Did chess.com get harder suddenly?

I feel like people at the same rating have been playing better lately and I have been losing a lot of rating points the last few weeks. Am I just playing worse? It could be because I quit caffeine but it's still a pretty big difference.

rating is just a number my friend. 

It's a very accurate measurement of strength. The only way it can go down without me playing worse is that everyone else got better, or lots of people left chess.com (because it sucks) and there was a rating deflation.

It's an accurate measurement of past performance. All things being equal, it can be used as a way to gauge strength and potential outcomes in future games.

Everyone goes through fluctuations in play quality and that is a more likely explanation.

Variance is a very real thing that is inherently a part of the nature of the rating system. Human player fluctuations are the normal - if there were not unexpected ups and downs, then that alone is suspicion of someone using a chess engine. Everyone has ups and downs or they are not challenging themselves. 

In my experience, it usually takes several "ups" into a rating and "downs" out of it again before one stays there for a while. For example, let us say a hypothetical player is trying to maintain 1500 blitz level. They might take a long time to reach 1500, when they do, they may stay there for a week, crash to 1400. Take a month to build up to 1475, then drop to 1450 - then climb to 1520 and then drop out of 1500 range again etc. This process is somewhat mocking, but one gets a bit used to it. After a few of these fluctuations, you will eventually stay more reliably above 1500 (or whatever other rating is in question).

Just keep learning and keep working on your chess, then the rating will eventually go up slowly over the long-term if you are doing things correctly

I am not so sure about this. Even if such a "purge" did exist, then I don't think the number of accounts taken out would impact the general rating pool enough - with literally millions of chess.com members, even many accounts taken down would likely change little since they make up such a small ratio to the "honest" members on the site. 

I think that the algorithm makes choices about who we play that are not Hobsons choice, next one up, type of thing. Also I think that it is like casinos, the house always wins. There is a thing about rats, I have heard of, that when two rats play fight, if the bigger always wins, the other won't play any more, so to keep us coming back to play, every now and then, we win a few, rating goes up. But not too much, then straight back down again. It is quite easy to watch it play out on the graph.

OTB ratings work the same way with ups and downs (improving players may have less of that in a particular rating graph). There isn't anything nefarious going on and is completely dependent on the player and any of the various situations in life that can impact quality of play.

The only difference between here and OTB is that it's a lot harder to get so many rated games in over any period of time. Here it's just a push of a button and the cycle of just one more game and I'll log off when my rating gets back to X 

You could run computer analyses on games from long ago, and games from today, at the same rating and using the same engine. My gut impression is that yes, the same rating from 1250-1400 requires slightly more skill than pre-pandemic.

Hey,

 Yeah,  I guess it did.  I just fell 200 points from no where.  I though I can't play anymore.  But I see that i'm not alone.   Currently wins 3 - loses 15  ( rating fell 1700 to 1500 )

I just could barely beat 1400 player he achieved  81% accuracy in 35 move game.

Harder faster stronger better work it do it makes us stronger faster smarter

Exactly how it's going for me.  I'm trying to fight my way out of an hole at the moment, every time I nearly make it to the top I slip back down again. 

Yes, this is relatable and you may also find humor in this older thread of mine

https://www.chess.com/forum/view/fun-with-chess/my-online-chess-rating-59795420 

Just tell yourself: "Dammit... flucked again!"

As with most game design these days, the match engine / opponent pairing algorithim has 1 priority / success criteria, and that's to keep up engagement rates. A company like chess.com can only really improve their product via the pairing algorithim, and things like UI improvement (which doesn't have much room to improve) + game analytics and lessons etc, so they will be pumping alot of resources into the pairing algorithim as it's one variable they have with which they can directly control our engagement rate.

Engagement rate increases with psychological heuristics like contrast effect, where hot streaks are much more dopamine fulfilling after a cold streak. Streaks are most likely in the pairing algorithim by design, and as others have pointed out "variance" is natural but I (as an AI & behvaioural psychology specialist in app design) would imagine the chess.com guys would be leveraging variance as best they can to improve engagement rates. They will have a lot of data on your addictedness level, as well as the global levels, and have figured out a "threshold" for how much losing you can tolerate before you never come back. If you have a low threshold, i.e. you have left the game for a while already, I imagine the pairing algorithim will more likely match you with someone that has a weighted average to lose against you. Rating is only the visible number the players can see that express the skill of the opponent, but things like existing streaks, real-time accuracy levels, how long the opponent has been playing in the last few hours (speculated concentration levels) etc would be taken into account, and the ratings could be "adjusted" behind the scenes to reflect the present strength of the opponent based on all data points they have accumulated. It's no accident that you will get paired with 85% accuracy + at 900 ELO for 10 matches in a row following a rise in the rankings. Without those cold streaks, the hot streaks wouldn't be as engaging and we'd get bored of the chess.com product faster. 

If the pairing engine was completely random and only based on the rating we see, there would probably be much less variance and streaks - and we'd also not still be playing despite the fact 900s have been playing like 1200's for the last 10 games. We are chasing that dopamine buzz of finally ending the cold streak, and we've been conditioned to realize that it does end and it feels great when it does. 

Also don't forget, everyone is playing more frequently because of the ease of access and things like Queen's Gambit show, so the global average skill level might be improving. 

I agree that players are getting better for their rating. Of course there are a number of reasons why. Quitting caffeine and playing chess, though... that's like jogging in ugh boots.

maybe your just tilting or simply not playing as well as you do before

When i took a long break from chess and came back, I played really well to my surprise, but after a bit, I got back down to my usual slightly dumb slightly smart self.

there’s something about approaching things with a fresh mind that makes a difference, but I don’t know how yet.