*** The Chess Thread ***

[OngBonga]

They've been using conditional moves on gameknot for years, it can irritate opponents which is good in a psychological game like chess. Whenever I face a conditional move I always take my time to respond. If it's in the opening or if my move is obvious or forced, it's easier to dismiss, but if they've basically predicted a move that wasn't obvious, I'd be a bit worried and would definitely take more time than if they just made their move as normal, so in some cases it might work against you. You can alert your opponent to imminent danger.

[Poopadoop]

If you really want to mess with your opponent's head, use half your clock on your first move.

[OngBonga]

I'm presuming the conditional moves are used in long chess, that is 1 day per move or longer, rather than classical chess or blitz. In short chess, they use premoves instead. It's basically the same except it has more strategic value, especially if both players are under extreme time pressure. Throwing a random "bad" premove in can often be met with a premove that anticipated a different move, and can result in a huge blunder.

I don't really like fast chess. I want to play the highest quality game I can, and that happens with proper deep analysis instead of quickfire mental calculation.

[OngBonga]

The difference between a conditional move and a premove is a conditional move is "if he plays x then play y" while a premove is "play x unless it's illegal". A subtle difference but they're not quite the same. Conditional moves don't really have a strategic value beyond perhaps irritating your opponent a little, while premoves can be used as last resort blunder hunters and save time.

[CoccoBill]

Right yeah, those are 2 different things. I see some crazy people queuing up 3-4 moves in blitz games, insanity.

Just cracked 1400 btw, and if anything my winrate is up. Been playing people rated 1300-1500 the past weeks and my winrate is 82%. I've managed to pay more attention and avoid the majority of stupid blunders.

[MadMojoMonkey]

Right yeah, those are 2 different things. I see some crazy people queuing up 3-4 moves in blitz games, insanity.

Just cracked 1400 btw, and if anything my winrate is up. Been playing people rated 1300-1500 the past weeks and my winrate is 82%. I've managed to pay more attention and avoid the majority of stupid blunders.

[OngBonga]

Nice work, keep going!

[Poopadoop]

Just cracked 1400 btw.

At this rate, in a few weeks it will be time for a Cocco-Ong death match. Or a chess game. Either one I want to see.

[OngBonga]

I'm pretty sure he'd give me a challenging blitz game already.

Long chess, no chance.

[CoccoBill]

Still haven't played a single blitz or rapid, I think I'd have a panic attack on turn 4. I still think I'd fair better in daily, but have no illusions of winning. 1500s already play goot, no one is falling for opening gambits anymore most of the time, and they often know the best responses to a bit more obscure openings too.

On that note, I've been dabbling with nimzo-larsen lately and liking it.

[OngBonga]

You're starting to run into people who are using an opening database. That's the next level and will force you to employ solid opening theory instead of hoping for a smash and grab with some gambit. I would hope you're using a database yourself, it's legal in long chess and if you're not using one then this is the kind of rating where you'll start to hit a brick wall.

[CoccoBill]

I've been using the chess.com opening database but it sucks, only includes master-level long games. Lichess has a far better one but most of the time I can't be arsed to use it, too much work to move the pieces to the correct position there, on chess.com while playing it's just available right there. Based on my stats (the paid subscription on chess.com gives pretty neat stats and analysis of played games) the openings are my strong part, endgames the weak part. I should probably read a book on endgame strategy or something.

[OngBonga]

Puzzle solving will help your endgame, more so those with fewer pieces. That's probably the easiest thing you can do, it's not even a chore.

Understanding zugzwang and being able to recognise when it's necessary to win, that will help too. Many endgames are won by forcing your opponent to make a bad move. The K+R vs K is an example, if the defending king could just refuse to move it would be fine but it has to move and that's why it loses. That's zugzwang, literally "compelled to move". Most balanced king and pawn endings that have a decisive result are won by zugzwang, and it's extremely satisfying to find.

Don't waste your time studying unlikely endings like the mating pattern for K+N+B vs K because you're very unlikely to ever get to such an endgame (I haven't). It's better to study balanced positions or ones where you're just a pawn ahead.

Queen and pawn endgames are ludicrously complicated and you're probably best leaving that well alone, just go by instinct when you get there are trust that your opponent is unlikely to be an expert in such positions either. Usually, what's technically a draw will offer winning opportunities for both, even at high level.

[CoccoBill]

The biggest hole in my endgames is probably simply deciding what to do. Should I hunt down his remaining pieces, defend against their pawns or attack with mine. Obviously there are clear situations but I find myself often not having a clear plan. If I have a passed pawn and my opponent doesn't, well there's no question what to do. Even situations though, it's one thing to not find the right move towards your goal, and an entirely different one not having a goal. I guess attack is the best defense. Tactics I'm usually ok with.

[CoccoBill]

Had to brag. Just 14 moves with the nimzo-larsen, managed to pin his Q to his K in 14 moves, he resigned. I'm starting to feel far too invincible again, I need a reality check. Too bad chess.com doesn't allow you to specify the opponent's rating.

[OngBonga]

Should I hunt down his remaining pieces, defend against their pawns or attack with mine.

I mean, all of the above. There's always an optimal move, and that might be an attacking move, a defending move or even a waiting move. I guess the easiest way to approach "balanced" endgames is to try to control more space, that will usually mean a superior position, so moves that either increase your control of the board or restrict your opponent's are what we want to play.

If the game is truly balanced, a theoretical draw, then our ideal moves are ones where our opponent has few good replies and many bad replies, increasing the chances of him blundering. Of course we also need to ensure our position is solid so we're not giving him winning chances, that's your priority in theoretically drawn endgame, but we do want to be thinking about making it as difficult as possible for our opponent to find the right moves to secure his half point. If you think you have a better position in a balanced endgame then avoid trades (edit - unnecessary trades, if it wins a pawn then go for it) and try to control more space, if you're under pressure in balanced endgames then seek to trade off pieces and blockade dangerous pawns or create a fortress for your king.

That's general advice and won't always apply, chess is wonderful in that it has its quirks. Sometimes logical assumptions in chess (controlling more space is good etc), they fail us. But it's rare.

[CoccoBill]

Update after 200 games. Winrate is going down, people at 1500+ aren't pushovers. Still feel there's some more elo to be gained.

[OngBonga]

Best win 1642, not bad. That average opponent rating is a bit sad though, you need to be getting that closer to your rating.

[CoccoBill]

Yeah the avg is gonna fix itself the more I play, I try to only play people rated higher than me. Then again I've also played like 30 games with a friend who's 600.

[OngBonga]

Well if that average is a legacy of your fast rating climb then it's not really an issue, it just looks like you're player weaker opponents which means you're not challenging yourself, but if you're usually playing people around your rating then your av opp rating will obviously increase.

You could play your friend unrated, but it's not like it matters what your av opp rating is, it's just a quick guide as to whether you're challenging yourself sufficiently.

[CoccoBill]

Chess.com doesn't allow you to specify a rating range which I hate, just very limited options related to your own. I use "-25 - +200".

[OngBonga]

Yeah that opp av then is very misleading and likely heavily weighted towards earlier games, or skewed by your games vs your friend. You might be able to adjust the time period that average relates to, which will be useful.

[CoccoBill]

Yeah you can adjust the stats period, for example the past 30 days my avg opponent rating is 1365.

[OngBonga]

That's a lot more like it. If that's also skewed due to your games vs 600 buddy then that's a very healthy average.

[OngBonga]

[wrong thread, can't be bothered to repost, nothing important nothing to see here gotta go to work bye]

[Poopadoop]

Can you get a very high rating by seal-clubbing alone? I mean your rating goes up every time you win right, just by less if you're beating someone with a much weaker rating.

[CoccoBill]

You can't, for example the games I play with my friend, for me a loss is like -14, a win 0. Not much incentive to play rated games with him.

[OngBonga]

I think the limit for rating gains is a -200 ELO difference, if you're playing someone with a larger rating gap they're basically freerolling you. Not 100% sure what parameters chess.com use but it's likely +-200 with a K-factor of 10 (how many points you win vs equal rated opponent).

[OngBonga]

You can't, for example the games I play with my friend, for me a loss is like -14, a win 0. Not much incentive to play rated games with him.

I'd expect your rating loss to be 20 if your gain is 0, that's the usual extreme. If it is 14, maybe their K-factor is 7, but that seems low.

[OngBonga]

To answer poop's question, you can inflate your rating a little if you play -150 rated players and basically perform slightly above expected over a large sample, but it will be tediously slow progress compared to just playing people closer to your rating.

You're also going to slowly run into better players, and if you're not challenging yourself and improving, your rating gains will decrease in value and ultimately start receding.

There's really not much incentive to club seals in chess, other than to massage your ego if that's your thing.

[OngBonga]

From what I can tell, expected win with 200 ELO difference is 76% for the stronger player, so if you're winning 80% of games against -200 players your rating will gradually rise. But as your rating rises, so too does the floor of an opponent at -200, so it's a self regulating system.

[CoccoBill]

I'd expect your rating loss to be 20 if your gain is 0, that's the usual extreme. If it is 14, maybe their K-factor is 7, but that seems low.

Against equally rated it's indeed +7 so checks out.

[Poopadoop]

I was just thinking it would be funny if someone became a GM by beating 10,000 players rated as <1000.

So if Cocco's average opp is 1042 how is he still climbing? I guess some of them he gains points on but a lot of them must be freerolling him as you say.

[CoccoBill]

I've gained 648 in rating in the past 90 days, 212 in the past 30. The opp avg just lags behind. I'm only playing people rated the same or higher than me (excluding the games against friends).

[CoccoBill]

Assume I'll play the next 50 games only against players rated 3000. After that my avg opp rating would be 1234, and my own rating would be 2247 if I won all games.

[Poopadoop]

Ok, let's see it.

[Poopadoop]

What does your rating start out as? Say you're playing your first game, are you at 1000 or summat?

[OngBonga]

Against equally rated it's indeed +7 so checks out.

I'm surprised it's this low, I'm almost certain it's K=10 on gameknot where I used to play a lot (haven't player a proper game for a while now), and figured that was the online standard. A low K-factor means slower rating response to form, and less extremes in rating fluctuations, so some would say it's preferable. But you don't want it too slow, so it's not as simple as the lower the better. 10 is unlikely to be optimal though, that's just nice round human numbers in action. Maybe chess.com are getting this right with 7.

[OngBonga]

What does your rating start out as? Say you're playing your first game, are you at 1000 or summat?

It's complicated. Again I don't know chess.com's specific parameters (and surprise surprise I can't be arsed to check) but on gameknot the first 10 games, you get a provisional rating that can fluctuate wildly in the range of like 800 to 2500 or whatever depending on the rating of your opponents and your results. It's basically designed so master level chess players don't have to grind their way up.

Gameknot's rating floor, I believe, is 800. Each player also has their own floor, which is -200 and then rounded down to a multiple of 100 from your rating peak, so my rating floor is 1700 since my peak is 1969. That means I can never drop below that rating, no matter how many games I lose.

[Poopadoop]

It's complicated. Again I don't know chess.com's specific parameters (and surprise surprise I can't be arsed to check) but on gameknot the first 10 games, you get a provisional rating that can fluctuate wildly in the range of like 800 to 2500 or whatever depending on the rating of your opponents and your results. It's basically designed so master level chess players don't have to grind their way up.

Gameknot's rating floor, I believe, is 800. Each player also has their own floor, which is -200 and then rounded down to a multiple of 100 from your rating peak, so my rating floor is 1700 since my peak is 1969. That means I can never drop below that rating, no matter how many games I lose.

Makes sense. What they probably do is start you at an "average" rating (whatever that is, let's say 1500), and the K-rating is very high for game 1, then gradually comes down to whatever is standard.

[OngBonga]

I think that "average" is 1200 on gameknot, but yeah before you actually play a game you do have a rating.

[CoccoBill]

On chess.com when you start you're at 800 but yeah it's not that simple. Say you start by playing some blitzes and win, then when you play your first daily you're probably gonna be 1200 or even 1600 to start out. I've played several people with an initial rating of 1600, where they have had hundreds of rapids or blitzes played but no dailies.

[Poopadoop]

On chess.com when you start you're at 800 but yeah it's not that simple.

So basically they assume you're a moron until you prove otherwise. lol.

Say you start by playing some blitzes and win, then when you play your first daily you're probably gonna be 1200 or even 1600 to start out. I've played several people with an initial rating of 1600, where they have had hundreds of rapids or blitzes played but no dailies.

Don't you have the same rating regardless of what types of games you're playing? I mean if it's all straight chess but with different time limits. Not including badugi or bukkake or whatever that weird variant is called.

[CoccoBill]

So basically they assume you're a moron until you prove otherwise. lol.

That's usually a good default setting in life.

Don't you have the same rating regardless of what types of games you're playing? I mean if it's all straight chess but with different time limits. Not including badugi or bukkake or whatever that weird variant is called.

Nope, separate ratings for all formats.

[OngBonga]

It's definitely correct to have different ratings for different time formats, although you don't have a different rating for literally every time format you can think of (potentially infinite!).

Formats are a bit arbitrary and might vary from site to site, but as a rough guide...

There's classical (standard) chess, online that's like a day+ per move, and over-the-board it's multiple hours, perhaps two days, for the entire game, then there's rapid, which is like 15 mins to maybe an hour for the game, there's blitz which is the 5-15 minute range, and then sub 5 minutes is bullet, which is chess for crack heads, fun to watch but insane to play.

They require very different skills and strategies, so it doesn't make sense to have a global rating for all formats.

[OngBonga]

I mean to achieve a very high rating in bullet, you don't necessarily need to be a master level chess player, you just need a solid foundation, know some tricks (opening traps, premove traps, etc), and be very fast at calculating and moving pieces. For classical, the lack of serious time pressure allows for more complex calculation and therefore a higher quality game, requiring expert level opening knowledge, positional play, tactical play and endgame knowledge to succeed.

[MadMojoMonkey]

That's usually a good default setting in life.

[CoccoBill]

A quick update, cracked 1600 some days ago. Playing against some Icelandic IM now, and haven't been embarrassed yet! He had the initial 1600 rating so was paired against me, dude's between 2200-2500 on the quickies.

[OngBonga]

1600+ is moving into "respectable" rather than 1500+ which I'd call "above average". Next up is 1700+ which is "very respectable".

1400+ "average", 1300+ "casual", 1200+ "hobby", 1100+ "weak", 1000+ "shit", 1000- "give up".

That should be an official ranking system.

[Poopadoop]

1600+ is moving into "respectable" rather than 1500+ which I'd call "above average". Next up is 1700+ which is "very respectable".

1400+ "average", 1300+ "casual", 1200+ "hobby", 1100+ "weak", 1000+ "shit", 1000- "give up".

That should be an official ranking system.

I like this better than the usual system, "grandmaster" sounds too much like "grand wizard."

[MadMojoMonkey]

1000- "give up".

*waves*

[CoccoBill]

I actually won the game, he just resigned. Playing him felt exactly like playing a bot. Really strong play coupled with some really erratic moves. He blundered a minor piece fairly early, by midgame he had evened it out, and a couple weak moves from him to start the endgame gave me the edge. I used passed pawns to get a solid material advantage, in the end I had a rook and a bishop pair against his rook and knight. Apparently he pretty much only plays bullet and blitz.

[CoccoBill]

[CoccoBill]

I've been stuck at low 1700s for a while, some chess fatigue going on I guess. I wanna break 2000 though.

[OngBonga]

I've never broken 2000, I've got close but never reached that target. I'll be very surprised if you succeed in that goal, but good luck, certainly not impossible. You'll need exceptional opening knowledge (or be using an opening database) and consistently make very few middlegame mistakes. Breaking 2000 is extremely difficult, at least it was when I last tried.