LiChess: Learn from your mistakes | Hacker News Reader

LiChess: Learn from your mistakes

en.lichess.org

639 pointsakkartik113 comments

lichess is legitimately one of the most impressive web applications I've seen and used. I've never been a big fan of playing chess, but I discovered the site several months ago, and it allowed me to get into the game in a way that I thought I never would.

It started with the tactical puzzles. They're derived algorithmically (using Stockfish analysis) from games actually played on the website. This alone is fascinating:

https://en.lichess.org/blog/U4sjakQAAEAAhH9d/how-training-pu...

My favorite thing about the puzzles though is that I can easily explore suboptimal solutions using lichess's analysis tools. If I ever got a puzzle wrong, my first question wouldn't be what the right move was, but rather why my move was bad. This is trivial to answer with the analysis tools. You can even see the game that the puzzle was derived from and see what moves the players actually made.

The site is just a goldmine.

Very cool project, I was blown away by this feature.

Did stumble upon a bug though. I had three puzzles which I thought were buggy. I forgot the URL and details on the first one (so it may or may not been buggy), and I was flat out wrong about the second one (a bishop was making sure a horse couldn't move which I thought could be used to take the piece supposedly causing checkmate).

My third one though is clearly buggy: https://en.lichess.org/training/61461

Its not over when the game says victory. Black has 4 (! EDIT: 3, but still) ways to avoid the checkmate, and Stockfish also suggests these moves.

Nevertheless a very nice project, including Stockfish integration. I can't repeat that enough! App of the year for me.

I don't think the goal of that puzzle is to achieve checkmate. "Victory" just means you solved the puzzle, not that the game is over.

The goal is just to find the most optimal move given the board configuration. In that particular position, white can use its knight to simultaneously attack the black king and queen. Since the king is more important, it must be moved, effectively forcing black to sacrifice their queen. So the goal of that puzzle isn't to checkmate black, but to use a temporary check against black to capture the black queen (at the cost of a white knight -- still a worthwhile exchange, since the queen is much more powerful than a knight).

Yes, I was wrong, and I concur. I have since my post completed puzzles which do not have the goal to reach checkmate. For example, I had the goal of 'deflecting' an attack where I was in severe disadvantage. Its refreshing to have different goals, and forces the player to think outside of the box for such a goal thereby increasing the difficulty.

FWIW, I do not agree is (always) more powerful than a knight, but generally they are (while a queen is more powerful than a rook or bishop which is almost always better, a knight has a unique utility).

I didn't like there algorithmically derived puzzles at all since almost 80% of the ones I tried where variations of similar puzzle. I guess I should admire that it is algorithmic but even then I expected proper puzzle.

When did you last try them? They changed over to a new selection process and a new batch of puzzles a few weeks ago, and they're much more real-world now.

LiChess is one of the most amazing hosted platforms out there, it's very fast, it has a minimalistic design, it has no ads and does the thing that it does very well.

Furthermore they sympathetically state: "Lichess mobile is developed and translated to 80 languages by volunteers. Just like the website, it's 100% free forever, and there will never be advertisements. This is humanist software, made open source for the love of chess and user freedom."

But there are two things on my wishlist 1) make the website work off-line 2) new variant to change the chess rules to accommodate a smaller board 5x6, 4x5 or 6x6 for even shorter games.

>new variant to change the chess rules to accommodate a smaller board 5x6, 4x5 or 6x6 for even shorter games.

That makes no sense. You can have games as short as you want. Just play using a different time control. I play 1+2 (minute plus seconds) and those games last 2 to 3 minutes at most.

BTW with variants, you're not really playing chess anymore. You're playing a different game with chess pieces. Not all the strategy and tactics transfer well if at all.

Yes, but variants are supported in the application, anti-chess, horde, atomic, three-check etc. Agreed, those are not really chess, as the rules or end conditions are significantly changed. But it's like chess960 chess, the pieces are on a different position or there are less pieces and a smaller board. The reason for my wish is that I want to play a small chess game with my son before bed, it should be a bit less daunting then a complete chessboard and only use a few pieces, so it would be both quick and easy. A time constrained regular chess game would be quick but less easy or usable for a young kid.

Can't you just make this yourself? Why does it have to be a part of the lichess platform?

Yes, of course. I was just explaining what the lichess platform would need (and why) to be perfect in my humble opinion. It would be nice to use the same chess app for everything chess related, with the social aspect working offline like a progressive web app.

Why do any rule variants have to be a part of the lichess platform?

Hi Derrick, I gave my personal answer and lichess gives this answer: "Chess variants introduce variations of or new mechanics in regular Chess that gives it a unique, compelling, or sophisticated gameplay. Are you ready to think outside the box?"

https://en.lichess.org/variant

I don't like time control, I can't handle the stress behind that.

I saw someone else above mention an iOS app, and it looks like there also is an Android app.

Hi Tom, yes, but mobile apps don't make a website work offline on windows, linux or osx unless you are proposing the Android app can be run through Chrome APK welder (if it works). But even if that were the case, I think the current Android/iOS apps don't allow you to asynchronously play chess, but only offline against the AI. Scenario's would be having a spotty or on/off connection for example, or playing one turn/hour or day. Progressive web apps can work offline and sync the state back to the cloud when online.

After some exploring I came across a "good enough feature": correspondence chess. On the website and the mobile apps you can login en logout a game (including the variants) in between turns and set the window from 1-14 days per turn! The only thing is that you don't get notified when a person has taken their turn, except when you are on iOS i read. Well, you can always exchange chess notation through Riot, e-mail or snail-mail :-)

https://play.google.com/store/apps/details?id=org.lichess.mo...

If your interested in cool and unique features, try resetting your password:

https://en.lichess.org/password/reset

This is clever but it can't be a good captcha. Computers are really good at understanding a visual chessboard and even better at actually finding the best move.

I had the same thought, but on further reflection, while it's very doable to get a computer to figure it out, it still seems like quite a bit of work to have it read, parse, analyse and play the move for this specific page, on this one website for a low value payout. Additionally there is natural language question that has to be understood for what you are supposed to do - it doesn't necessarily have to be mate.

Since security is fundamentally increasing costs enough to outweigh the prize, I feel like this makes it.

There has been a whole thread concerning this https://news.ycombinator.com/item?id=5351734

lichess.org and the corresponding open source project[1] are really quite something. If you're at all interested in chess or code, I recommend following it -- their great execution will inspire you.

[1] https://github.com/ornicar/lila

This may sound naive, but what are the incentives and/or how are the developers of Lichess making a profit. The site is one of the best executed webapps I've ever used. What motivates the creators?

Thibault simply wants to deliver the best online chess experience. Period.

It's purely donation driven and he pulls a small salary from whatever's leftover after buying enough servers to keep the site up.

detailed cost breakdown

https://docs.google.com/spreadsheets/d/1CGgu-7aNxlZkjLl9l-Ol...

( link came from here https://en.lichess.org/patron )

I'm supprised that they get $47 mn in donations. Rarely see the patron icon when I play.

$47k, not million. Amounts in that spreadsheet aren't in thousands.

Apologies, thanks for clarifying

mn?

Has the site been around for long? I've never heard of it but it looks like a large-ish community.

It was launched in 2016, with a much simpler interface (no accounts iirc), which evolved a lot to come to this day.

Its source code is also really interesting, it makes me want to learn Scala.

> It was launched in 2016

It's been around at least as far back as 2010. Here is a 6 year old HN submission about it: https://news.ycombinator.com/item?id=1927188

Yeah, I borrowed some css for a chess board from it for a (never published) project I was working around 2010 or perhaps slightly earlier!

I should reread my posts before sending, this was a mixup between "6 years old" and the launch year. Thanks for the correction.

The Lichess author gave a tech talk last year where he discussed the design of the client and server.

https://www.infoq.com/fr/presentations/lichessorg-open-sourc...

My big problem with lichess is that you are allowed to pick white when you set up your game request and/or if someone starts a match with you and they have black pieces they can cancel game after you make your first move with no penalty. If you query the top players games in blitz or lightning their ratio of white/black games is ridiculous.

Maybe they should display win rates for black and white separately. Or display the midpoint between the two (i.e. counting each color as 50% even if the number of games is different).

I agree it's annoying. But at least if you use the new "Quick game" pairing system there is no option to choose a color.

So now it's just people who quit half the quick games they start with you?

There is anti-boosting logic that disables your account if you abandon too many games

I haven't encountered this problem personally. I think most people are just playing for fun not to get their number higher.

It's interesting that even the best chess engines are still worse at analyzing openings than humans.

They're also terrible at recognizing fortress positions. There are many such positions where decent human players immediately recognize a draw but an engine will give a non-zero evaluation. The engine can then be manipulated up to the 50 move rule whereby it might weaken its own position in order to avoid the draw.

Is that actually the case, or perhaps humans try a lot of sub-optimal opening lines?

The engine dislikes the King Gambit, which has a pretty lengthy history, dating back to the 1600s [1]. However, it's not played much anymore at very top levels, and much has been written about its weaknesses [2]. So in some sense, humans have been "trying" the King's Gambit for centuries and have only recently come to the conclusion that it's not optimal.

Many players enjoy using the opening simply because it leads to dangerous and exciting games, so it makes sense that Stockfish and other engines would prefer conservative opening lines.

[1] https://en.wikipedia.org/wiki/King's_Gambit

[2] http://chess.stackexchange.com/questions/94/why-is-the-kings...

So what happens when you force Stockfish to play itself in a Kings Gambit opening?

I have a GM friend who regularly practices against a computer with the opening repertoire turned off. One of his comments : "The bloody thing is absolutely clueless in the opening, and I still can't win. It's just so good at defending a worse position, and in the end, I will make a mistake."

I think the case is more that opening theory is very deep. A computer without an opening book is having to work out what humans have spent hundreds of years learning/memorizing. It may also be that since the eval function of an AI can skip opening/endgame with tables, it doesn't need to focus on some of those concepts as much in favor of being optimized for general midgame positions

Sub optimal from the perspective of the AI, some positions that seem easy for an AI looking 10 moves ahead might be tricky for your opponent.

Sorry, but this makes my head hurt.

>>> best chess engines are still worse at analyzing openings than humans.

>> Is that actually the case, or perhaps humans try a lot of sub-optimal opening lines?

> Sub optimal from the perspective of the AI

There's no perspective involved: optimal play is perfect play, assuming the opponent also plays perfectly. If a move is sub-optimal for White that means White has a better move, no matter what the opponent plays.

Sure, in practice a non-perfect-player might do better playing a sub-optimal move A rather than a better move B, because they don't know how to continue after B or because their non-perfect opponent ends up replying even more sub-optimally to A.

But just because I can manage to pull a fool's mate on someone that doesn't make it an optimal opening. One could say I might be "optimizing" for fun or for time, but if you redefine optimality arbitrarily then the original claim might as well be "chess engines are worse than humans at analyzing openings because they play better than humans".

PS: I'm not claiming that chess engines are better or worse than humans. And, incidentally, playing an inferior move while hoping the opponent makes a mistake is bad even if it works, as it really hinders your progress.

Of course there's perspective involved.

Restricting yourself to perfect opponents is very silly firstly, because there are no perfect opponents because chess is unsolved. So right out of the gate our definition assumes the counterfactual for all chess games every played and is both inaccurate (does not describe actual chess) and impractical (cannot be implemented)

Secondly, if there were such an opponent (which is not at all clear there ever will be), there is no way within the rules of chess to detect whether or not we face it in a given game. Imperfect opponents are allowed to play chess, and assuming they are all perfect while a defensible strategy is merely one of several.

Thirdly, it leads to very counterintuitive results. Consider the case in which chess is solved and well-balanced (e.g. a perfect opponent P can force a draw). We modify P to create opponent NP. NP behaves as follows: if NP's opponent is consistent with P, NP will implement P (e.g. force a draw), if NP's opponent deviates from P, NP will lose (forefeit or make bad moves). Therefore everyone can beat NP except "optimal" algorithm P, which can only force a draw. So now we have an "optimal" algorithm with a 0 tournament rating.

In reality the whole "perfect opponent" thing is mostly a notational convenience for Zermelo's Theorem. Which is very interesting and useful if you are writing a proof, less interesting if you are trying to win a chess tournament. A more practical definition of "optimal" is something like "wins a lot of games".

You're going off a tangent here. There's nothing in my argument that relies on having an optimal opponent only; forget about the "assuming the opponent also plays perfectly" part for a moment, the key part is getting the best move no matter what the opponent plays.

Exactly. In my opinion, chess at its core is all about presenting challenging puzzles to your opponent and solving the puzzles put to you by your opponent. An engine may show that one side is way ahead but it could all depend on finding one series of precise moves and therefore not be much of an advantage at all for a human under the pressure of the clock.

Isn't positional play all about recognizing that some positions are better by virtue of giving some soft advantages that will only be realized further down the line?

If your goal is to win, wouldn't you want to play the move that always leads to a win 10 moves ahead, rather than a move that looks immediately tricky but is actually escapable for the other player?

And given that you want to become a better chess player, isn't it better to start learning the best lines from the start?

If your goal is to win, you should play moves that lead to situations where you will do significantly better than your opponents, given what you know of their strengths and weaknesses. Whether they are good or bad from an 'optimal play' perspective doesn't matter.

For example, if chess is a win for white, you play black, and you have a halfway decent opponent, he will know what black's best defense looks like (that is: the one that takes him the most moves to win). Replaying that gives you zero chance of winning or drawing. Deviating from the beaten path cannot decrease that chance, and may increase it.

It is similar when chess, played optimally by both players, turns out to be a draw, and you want to win, but in that case, you may not want to give up your assurance of a draw, so you may play less weird moves. You still will want to play moves that lead to plays your opponent is weak at, though.

> If your goal is to win, wouldn't you want to play the move that always leads to a win 10 moves ahead

Sure, but the King's Gambit is just an opening, and you'd have to rely on lazy mistakes to guarantee a win 10 moves ahead that early in the game.

Right, but game theory tells us that for every single position, perfect play either leads to a win for one player or a draw, and for every position, there is a best reply.

To the extent that an AI approaches perfect play more closely than a novice or masters from the 19th century, shouldn't we go learn from the AI?

One interpretation of Stockfish here is 'I don't think this is a very strong opening'. To the extent that it's correct, stronger players will not play that opening very much, so spending time to learn it seems potentially wasteful.

Chess engines (I suppose that could be called AI) as well as databases of games have developed opening chess theory very significantly because of these reasons.

For perspective, Bobby Fischer created a version of chess which randomized the game because even in his day, he was annoyed that some players that he felt were far inferior at playing chess, could actually better memorize opening lines and enter the mid game with an advantage.

So the answer is that players very well have been learning from AI.

>To the extent that an AI approaches perfect play more closely than a novice or masters from the 19th century

AI has become better than even the top chess players of today.

Lichess is so awesome, it really changed my life. The features I use most: 1. The tactics trainer is incredible and fun. You get tactics ELO and the tactics stem from real games. Hence you can replay the game and see how this position arose. 2. The new tactics trainer from your own games is wonderful (but many other desktop programs had this before). 3. The analyse game function using stockfish is wonderful. Even continuously with a few lines + arrows in the game, you get immediately what you have missed. 4. The study section is awesome. I tried to build up a complete new opening repertoire with it in parallel to the leading chessbase tool. And guess what: It's free and a nice competitor to chessbase. All study pgns can surely be exported and imported to chessbase. 5. I really like the simul section. You find many dozen simuls a day. 25-50% of players are with 2000-2500 Elo. 6. And if you watch the lobby in the diagram mode you always find your right partner, right time, right chess variant. 7. And of course, the mobile version is also great. Keep on guys! We love you.

I'd like to have similar tool for the game of Go!

Crazy Stone deep learning has analysis mode where it identifies your mistakes and recommends alternatives:

http://www.unbalance.co.jp/igo/eng/

the game engine is strong, 5d running local on my phone, no server!

here for android: https://play.google.com/store/apps/details?id=jp.co.unbalanc...

i hear it exists for iOS too

I was thinking precisely the same. I love both games, but I'd love to see the equivalent for Go.

Many Faces of Go has some analysis tools that might be helpful: http://www.smart-games.com/manyfaces.html

I found it worth the price. And the lower kyu AI was actually fun to play.

I'm a big fan of lichess. Great UX and analysis features. The only issue I have with it is that disconnection auto-forfeits the game. Would love 30 seconds grace to reconnect (like they give you on chess.com).

The future grand masters who are 10 years old atm will be far in front of the current crop of grand masters due to improvements like this.

Gotta ask and I ain't trying to start a flame war here.... how's the community compared to say, ICC? I was a big time user of ICC back in the day simply because the average player was high caliber after FICS started to become a bit of a dead zone for standard players (I like 45 5 or higher timed classic games )

There are more high level players on ICC because that's where the high level players are. Chess.com also has a decent interface, but they particularly excel in chess news, event coverage and community support.

That said there are plenty of verified expert-level players there, whom are probably at least 1 standard deviation above and beyond satisfying 99.99% of the world population of people looking to play chess.^1

[1] - https://en.lichess.org/stat/rating/distribution/blitz

> There are more high level players on ICC because that's where the high level players are

It appears this is not the case, anymore. I logged in to both servers this afternoon, and counted the number of players rated over 2350. Lichess had 50, ICC had 6. I logged in again tonight and saw roughly the same ratio.

Those data points could be valid, but it's important to note that chess ratings work relative to the players involved. A site that only allowed pre-schoolers, for example, would still have 2300+ rated players.

WHAT ratings though? Ratings are not comparable across different player pools.

My Lichess & ICC ratings are typically within 50 points of each other. It's the same for a friend of mine who plays on both servers, too.

ICC is basically dead, even top players are leaving. Takes forever to get a game.

I'm not sure about the quality for standard, but the overall chat community is more like modern-day Reddit than 00s ICC. That place was a garden.

havent used this, but icc ftw!

The LiChess guys are just the best. The BEST!!!!

how does this compare to chess.com ? that site also has a post-game analyse feature which presents alternative lines

LiChess analysis is free and better. I don't know what the Chess.com paid analysis is like, but on LiChess you can just run Stockfish on your games and analyze different lines as you like. On Chess.com you have way less flexibility (at least with the free option).

LiChess also provides post-game analyse features, allowing Stockfish to determine alternative lines, and also learn from your own mistakes.

I never used chess.com, so I cannot really compare. I'm an avid user of the linked site, and use it play both with friends and users around the world. So far it's been a great experience, and I still have lots to discover : chess studies, tournaments and alternate game modes.

I've used both, lichess' free is better than chess.com's premium.

Most of the times I get this message when I open lichess.

Your connection is not private

Attackers might be trying to steal your information from en.lichess.org (for example, passwords, messages or credit cards). NET::ERR_CERT_AUTHORITY_INVALID

This resolves after 2-3 days and then within a week it again reappears.

That is suspicious. I play on lichess frequently and have never seen an invalid cert there.

Maybe an expired certificate? -- is your OS fully patched/updated?

check with your ISP/vpn, works fine for me

Ha how ironic that this was just mentioned in another thread! +1 to you sir!

Now with that said, this looks quite good and really broaches chess in a modern way in terms of learning. I sure have stacks of books that I wish were just software, mostly on studying chess tactics

I haven't tried it yet, but there's an app for "running" chess books, i.e. you can play the positions described in the book right in the app. No need to have a chess board. According to the site, there's a big list of books that have been converted to work for that app.

http://eplusbooks.com/

> I tested the new feature on my favorite opening: 1.e4 e5 2.f4!, a.k.a. the mighty King Gambit. Stockfish hates it, and asked me to review it like it was a mistake! Yet it's considered playable. Actually Carlsen played it against Arionian in 2015.

Not acceptable, Mr Stockfish!

Actually, the King's Gambit has been (essentially) solved: https://en.chessbase.com/post/rajlich-busting-the-king-s-gam...

There is only one move that doesn't lose by force after 2. ... exf4, and that is the strange-looking 3. Be2. Does that mean this opening is a "mistake"? Well, maybe not, obviously a human wouldn't know the perfect lines, but personally I wouldn't play it :)

Either way, I'll certainly be using this feature a lot. Kudos to the folks at Lichess.

The King's Gambit hasn't been solved nor is it likely to be solved any time soon. That article was one of Chessbase's annual April Fools posts. https://en.chessbase.com/post/the-chebase-april-fools-revisi...

I've been under the impression for years that the King's Gambit, despite it being the best variation/opening of all (IMO), was obsolete at the highest levels. A chess master/instructor once assured me of this, despite it being his favorite too. He referred to it as "romantic" and said it was almost never played at GM level.

It is rather rare at high levels because someone who knows the book lines will equalize fairly easily as black. That said, it isn't losing for white, nor is it particularly easy to play as black if you've never studied it in depth.

reminds me of https://www.chess.com/blog/eXecute/a-capablanca-story

That's an April Fool's Day prank. The King's Gambit has not been solved.

That article is an April Fool's joke: https://en.chessbase.com/post/the-chebase-april-fool-s-prank

> Actually, the King's Gambit has been (essentially) solved: https://en.chessbase.com/post/rajlich-busting-the-king-s-gam....

> There is only one move that doesn't lose by force after 2. ... exf4, and that is the strange-looking 3. Be2. Does that mean this opening is a "mistake"? Well, maybe not, obviously a human wouldn't know the perfect lines, but personally I wouldn't play it :)

It was an April Fools Prank. Apparently a good one.

What gives it away as a prank? I surely couldn't tell. Anything more than the reference to March 31?

> If at least two masters have played a move in tournament situation, then the move must be playable! So that's how it works now.

Considering how much Stockfish is stronger than those masters, those moves could be really wrong despite being playable in human vs human games.

http://www.computerchess.org.uk/ccrl/404/cgi/compare_engines...

Hoowever we don't know the estimated ELO of Stockfish as used by Lichess.

No, there are positions where computers are known to make the wrong decisions. Computers use opening books instead of evaluating the moves themselves because they don't evaluate openings well.

I wrote something similar a while ago. http://blunderchess.sf.net . Actually, the portion which broke down PGN games into FEN positions is what took a while.

I found a good way to improve myself is to play against myself in my head.

Can someone recommend excellent books and/or tutorials to learn chess? I know only basic rules, but no tactics/strategy/opening/etc.

For tactics I've never seen anything better than chess tempo [0], which is free, or for a small yearly subscription you can absolutely go nuts breaking down tactical themes and combinations of themes, and the tracking is very good. I did several thousand tactics puzzles there and it helped quite a bit.

For books, Beginners Mind by Jeremy Silman is pretty good, as is his endgame book. Logical Chess by Cherenev is also a great introduction to how skilled people think about games but takes a different approach than Silman.

[0] - http://chesstempo.com/

Tactics: https://en.lichess.org/training Video library: https://en.lichess.org/video

I found http://www.chesscademy.com pretty helpful. It takes a MOOC approach to chess, while keeping things fun and manageable.

I learned a lot from the tutorials by Joshua Waitzkin in Chessmaster: Grandmaster Edition. I heard that lichess has some puzzles as-well, but never tried them.

for books starting from scratch or nearly from scratch, I have enjoyed the "winning chess" series by Yasser Seirawan

I just poked around on this site. I think this is what I need to take my chess game to another level.

Can this only be used for games played on LiChess?

Or can I upload a PGN from, for example @fbchess?

It says it works on imported games.

I actually just signed up for this last week.

Nice to see lichess.org on here. I'm a huge fan of the iPhone app. I usually just try to solve the puzzles of the day to and from work while on the train. It has helped my chess game a good bit.

It can be rather addicting, also. On more than one occasion, my wife has gotten annoyed with me just sitting on the couch for hours trying to solve the puzzles :P

I'm not a chess player and I lose 9/10 games I play on lichess, yet I still occasionally go there when I don't feel like playing mahjong, partly because the website is so pleasant to use.

Reply on news.ycombinator.com