Hacker News Reader

Comment by rectang | Hacker News Reader

Blitz online chess is one of my vices. To encourage wild games, I play the Fried Fox as black and the Wandering King as white, sacrificing soundness for excitement. Both involve moving the King's bishop pawn ahead one, and then moving the King into that spot.

It probably chops 100 points off of my rating, but who cares? It gets away from book lines and thwarts boring drawskis.

Should enough people migrate to chess960 that opponents could be found readily, I might not have to resort to such silliness.

View on HN

Chess master and bullet addict here. It sounds like you're as into avoiding book lines as I am. May I recommend my secret opening weapon: the Hippopotamus. It is quite possibly one opening to rule them all:

- play it with both sides

- guaranteed to get people out of book

- successfully deployed to defeat none other than super GM Etienne Bacrot (in blitz, but still): http://www.chessgames.com/perl/chessgame?gid=1479768

[edit] it's not a particularly sharp/aggressive opening. Tends to be more slow positional grinds. But not always - there are lots of possibilities for breaking things wide open with center pawn breaks.

How did you become a chess master? Can you tell me more of this hippopotamus?

This is going to sound really egotistical, but honestly I don’t remember much about becoming a master as I was 10 at the time. What I do remember is playing a shitload of games, rated games most weekends. And analyzing my losses with a soviet chess coach to identify and work on my blind spots. Rinse and repeat. Playing experience way more valuable than book learning.

I’ve been working on some training software and will hopefully be releasing a beta version on HN this summer, pending wrangling a bit of funding and/or a few devs willing to work for equity/glory.

I recommend checking out the mighty hippo’s Wikipedia page for some decent commentary and sample games.

It doesn't sound egotistical. It just seems like you had good parents and you put in a lot of hard work around a difficult subject and it paid off.

I wish I had put in more time as a kid for chess and that there was more STEM materials for computing and robotics. The 90's was a wasteland between the time of the BBC Micro/Commodore era and what we now have today with the Raspberry Pi and the maker movement. Windows 95 just wasn't very good for a kid with no idea where to start.

Is chess training software not a saturated market? I'm not super familiar with this to be honest.

Haha, no, like most scholastic chess champions I had a super dysfunctional home environment. One parent definitely supported/pushed my chess, and I give credit there where it’s due, but I could only apply the label “good” to either of my parents with some major qualifiers.

Never too late to start training! Neuroplasticity is a thing.

Chess training software is super not saturated. The dominant program, ChessBase, basically hasn’t changed in 20 years. Still no Mac version or decent web version if I’m not mistaken.

There’s a lot of cool things happening in live play and steaming right now with LiChess, which is phenomenal, and chess.com is starting to up its game as well.

But serious training software is stuck in the 90s, hence opportunity. Not that there’s any money in chess, but that’s another topic.

Well I hope everything is going well for you now in life! I didn't mean to imply your past was ideal and am sorry if it came out that way.

Thanks for the background on chess training programs. I had a friend with Houdini or some other engine maybe 10 years ago and it had a pretty extensive training mode or so I thought. It was fully 3D and could explain every move and suggest better ones...etc. Houdini is pretty expensive iirc (edit: just looked it up and it is ~$100 range), so maybe something more reasonable would satisfy a need? Also, I suppose the training mode was not as fully featured as real training software? I can't tell if this is the same "chessbase" thing you're talking about.

I’m up for helping, I have a load of ideas in this area...

Awesome! I just emailed you.

That opening is amazing! I’ve never seen it before. Super aggressive and very aesthetic!

Yes on the aesthetics. More of a counterpuncher choice than overtly aggressive IMO. Usually lines don’t open up until move 10-20.

As an aside, cool to meet the person behind ironworker. I used your service many years ago to launch my first startup. Helped me with some cron job shenanigans that otherwise would have caused me a lot of grief. Nice work!

Is that not simply hypermodern school with the fianchettoed bishops?

Yes, it’s like the ultimate expression of hypermodern.

Other pro tip: if you do the full hippo with all the fixings (double fianchetto, pawns e6 d6 h6 a6) it also allows you to delay castling until you know which side your opponent commits to. This can lead to an awesome attack that comes out of nowhere and completely blindsides the hippo-unaware.

For clarity, is the attack you're referencing a castle to the opposite side of the opponent and then using your two rooks on the opponents side to push down the pawns?

Kind of. Waiting to commit works both ways:

1. If they castle one side, you can do the other and then launch an attack, usually it’s the pawns that clear the way and open lines, then rooks and minor pieces can get in the action. (As opposed to using rooks to push down pawns, if I’m reading your comment correctly)

2. You can also see if they are preparing an attack on one side, usually watching pawn structure or where they’re putting rooks. And then you can castle on the opposite side to effectively neutralize the attack.

Note that it’s definitely possible to launch an attack on the side you are castled on, especially if the center has a closed pawn structure - it’s just more risky and thus less common.

Why didn’y Bacrot just play 31. Bg7!! And win the game?

That's discussed in the game comments.

I don't understand.

Below the game record are some comments. One of them asks "why not 31. Bg7?" and some discussion ensues. One commenter points out a line beginning with 31. ... Ke7 and ending up with W gaining the exchange but B having several extra (passed) pawns. (I haven't attempted to verify the accuracy of any of these comments.)

What percentage of your blitz games end in draws? I just checked my stats and over my hundreds of blitz games only about 5% of them end in draws. I checked the stats of some friends who are both better and worse than me and all of them were around 5% draws.

I fully appreciate the idea of wanting to play more aggressive lines to get a different style of game, but I don't think it is necessary by any stretch if you are looking to get decisive games.

2.45% of my blitz games end in draws.

Playing a weak, non-standard opening is surely not the only way to get interesting games, but it's reliable and easy for someone of my skill level (around 1850 on blitz.chessbase.com). It's much more reliable than playing any well-known gambit, for example.

This article starts off with an appreciation of the King's Gambit. However, there are many known lines which counter the King's gambit and bring it back into a boring, drawish middle game.

Chess is unfortunately a game that rewards people who play to avoid losing rather than play to win.

Tournament chess could change this style by counting wins as more than twice a draw (say 1.5) but where a loss is still 0.

Your idea is normally implemented as football scoring (so 0 points for a loss, 1 point for a draw, 3 points for a win). It has been tried (with limited success if I remember correctly) in top tournament chess. On mobile and reluctant to try and track down the details of when and where.

The Bilbao Chess Masters [0] used a 3-1-0 scoring format.

[0] https://en.wikipedia.org/wiki/Bilbao_Chess_Masters_Final

Me too. I used to play the Polish opening for white a bunch, but against anyone who can withstand unusual positions, it ends up in a boring midgame. My strategy now is to just play the same 2 openings every game, so that I get to an interesting midgame as quickly as possible with as much time as possible. fwiw my draw rate is 3.5%

Also, no shortage of games on lichess.com (their horrible UI update notwithstanding).

If you like crazier games that avoid book lines, I highly recommend the chess-like abstract board game Onitama [0].

It’s a smaller chess grid with each player getting four pawns and a king, and each player has two random cards that specify the move that all their pieces can make. If you take a move using one of the cards however, you exchange it with a neutral center card, so next turn your opponent will get that card after their move. It makes it very very hard to see more than a few moves ahead and the random cards basically ensure every game plays like some novel endgame neither player has ever seen.

[0] https://boardgamegeek.com/boardgame/160477/onitama

> you like crazier games that avoid book lines, I highly recommend the chess-like abstract board game Onitama [0].

Or just play Go.

>Or just play Go.

Or just play a game where the pieces themselves change the rules of the game (magic: the gathering comes to mind).

Both Chess and Go suffer from the problem that a "correct" play can be computed and both games will eventually be solved (and yes, Go has many more possibilities than chess of course).

To be a real game worth playing for me, a game needs to have randomness and hidden information as well as some way of altering the rules of the game itself mid-game.

I’m also an avid Magic player haha. Onitama scratches a similar itch for me as opening a hand and solving it relative to the boardstate. Lots of predicting opponent moves / options and leaning into lines of play intuitively, since the cards are randomized each game.

Me too! One thing I am noticing is that because I often just use it to destress, I am not improving really, because it's all instinctive on 3 minutes.

The problem with improving seems to be that while I can play a few blitz games, taking time to work on game is more a separate occupation. I occasionally do the lichess practice thingies but without seeing much impact. Anything that can be done that helps but that is on the timescale of a few blitz games?

You have to be good at slow chess to be good at fast chess, because slow chess builds that intuition you then cash out on in time trouble.

If you play on blitz.chessbase.com, the position is continually evaluated -- they won't suggest moves while the game is live, but they will tell you who has the better position and by roughly how much.

Whenever there's a position where you don't see the reported advantage, or when a change in the assessment occurs that you didn't expect, that's an opportunity to go back afterwards and just look at those specific points in the game to see what the engine suggests.

If you really want to get good of course you need to be much more methodical in your study. But I started to find chess much more enjoyable once I left behind any long term ambitions.

Interesting feature. I figure it still provides something of a hint if the estimate is updated immediately—e.g. if your opponent plays a move and suddenly you're several points ahead of where you were, then that's strong evidence that your opponent has let you win a piece, and you should look for a tactical combination to do that.

Which isn't necessarily a bad thing, it just changes the game. If we embrace this rule, then I humbly suggest that a large change in evaluation (due to a blunder) be accompanied by an "Oooooooh!" sound effect; that's what semi-restrained human observers might do.

Play through a grandmaster’s game with some insightful commentary? Bobby Fischer’s favourite games for example.

Many years ago I was playing in Rossolimo's chess studio in Greenwich Village, and he pitted me against a master, who played this variation 1 f3 … 2 kf2. It took me three or four games to be able to figure out how to use the tempi well enough to get a successful attack. After that Rossolimo allowed me to play him. After achieving what felt like an opening advantage, I managed to blow the game.

Why not play on lichess.org

Reply on news.ycombinator.com