matching algorithm

It seems to me that there is something wrong with its algorithm.

I noticed that too.

When I had bad day and I started to lose one game after another.... chess.com matched me with opponents with lower rating than mine.

And it irritates me, of course.

That's it.

On lichess there is not the such thing... and every new players connection looks like quite random.

Good point mate. 

Please test it if you have time to figure out algorithms with brute-force. I have a full time job and I play chess for recreation purpose only.

Does any casual player have time/resources for detecting an algorithm with brute-force?!

This is just start of new era. Welcome to the world for AI rising. 

Cheers,

AB

stop making excuses i have a full time job and overtime.. just sleep less.. just pulling your leg mate, yes, we could hire someone to or we could put the effort or do it passively, a little at a time.  take care mate

Do you people honestly think that all the algorithm is doing is matching people by ratings and that's it?

Consider the years in development of the Chess.com website, the fact that it is a BUSINESS trying to make money and player hours clocked means more money...

That's pretty much it. There are some attempts to prevent players with high lag from playing ones with low lag, there's a poor sportsmanship pool where those players tend to get paired together, and it will initially try to pair premium members with other premium and move to a wider pool if a compatible pairing if one isn't found quickly

Aimchess training will show you all of the points that chess.com can analyse about a player's style regardless of their rating. There are many. You don't even need to use aimchess training to see it, most of it is right there on your chess.com account anyway. Things like how often you use an opening and how often you win with said opening vs. how often you lose to certain openings, how well you do in an endgame situation etc. is all stuff they can feed into an algorithm to pair either favourably or not. 
It may well be that there is a trigger that is switched after so many wins or losses that it pairs you with players who's styles are more likely to win against you. That is not to say that there is anything "sinister" going on. Quite the opposite to be fair. If anything, it is the best training you could hope for. Something that will let you win sometimes and then when you are feeling on top of the world, will kick you back down again and make you work for your rank.
Why anybody would suspect that chess.com DON'T use some kind of algorithm like this is, in my opinion, quite naive. As others have pointed out, they are a business and they will seek to maximise customer engagement whenever possible.

I am certain you are wrong in general.

Amusingly, this can be essentially proven, because you have failed to notice that your guess means that half the players would be being matched against players who are LESS likely to win against them. Why??

You can be sure they rely on ratings to indicate strength and to decide who gets matched to whom.

The more conditions added to a pairing algorithm, the slower it will be to find a compatible pairing. The site wants to pair members quickly in the vast majority of cases, so adding additional checks that will slow that process down or limit the pairing options, isn't going to be used.

Especially such bizarre and pointless modifications.

All I know is it hates me and I'll forever be my rating, bet you, so I've trained my mind to just place a 1 in front on bad days and a 2 in front on good days. Tell it, it will always be number one in my book, if you know what I mean,and play on cuz complaining only makes it worse.

Wake up, players.

We are being played.

it might be also likely that whenever you climbed a bit, or hit your all time high, then the algorithm might be sending you very experienced players which have played a lot of games. Imagine there are 2 players, both are 1500, one with 2K games the other is 20K games. I believe generally it is harder to beat the latter, therefore if the algorithm sends you the latter more frequently, you will start losing your rating. Then you will try to get back to the rating you had in the past and you will find yourself in a loop. I am not claiming that this it the method chess.com uses, but if I were the owner of this platform I would be doing this way. This partially explains the rating fluctuations we all are facing.

Number of games played has no bearing on pairings