WEBVTT 00:01.170 --> 00:02.130 Hi there. 00:02.160 --> 00:07.830 In this introduction, actually, I just want to go over the course index and conventions generally. 00:08.220 --> 00:15.390 So I talk about ingredients because I like those Master Chef people on TV telling you about these herbs 00:15.390 --> 00:17.880 they're adding, which you may never have heard of. 00:17.910 --> 00:24.690 These ingredients, these kind of secret ingredients that often differentiate, like the master chefs 00:24.690 --> 00:28.170 from lesser masters of cooking. 00:28.320 --> 00:33.540 So in chess, yeah, they're all attacking ingredients to be aware of. 00:33.690 --> 00:41.550 And it's good to refine how we evaluate positions and maybe leverage some of these ingredients like 00:41.550 --> 00:43.050 Bishop without a counterpart. 00:43.050 --> 00:46.150 When you've got a certain bishop without a counterpart. 00:46.260 --> 00:49.350 It can be dangerous or form pawn. 00:49.950 --> 00:57.210 So there's a lot of concepts in this course which you may get a chance to use in your own games. 00:58.290 --> 01:04.230 So we're also improving our evaluation of positions, not just our calculation, but by having these 01:04.230 --> 01:13.710 kind of tags, ingredients, you call them patterns, call them what you want, but they're items of 01:13.710 --> 01:17.220 interest, which can affect one's evaluation of positions. 01:17.490 --> 01:20.610 And so makes things more powerful. 01:20.610 --> 01:26.070 When we calculate, we can look at the resulting positions, especially at the end of our calculations, 01:26.070 --> 01:32.070 and say, Oh, well, we've got better our accounts by Oh, well, there's a form pawn, Oh well we've 01:32.070 --> 01:34.170 got this, which is an attacking ingredients. 01:34.170 --> 01:40.950 These signposts for attacking ingredients improve our evaluation of positions and evaluation is very, 01:40.950 --> 01:43.980 very important because we can't calculate that many moves ahead. 01:43.980 --> 01:49.040 We need to rely heavily, quite often on intuitive evaluation. 01:49.050 --> 01:55.320 So we're trying to get these ingredients, you know, how did our clinical staff get these fantastic 01:55.320 --> 01:55.740 positions? 01:55.740 --> 02:00.900 What were the ingredients which want to show examples for each potential ingredients? 02:01.830 --> 02:09.270 And the early sections of the course cover general aspects of king safety, like the Castled King or 02:09.270 --> 02:11.730 the weakened King or opposite side castling. 02:12.420 --> 02:17.040 So those kind of conditions for where the king is are very, very important. 02:17.040 --> 02:22.050 But then we go on to more sophisticated ingredients, which you may never have heard of before. 02:22.110 --> 02:24.480 When I use seeds, that's a shortcut. 02:24.480 --> 02:27.720 I want to give you maximum space for showing the theme of the game. 02:27.720 --> 02:33.450 Seeds means the number of collections a game is in at Chess Gamescom, which is one of my favorite sites 02:33.450 --> 02:35.190 for checking out all the games. 02:35.190 --> 02:40.330 And often I find the more collections the game is then, the more iconic it is, the more celebrated 02:40.330 --> 02:40.860 it is. 02:40.860 --> 02:45.420 And often it's kind of really instructive and compelling and dramatic. 02:45.600 --> 02:46.740 People love drama. 02:46.740 --> 02:47.720 I love drama. 02:47.730 --> 02:52.890 I'm sure you love drama, but it's good if drama is also instructive. 02:52.890 --> 02:55.140 It's good if drama is nutritious. 02:55.140 --> 02:59.130 That's a great aspect of attacking chess. 02:59.130 --> 03:06.600 It can be dramatic, but also instructive, especially if we take time to find and ask the questions. 03:06.600 --> 03:10.050 Or how did this attacking position exactly happen? 03:11.130 --> 03:17.540 So basically, yeah, I want you to kind of take in these patterns, these tags, these collections. 03:17.550 --> 03:24.210 The key is so the insights labels, they that are available in this course, hopefully they'll start 03:24.210 --> 03:30.630 flowing more easily in your games and more useful to establish the attacking ingredients. 03:30.630 --> 03:35.580 And the game usually has multiple attacking tags within it. 03:35.580 --> 03:41.790 It might not just be one advantage, it might be several advantages like stronger center of protection 03:41.790 --> 03:42.890 of E5 Bishop. 03:43.470 --> 03:46.680 It might be several ingredients all working together. 03:46.680 --> 03:52.050 And you know, when I was studying a lot of neural network games, especially, you know, alphazero 03:52.590 --> 03:57.750 kind of a game changer, it did show us well for me especially showed us a lot of key elements which 03:57.750 --> 04:03.580 people were not really talking about in the past literature like Form Pawns or Bishop Mountain counterparts. 04:03.600 --> 04:10.890 I've tried to introduce these and emphasize these over protection was from Nimzowitsch, though some 04:10.890 --> 04:14.130 are from Nimzowitsch in famous Nimzowitsch. 04:14.130 --> 04:17.850 My system is a classic book, but there's some fun. 04:17.850 --> 04:22.970 Like themes like a Queen in Siberia is like saying if the opponent's queen is away from their kings, 04:23.130 --> 04:24.150 it's a way of saying that. 04:24.240 --> 04:32.940 So we want to basically improve our attacking evaluation of positions, our sense of the attacking potential 04:33.030 --> 04:34.290 of situations. 04:34.290 --> 04:37.890 That's really important in chess, because checkmate ends the game. 04:38.160 --> 04:39.660 It's not all about material. 04:39.660 --> 04:44.040 It's not a game just for the accounts and chess about accumulating material. 04:44.040 --> 04:47.340 You know, we can we can checkmate from being several pieces down. 04:47.340 --> 04:47.850 That's the key. 04:47.850 --> 04:50.460 Beautiful thing I find about chess. 04:51.240 --> 04:55.860 The reason why attacking patterns from neural networks are very, very interesting is they're also kind 04:55.860 --> 04:59.670 of verified they were beating some of the much more analytical engine. 05:00.500 --> 05:06.700 The classic AB engines, alpha beta engines, which were analyzing often a lot more positions. 05:06.710 --> 05:12.830 We want to have that ideal where we don't really have to calculate that much and wear ourselves out, 05:12.980 --> 05:15.920 and we've got a greater instinct for attacking gradients. 05:15.920 --> 05:17.270 That's really great. 05:17.300 --> 05:19.880 It means we can calculate less sometimes. 05:21.320 --> 05:28.970 We also can say, Well, it's just all about evaluation of the core, the core elements, the four core 05:28.970 --> 05:32.330 elements that many test websites allude to. 05:33.170 --> 05:40.760 King Safety versus material, pawn structure versus piece activity four So we can often sacrifice material 05:40.760 --> 05:46.760 if we can reduce the opponents king safety or we can often sacrifice the structure of our pawns or even 05:46.760 --> 05:53.780 sacrifice pawns generally in order to maximize our piece activity, because pawns are the fundamental 05:53.780 --> 05:58.190 constraint of the pieces, especially if you read Michael Stephen's, you know, simple chess. 05:58.190 --> 05:59.750 That's one of the things he said. 05:59.750 --> 06:06.710 So the evaluation aspect, sure, you've got the main core elements of positions, but these other elements 06:06.710 --> 06:11.480 alluded to here are more specific and worth thinking about in my view. 06:11.690 --> 06:18.830 And I personally make use of these attacking themes in my games continually, like every day online. 06:18.830 --> 06:23.300 And I just love attacking Chancellor, I love winning with these attacking ingredients. 06:23.300 --> 06:28.910 So it's like sharing the attacking ingredients and also trying to build up my own attacking game as 06:28.910 --> 06:29.420 well. 06:29.420 --> 06:31.180 So this course is a work in progress. 06:31.210 --> 06:35.900 If there's any other important attacking things that crop up, sure, I'll create a new section I'll 06:35.900 --> 06:42.170 put examples in to reinforce, because this is all about our evaluation, our instincts being built 06:42.170 --> 06:49.010 up in this course for these themes, these tags, these patterns which create the ingredients, the 06:49.010 --> 06:55.250 exciting ingredients we're looking for to get those winning attacking combinations, especially over 06:55.250 --> 06:58.400 the opponent's king, not just to win material. 06:58.400 --> 07:00.290 So okay, so the course. 07:00.290 --> 07:00.580 Yeah. 07:00.590 --> 07:06.350 Is basically celebrating the ingredients of attacking chess. 07:06.380 --> 07:10.850 And I hope a lot of those ingredients you'll be starting to use in your own games. 07:11.000 --> 07:11.900 Okay. 07:11.930 --> 07:15.590 Here is an example game from the world of neural networks. 07:15.590 --> 07:21.350 Komodo was playing with 64 bit Comodo four cores. 07:21.380 --> 07:26.810 This is on a fast time limit, but four engines, you know, they can they can crunch so many positions. 07:26.840 --> 07:31.880 This was 5 minutes per game with 10 seconds per move increments. 07:32.030 --> 07:36.410 So not not such a quick time control but from this position. 07:36.410 --> 07:43.160 Yeah, Lela with the black pieces played a brilliant foresighted sacrifice here. 07:43.550 --> 07:45.500 So can you guess what that is? 07:46.780 --> 07:54.580 So the leading neural network calculating far less than its alpha beta engine classic engine counterparts 07:54.580 --> 07:57.190 played nine times G two. 07:57.220 --> 08:01.340 So this has all sorts of attacking implications. 08:01.360 --> 08:02.800 Let's just see the game. 08:02.800 --> 08:04.930 KING Tanks, G2, Bishop Africa. 08:05.830 --> 08:13.090 King G1 and having this huge confidence that this attack is is working is justified. 08:14.080 --> 08:18.400 So rook D1 rook f6, you can see the resources are flowing to the opponent's king. 08:18.400 --> 08:23.800 That weaken king position does imply that actually there's a lot of responsibilities and roles that 08:23.800 --> 08:26.980 can be taken up with these attacking pieces. 08:26.980 --> 08:31.810 So we have H for Rook H six King if on the king tries to flee. 08:31.840 --> 08:35.260 But now actually there's an element here left over. 08:35.260 --> 08:37.480 There's a there's an outside pass pawn. 08:37.480 --> 08:43.030 So even though the king is not being mated, that's not always an outcome sometimes which is needed. 08:43.030 --> 08:45.220 We don't even need to make the opponent's king. 08:45.220 --> 08:46.810 The team is on full blast. 08:46.810 --> 08:48.850 Every piece is working hard here. 08:48.950 --> 08:57.640 Now the h pawn is coming along and the IVB engine is struggling here as leader is improving and improving 08:57.640 --> 08:58.720 the position. 08:58.840 --> 09:05.740 So making way for the pawn here and conceding F2 now, which goes into this nasty pen. 09:06.130 --> 09:12.550 And so the pressure's building up the pass pawns Optimus now C5 Queen G friend the game actually ended 09:12.550 --> 09:20.140 here so what an amazing you know winning attack basically which created that winning outside pawn if 09:20.140 --> 09:28.270 King C as an example, then black is queen ing and then bishop takes E2 and the attack rages on here. 09:28.270 --> 09:34.240 So this would be absolutely a winning position, as example, where black is going to be material up 09:34.420 --> 09:36.820 with still a vicious attack. 09:37.060 --> 09:42.610 So yeah, this is just an example to show, you know, it'll be hopeless here given this is the case 09:42.610 --> 09:45.190 and this is time for desperation. 09:45.490 --> 09:52.900 So, yes, these intuitive sacrifices were shown in neural networks, and neural networks are essentially 09:52.900 --> 09:57.610 quite often it's about pattern recognition, pattern matching. 09:57.610 --> 10:02.590 And for me, yeah, there were great source of attacking themes and general themes in chess. 10:02.590 --> 10:09.640 So they for me are a kind of game changer because if you consider the number of calculations was far 10:09.640 --> 10:11.890 less than the traditional engines. 10:12.460 --> 10:17.950 So it's less about playing the position like a supercomputer, but more about intuition, intuitive 10:17.950 --> 10:21.400 factors that we can kind of identify. 10:22.300 --> 10:29.500 And I've made use of form pawns, for example, in so many games since studying neural network games. 10:29.590 --> 10:33.010 I just found them incredibly instructive. 10:33.010 --> 10:41.740 So sometimes if you're following an example of a model player or a computer, if the computer is not 10:41.740 --> 10:46.390 really computing that much, it could actually be instructive, like in the case of neural networks. 10:46.390 --> 10:49.810 Now, in this course, we're lucky to have great technology. 10:49.810 --> 10:57.310 This course in the latest Stockfish does incorporate neural network aspects so we can learn all aspects 10:57.310 --> 10:59.200 of the game even from the opening. 10:59.200 --> 11:05.890 And that's important because often the ingredients for attacks are born out of the opening advantages. 11:05.890 --> 11:12.280 So it's important to make notes and be very nuanced in our analysis of most games to see how advantage 11:12.280 --> 11:16.630 is accumulated, what advantage is accumulated, etc.. 11:16.630 --> 11:22.390 So yeah, we can learn a lot from the neural networks and this is just one game example, just to show 11:22.390 --> 11:27.160 you that sometimes even if we don't checkmate an outside pass, Paul might be winning. 11:27.160 --> 11:35.920 So okay, so Layla and Alphazero also potential attacking heroes to learn a lot from in the engine world, 11:35.920 --> 11:41.920 as well as all the amazing attacking players of our rich and cherished chess history and culture. 11:42.220 --> 11:43.030 Okay. 11:43.620 --> 11:44.380 And so much.