:02.359 the most points. So so you're you're handicapped even further 0:24:02.480 --> 0:24:05.080 than you were just from playing the game fair and 0:24:05.119 --> 0:24:12.359 square even. You know, another interesting thing, and another interesting 0:24:12.400 --> 0:24:16.119 point you made that was nice um in the Uh. 0:24:16.840 --> 0:24:20.120