The Turing Technique: How I Learned to Get Better At Chess and Its Variants

The Turing Technique is summarized as follows: the fastest way to learn a game that a computer has mastered… is to play against the computer. And memorize as many positions, moves, and patterns as you can. To be clear, this is not just about memorizing the beginning moves (also called opening theory by chess players), but also about improving anything in the middlegame and endgame.

This technique was life-changing because it helped end a gaming addiction consuming nearly every moment of my day. And I got it from the famous author Nassim Taleb (thank you, eternally grateful). Other mental techniques I used to solve the addiction were affirmations and gradually replacing addiction with a meaningful hobby.

I’m calling this the Turing Technique because this learning hack makes heavy use of the computer as a shortcut to mastery, and Turing is famous for pioneering computer science.

The Turing Technique is a rather bold claim. However, I speak from deep personal experience. This increased my crazyhouse (a popular chess-like game) rating by 217 Elo points (from 2233 to 2450) in only 4 hours of work for 5 days, or 20 hours total. And it finally helped end my crazyhouse addiction that I was desperate to end.

I’m certain the Turing Technique is the best way to learn chess — and I am certain beyond the shadow of a doubt that the Turing Technique is highly effective for simple chess-like games like crazyhouse due to the personal experience that I write about in this essay.

Some people are immediately going to question whether I actually know what I’m talking about. Before crazyhouse, I’ve played chess for over 10 years and quit at National Master as I personally no longer cared about “chasing rating points.” Way back in the day when we were growing up, I’ve played International Master (IM) Eric Rosen and IM Kassa Korley and won both games (to be fair, it was a LONG time ago now) — I mention because it seems they’re both popular streamers now (I sometimes wonder if I went down that path…). And I’ve played Marc Esserman at Harvard Square, where he gave me an extra Queen and I still lost horribly (I was already a couple of years retired from chess though 😅) — and then later he tried to sell me his Smith-Morra book. I know way more about the game than I probably should.

Skeptics are going to say the Turing Technique won’t work with chess, as the game is too large “to be memorized”. At the end, I have a full section of rebuttals addressing concerns such as these.

First, Thank you Nicholas Nassim Taleb

First, let me give credit where credit is due: Nassim Taleb deserves 80–90% of the credit for the Turing Technique. He might not remember these two tweets, but it totally changed my life completely (by putting an end to my gaming addiction).

My point is that if you want to practice skiing, you learn disproportionally more from hard runs (tail events) than from slopes for a beginner. Just as you gain more from lifting extremes (from convexity) than regular weights.

You learn disproportionally more from tails of distributions than you do from their center.

https://twitter.com/nntaleb/status/1060384417024217088

The Epiphany

You see it’s not a far jump from Taleb’s tweets to the Turing Technique. My first thought was: “What is the chess equivalent of a hard ski slope? Is it playing against Grandmasters?”. A moment’s pause. No…it’s playing against computers! What could be tougher than playing against computers? And it struck me that this must be correct. It must be the fastest learning method.

It’s like solving Rubik’s cube, which has thousands of combinations, but you only need to memorize a few key patterns to take you from start to finish.

How did this change my life? At the time I was extremely addicted to a chess variant called crazyhouse, spending 13+ hours a day playing the game. I had just successfully paused playing the game using mindset hacks, which I write about here and here. Then…I read Taleb’s tweet…commencing a funny “Angel vs Demon” dialogue inside my head:

• Devil: “I know I want to stop playing this game, but…….what if I tried one more thing…”.
• Angel: “That’s literally what every addict says. Do you really want to end up playing 13+ hours a day again? You just managed to stop playing, and now you want to jump back in?”
• Devil: “But I think this time it’s different!”
• Angel: “Aren’t those the most dangerous words in finance? This time it’s different?”
• Devil: “But this time I think it’s really different!”
• Angel: “Oh really? How is it different?”
• Devil: “It really is different! I’ve never tried learning by playing against computers, and only computers…”
• Angel: “OK, I guess you have a point. And if this learning technique doesn’t work? What’s the worst case scenario?”
• Devil: “Then I guess it’s going to be 40+ hours of my life gone while trying to knock the addiction off again…”
• Angel: “Sounds like you know the devil’s bargain. Your choice.”
• Devil: “I just have to try it…Taleb’s tweet makes too much sense. I think this could really work…”

Taleb’s idea just made too much sense-the signal-to-noise ratio of learning from the computer must be very high…

The Turing Technique in Action

…so of course, I went ahead and tried the Turing Technique.

The plan was simple.

The “test procedure”: How to test my knowledge

1. Go to lichess.org and play crazyhouse against the lichess Stockfish computer
2. Allow myself as many take backs as I wanted.
3. Lose horribly despite all the take backs.
4. Resign the game and open the analysis board, which shows computer analysis. Find all the mistakes I made. Use curiosity to explore anything interesting.
5. Go to Step 1. Repeat.
6. Occasionally switch it up by playing taking the other side: play against the line I was trying to learn.

The “learn procedure”: how to learn

1. Open lichess analysis board, turn on Stockfish computer.

2. Play out opening and all moves I was interested in analyzing.

3. Then to test the new opening knowledge, run the “test procedure”.

   

   Then finally, the “real world test procedure”: play a few rated games against humans-but only a few…majority of time should be spent with the computer. This is to get real-world feedback on what to learn next. If a particular line was weak, then you know where to spend more time with the computer.

Results? End of the week report.

It was very consistent progress. Every single day I would put in 4 hours of training against the computer, then a few games against humans, and I would gain 50 points. By the end of the week, only 5 days later, my rating went from 2233 to 2450. That’s when I knew, this time in my gut, this visceral feeling…that I was done with crazyhouse. I was really done. My addiction was finally over. Because…if mastery means playing as close to a computer as possible, then I don’t want to play the game. Let me do something that computers cannot master. Let me be human.

Incidentally, my “Let me be human” desire resonates extremely similar to this clip I found about Bobby Fischer hating chess…

Why Bobby Fischer Hated Chess:

This section was a last-minute addition to the essay, but I just had to add it because it fit so perfectly. I just watched this great clip of Bobby Fischer hating chess, and it ties in so perfectly with this essay. Why does he hate chess? Because of computers. Good quotes from the clip:

• “I hate chess because I know what chess is all about. It’s about memorization, it’s all about pre-arrangement, creativity is lower down on the list!”.
• NOTE on “pre-arrangement”: a terrible word choice by Fischer that implies conspiracy and collusion by players. This is NOT what he means. He’s referring to the grandmasters analyzing and memorizing the crap out of the first 20–30 moves in chess. His complaint is players aren’t finding moves over the board, but just playing from memory. A better word choice that chess people will understand would have been “extremely excessive opening theory”. Here’s another guy’s explanation of the same thing.
• “Chess just the last few years has changed dramatically with all this computer stuff”
• “But why would you want to get involved with something that is mainly rote and pre-arrangement?”
• Nakamura comments: “When he talks about chess and how it’s getting harder and harder, that’s true of every game…he should be angry at computers, not angry at the game”.
• I think Nakamura misses the point — I don’t think Fischer is angry about “chess getting harder”, but that it’s no fun playing a game that’s largely based on rote memorization and massive opening theory.
• Tangential point: this is why he spent a lot of time trying to revive chess with FischerRandom, which throws away all the opening theory by randomizing the starting position.
• And lastly, there’s this great Youtube comment that resonates so much with me that I’ll quote it here:
• “I quit chess at 2200 for this reason. I “looked ahead” at FMs IMs GMs and saw what they did. They just memorize. They know all the endgame principles, are very good at calculation and objective evaluation, but ultimately… all that separates 2200 from 2500 is how many hours he spent on his engine memorizing a line in the Sicilian. Chess at the beginner level is its most pure form of game. Wild sacs, fun mates and silly opening traps, kids smiling, and old people sipping their coffee. Thats what chess is about, friends and community.. not the hyper-competitive memorization and brute force analysis of tournaments today.”

There IS one thing I’d like to comment on Fischer Random, since Bobby Fischer thinks it’ll bring back chess: the optimal way to learn Fischer Random is still ……playing against the computer! This time it’s not about memorizing a lot of the beginning moves but about recognizing patterns really well. Still, it feels so lame that the best learning strategy is to bang your head against the computer. Why not do something that only humans can do?

Rebuttals to potential arguments by skeptics

But how are you supposed to learn basic chess concepts by playing on the computer?

Fair point. I’d say you still need to know the “big ideas” or “mental models” of the chess game in order to hang all the learnings on those hooks. However, the middlegame concepts can easily be covered in a book, while your specific opening patterns can be covered in a book as well. The rest of the time…?

• How would you learn chess tactics? By learning how the computer crushes you with those tactics
• How would you learn chess openings? By learning how the computer crushes you in those openings
• How would you learn positional play? By learning how the computer crushes you in positional play.

Seeing a pattern?

You can still learn those good moves by learning from your losses against the computer.

If you really believe this is the best way to learn chess, why don’t you become a chess grandmaster (GM)

For the same reason, Bobby Fischer hated chess, dear reader. Why get involved in something that’s largely based on rote and memorization and opening theory? Scroll up to see the quotes from Bobby Fischer on chess.

And besides, I already proved for myself that the Turing Technique drastically speeds up learning in crazyhouse. That’s good enough for me. And I really don’t want to be like a computer. Let me be human.

Secondly, I’m sure out of the thousands of people who’ve read this blog post, there’ll be one person who takes it to heart. That person will succeed and write a blog post going viral. You can let me know, and I’ll write the blog post publicizing it for you.

But chess is much more complex than crazyhouse. You can’t learn chess just by playing on a computer.

Readers who are chess fans will have disbelief that the same rapid growth of 200 Elo points in such a short time (~20 hours over a week) could be achieved in a game as big and complex as chess.

Let me preface this by saying I’ve spent many years of my childhood playing chess. I was only a few points away from the National Master title, when I purposely quit when I realized the futility and pointlessness of these chess titles and rating points. I quit in protest and rebellion of the whole darned system.

Let me also say that I agree chess is a much bigger, complex game than crazyhouse. Chess has much more to memorize. Crazyhouse barely has an opening and no possibility of an endgame.

However, I firmly believe the Turing Technique applies to chess as well. Just wait a moment: you realize grandmasters are famed for having insane recall of literally thousands of games? They can recall games played decades ago. There was one video that went mega-viral testing Magnus Carlsen’s memory.

Now, why wouldn’t you expect the GMs to just memorize against the computer? It’s a miracle of the computing revolution that even cheap pocket phone computers have more processing power than the famous Deep Blue computer that beat Kasparov.

I would advise an aspiring chess player to learn from a book the “mental models” of chess such as pawn structure. But all the ideas can be easily covered in a single book (How to Reassess Your Chess by Silman was one book I often read in my childhood). Then, spend a ton of time with the computer, tying specific examples to the mental models.

It’s not going to be as quick as crazyhouse…it won’t take only 20 hours and 2 weeks of work. It might take 6–12 months to gain 200 Elo points. But the Turing Technique is without a doubt the fastest possible path.

Think about it. Do you really think Magnus Carlsen (or whoever the World Champ is) is not using a computer? Do you really think none of the grandmasters are using computers to analyze the crap out of everything? I remember Marc Esserman spending his entire Harvard winter break analyzing the hell out of the Smith-Morra, which is how he became the expert on that variation (By the way, I have actually met and played Marc Esserman at Harvard Square! Pretty chill dude 😊). Why has the age of “world’s youngest GM” been steadily declining, now to only 9 years old?

Rhetorical questions. The answers are obvious.

A related point: I believe one of the best things an aspiring chess player can do is to take a couple of well-known opening books and just memorize the crap out of it. One of the most common (bad) advice is advising beginners to focus on the middle game principles instead of studying openings. I had the most rapid growth (200 Elo points easily) in my “chess career” simply by studying the heck out of 3 chess opening books, every single line, and position. Usually, middle game performance improves from understanding the opening so well.

Why not play a human then analyze that game with a computer?

Some people might say the Turing Technique doesn’t sound that different from playing a game with a human and then analyzing it with a computer.

Au contraire. Couldn’t be more different.

Forget humans. You should literally play against computers, and mainly computers. Why? The signal-to-noise ratio is far higher. Even the slightest mistake is punished and highly visible. And it is easy to explore many variations simply out of curiosity, with the computer showing you the best moves every step of the way.

The main benefit of playing against computers is being able to focus on exactly the problem area. It’s focused training. Playing against random humans on the internet…well you can easily go a whole hour until you get challenged in that one line you have trouble with. Playing against computers, you get heavily challenged, 10 times in a row, in a very short time.

Why not study Grandmaster Games?

This is sort of the same idea as the Turing Technique, in that the goal is to study what the good moves and patterns are. However, the major missing weakness is the Grandmaster is not there to answer what would happen if a different move was played.

If you’re going to study GM games, better have computer analysis nearby so that you are able to explore questions easily. It’s a mistake to study GM games in isolation otherwise.

There is no crazyhouse book. Hence no one has any idea what the “best moves” are. You can try copying the high-rated players, but what happens if you get taken off-variation…?

How to Learn Chess Openings?

A combination of an opening book and Turing Technique. One book per opening is probably fine. The opening book is mainly to learn the “mental models”. Then test and expand your opening knowledge fast and heavy with the Turing Technique. Play against the computer with both colors on the same opening.

Final Thoughts: Dear Younger Self…

I’m writing this as if I were writing a guide to my younger self.

Several years ago, I was severely addicted to crazyhouse and desperately wanted to shake off the addiction.

Here’s what I’d say to my younger self several years ago: “Son. It’s simple. Just use the Turing Technique. Just play against the computer and memorize everything. That’s all there is to it. You can go do something else now. Don’t be addicted.” Followed by a wholesome hug 🥰

Don’t believe me that this Turing Technique works? Why don’t you try it out for 10–20 hours and come discuss it with me on Twitter.

I also have a theory on how to master first-person shooter games like Halo rapidly. However, I haven’t tested it yet. If it works, I’ll write about it 😊.

Random Thoughts I Couldn’t Fit Into Essay

• The corresponding “hard ski slope” for language learning would be going to a bar in France and attempting to converse using French. This thought is also inspired by something Nassim Taleb said in a tweet, which I am too lazy to dig up at the moment (readers, help?).
• Computers are significantly cheaper and more widespread, leading to ever-younger chess GMs from even countries with lower median income.

(NOTE: If you disagree with me, I’ll be willing to listen on Twitter 😊: @SpeedTestDemon).

Also published on https://speedtestdemon.com/turing-technique-fast-learning-trick-fo-chess-like-games-mastered-by-computers-written-by-a-fully-recovered-chess-addict/.