Did chess.com get harder suddenly?

Brilliants are sacrifices that are good or best and can be influenced by rating.

My game seek is set to -400 and ∞ I was thinking maybe -25 and +200 would be better.

lol ok

I was wonderig the same thing , got off to a decent start ,won 4 games in a row when I started here last week, then ws up and down, but the last few days have been dreadful overall. Then again my mind has been very preoccupied with other things that have been stressing me out like crazy ! Perhaps some of the same things that might be stressing others out here as well !

Look at percentile instead of rating.  The mean rating has dropped from 1100 to below 800 in about 3 years.   I am now about 93rd percentile but with a rating of only 1360.  Years ago my rating was about 1450 but my percentile was below 90th.  I think this is due to new players coming in at lower starting ratings.  The system used to start at everyone at 1200, which was considered average back then.  The distribution has been skewed to the left, 

I recently dropped 250 points in a month, and then had to work my way back.  In my case, I had been doing a lot of the tactics trainer puzzles.  Good training for offense, but I realized (meaning I got schooled) about also playing defensively.  Once I started paying attention to defense, I stopped losing as many games.  That is just what happened to me - YMMV.

Players in the 1200-1500 range are definitely getting harder to beat, Evidence being that I just beat an 1800 bot quite quickly and comfortably, I did not used to be able to do this, so I have gotten better, but rated about the same.

The bot should not be rated 1800 but I guess thats another question.

Making the bots move in proportion to what their rating is, is kind of hard.

The lower ratings are easier. Makes weak moves, but doesn’t just move randomly if you blunder your pieces.

more average ratings are easy too. Plays stronger moves, punishes you for your mistakes, but doesn’t have a very high depth, and makes moves that are more progressive rather than being a static threat level to the player.

once you get into higher bot ratings though, it gets harder. They play a lot smarter, but it’s difficult to make sure that they aren’t smarter than what other players at that rating play. Since it varies and changes frequently, the bot has to purposely make moves that hurt itself, IF the player can see those planted mistakes.

My idea is that, all of the bots are adaptive, like Jimmy.

But the ones that have ratings are still adaptive, but stay in its relative rating area. So it can have a more realistic style of playing.

Interesting find!! coffee fast anyone!

Yeah I've noticed that too. When I first started chess I started with playing the Martin bot, and I absolutely destroyed him because he never threatened my pieces and never protected his pieces. Just recently I couldn't find an opponent so I decided to go with a bots speedrun, starting with martin. I played much better that game, but so did martin. He still blundered a few times and only took like two of my pawns, but it was much better than his previous performance. There us something I noticed though both times I played him: 

He doesn't take hanging pieces unless in check, meaning that if you hang your queen right in front of his pawn, he won't take it, most of the time, but if its also check, he MIGHT take it.

You should see the state of chess.com in 2015😵

There were FMs barely past 1900😱

By “my idea” I meant as in a suggestion they can add. The bots currently don’t have that (at least I don’t think they do).

jarrs123  WROTE

''Look at percentile instead of rating.  The mean rating has dropped from 1100 to below 800 in about 3 years.   I am now about 93rd percentile but with a rating of only 1360.  Years ago my rating was about 1450 but my percentile was below 90th.  I think this is due to new players coming in at lower starting ratings.  The system used to start at everyone at 1200, which was considered average back then.  The distribution has been skewed to the left, ''

I have been wondering about this for a while, and I am surprised that chess.com do not make a serious comment about this, as opposed to their key workers usual ironic and allegedly witty evasions

Why not just let the BOTs play some 3/0 blitz tournaments? Then we will know their "real" rating.

I also feel as though the search algorithm is choosing opponents differently. I'm a beginner, but I went straight from a 9 game winning streak to being unable to beat anyone I'm pitted against. Started a couple of weeks ago. 🤷‍♂️

That's my rapid ELO history here. Player fluctuation is real. I have long 1500 plateaux, but every once in a while I had a little crash into the mid-1400s. One noteworthy exemple in my case is when I stopped playing chess in 1 month, was a little bit rusty and lost 100 ELO points in a short amount of time, «reseting» into the 1450s.

In my personal case, it is cyclic, and the fluctuation matches my health and my brain focus. But I never go lower than the 1400s. I don't get worse : I notice I have a floor I can't cross even when I am at my worst, and I always get back into my natural 1500 habitat where I stay most of the time.

I also know that I will have a fluctuation one day (this year? next year?) or another that will get me in the low 1600s, as long as I work to improve my chess. Because 1500s and 1600s have ELO fluctuations like me and their human strengths vary everyday, and because I have already reached the 1550-1570 levels many times before.

As humans, we sometimes have an inconscious personal bias to compare ourselves with others as if we we not changing. But we need to understand that we get through fluctuations periodically, like other people we compare ourselves to.

Chess players also change slightly their playing behaviour as they play. You may have developped over time some habits that make your games harder for you.

The only way that would really work very well is if they played a sufficient number of games against members with established ratings and people didn't know they were bots.