are ratings being rigged at chess.com to farm engagement.

Over the course of several thousand games in the span of a year, neither a person's lowest rating or highest rating will be an accurate reflection of their chess skill, so it makes little sense to base pairings on those outliers.

It is true that a 1700 with a peak rating of 2000 is probably better than a 1700 with a peak rating of 1850 ( but there are valid reasons why that may not be the case). But over the span of a few weeks a person's rating will return to its "true" score.

Over the course of many games, a person will play just as many people who are "underrated" as "overrated" so the long term effects on that person's rating will even out.

Given the very low stakes of the vast majority of games on chess.com, I think efforts to somehow include an estimate of "true" rating hardly seem warranted. Especially because that kind of exercise would need assumptions about how to estimate a "true" rating, which could sometimes result in even worse discrepancies than the current system.

In your last 20 games, you had 3 wins - 1 loss - 2 wins - 3 losses - 3 wins - 1 loss - 3 wins - 2 losses - 2 wins.

This is the exact opposite of what you're claiming. No long streaks at all.

I said recently as opposed to "In the last few games I played before making this post"...

But to be fair I did try to find exactly how long the last streak was on my results listings but I couldn't see how to get past one page for 960 games. If anyone can help me out on that you'd be able to see.

Click the right-arrow: >

Then click the heading that says "Live".

Had not even considered this. Lost 12 Rapid straight. Lost 24 of 42 blitz games the same weekend. A complete ELO crash…..

put it down to just playing crap and not stopping when clearly tilted ….

Makes no sense to rig ratings - more likely to fustrate and exit users….

Hey, thanks for that!

So on the 2nd page of results I lost fifteen consecutive marches between 22-25th of March.

On the 6th page I lost fourteen consecutive matches, again with a similar 960 ratings plunge from 1350 to 1250.

And there were definitely further times this happened.

I just find it odd. Sure, you can be on a bad spell and lose more than you win, but these runs just seem daft to me.

the main problem on this site - no possibility to play with more rating then you have. Even in tournament you are forced to play with your rating 99% time.

I won't rule out that chess.com is rigging ratings in some indirect ways, but fwiw 300 point Elo swings are pretty common among players who aren't concerned with maximizing their rating. I don't care at all about my lichess ratings, it's my playing at 3am or while distracted watching Daniel Naroditsky account. I've lost and regained several hundred points several times.

In particular, I see a lot of people that play instantly during rapid. Their clearly strong players, but just impatient or cocky. These players feeling more or less patient during a week can swing their Elo massively.

In coin flipping, the probability of getting 15 tails in a row is (1/2)^15, which means it will happen about 3 times for each 100,000 set of 15 tosses.

For the sake of simplicity, assume each person on chess.com has a 50% chance of losing each game they play.

Wikipedia says chess.com has about 11,000,000 million daily active users. Assume each plays 15 games a day. That means each day about 330 players will have a 15 game losing streak. And that's just based on randomness, not taking into account how many people are on tilt with losing chances of more than 50%.

Obviously this model is far too simplistic, but the point is, as millions of people play billions of games, it is a certainty that some of those people will have exceptionally long streaks.

1900-2100 range is weird. Sometimes(in 10+0) people play 95% but others barely 80%. Some games are easy but some are impossible

Your math is wrong. Precisely because a person with live rating say 2248, when he/she plays opponents of same level, the percentage of loss is 50%. But as he/she reduces in ratings, his/her probability of success rises more and more. And I have seen a player with peak rating of 2284 something in 1700s. so doing your math, the person should be on a tilt 66 times . That means, probability of 1 in 10^20. And it isnt just one player.

That person had a tilt of 500 points, it should be 7 in 10^20. Now do that with a tilt of 200 points, 300 points and 400 points, for 400 points, it should be 9 in 10^16, for 300 points tilt,7 in 10^12, for 200 points tilt, it should be 3 in 10^8. Now multiply that by every person whose live rating is 300 points below their peak rating . How can you explain that by math?

I will never understand why so many people here have so much of their self-esteem invested in a rating. They are not even playing under their real name, so why does it matter so much? Nobody in the real world knows or cares. And anyway, rating is not an absolute. It depends on the games you play. It is not going to continue going up or down. It simply measures your strength at a particular moment. If someone becomes so totally fixated on a chess rating, that's an unhealthy thing. Maybe find a less stressful hobby.

I guess the OP decided to find a less stressful hobby...

there is definitely something fishy going on, cause threads like this, where people are complaining about massive losing streaks and not being able to win a game at their rating have been a daily occurance since the candidates started, so chess com better get their act together cause some manipulation is sure going on, otherwise there would not be so many frustrated people making these kind of threads

Sure everyone gets loosing streaks including me

I don't think someone loosing too much means it's rigged even if there's a 0.0000001 percent chance to loose 15 games in a row that's still at least a few hundred people times the amount of games means a good chunk of chess.com will get a loosing streak at some point(very sloppy stats but I hope it's a good enough example) most people who grind a lot will get a loosing streak at some point better to cope with it then blame the system

Indeed^

So many players like to blame others for their own mistakes.

"I lost - therefore, it's someone else's fault, not mine!"

So eager to point the finger at everyone else, to accuse the "big bad chess website" of conspiring against you ...

When I lose - I blame myself. Then I analyze the game to see what I could've done better. Nearly always, the fault is my own. I played too slowly. Or too recklessly. Or I missed a simple combination. Or I weakened my structure ...

If you want to stop losing so much, you need to take the time to figure out the reasons you've been losing.

Most of the time, the culprit is quite simple: it's the moves that you've been making on the board.

I get streaks in over the board chess. It happens.

I get the impression that many people who create thread after thread about losing streaks don't really enjoy chess that much. For some reason they are just hung up on the number connected to their fake name. Study more, play more, study more. But so few people want to be told that real improvement requires dedication and hard work. Improvement does not just come automatically when you play a certain number of games.