Why online chess ratings are inflated/deflated

these are some simple reasons why,

1. cheating, it would be very easy to do.

2. Players can study their games as long as they want, so there could be 1 hour spent into a move

3. some players just move the first move that pops into their head.

could you please post any other thoughts on this.
CP 

true, but this could explain why some ratings are low. but how do you get a player who has play thousands of games in both standard and online chess and has a gap of 500 points.

Compare percentiles and ratings. It will answer your inquiry.

You are right. I have a higher percentile by almost 2 percentage points in live standard than correspondence yet my rating is 160 points lower in live standard compared to correspondence.

The reason correspondence chess ratings aren't reliable is that you have a very diverse pool of players.  There are players who only play "online" chess because their connections are lousy and would really rather be playing blitz.  They play almost as quickly as they would blitz as you pointed out.  At the other extreme you have your traditional correspondence player who uses all of the legal tools available to him.  He uses databases and opening books and looks at videos and sets up analysis boards and sweats every move.  NM  Aww-Rats (one of these days I'll learn to spell his handle ) recommends 2 to 3 hours of intensive study for each move and believes strongly that's what got him an OTB title.  Then you have folks in the middle who want to study the position but not use books or databases and basically are trying to replicate an OTB experience.  They'll take minutes to look at the position but not hours.  Those are 3 very different approaches to playing and the ratings reflect that diversity.  It's perfectly reasonable for someone who spends hours on each move and takes advantage of the legal resources to have an "online" rating hundreds of points higher than his live rating.  Now if it's 1000 points, then maybe there's a problem.  More of my $.02 from an apparently infinite store of pennies. 

Timeouts are probably the biggest reason in general.

But also the disparity in effort one can make will affect individual differences.

Inflated correspondence ratings can be explained by the ability of launching many games at once and then simply abandon them. If such a member closes his/her account it creates rating inflation among other members benefiting from the abandoned games. But this is not an official explanation ( if there is any ) just my best guess.

You are making good points but diversity in itself can not explain the fact that the _average_ rating in correspondence is higher.

I think the Glicko formula has some bearing. I have continuous correspondence games going, and have had since joining this site. But, I play slowly in many games and keep my game load under one dozen games (often half of that) and, consequently, my RD hovers around 90. I gain more from each correspondence win than I do in blitz where my RD is low.

Well that also affects games you lose so may make your rating less accurate but no more inflated.

True, but added to other factors--percentage of timeouts, wins against banned players early in the game, taking an extraordinary amount of time per move--that all tend toward a much higher winning percentage, and the distortion creates the appearance of inflation.

I don't believe in inflation, as it presumes a logic of comparison across pools that has been shown repeatedly to lack merit.

Yeah, there's no objective ability of a 1300 player, it's just someone who scores 66% against a 1200 player in that pool.

Just thinking about it (so may be wrong) seems the biggest factor of inflation would be new players who join, lose 5 games out of 5, then never log back in again.  i.e. they inject some points into the pool then leave.  Doesn't sound like much until you wonder if all 8 million registered accounts only played a few games each what it would do to the regular player base of...  lets call it 50,000.

my Online rating is around 1800 while Live/Blitz rating in the range of 1400-1500. This can be attributed to a number of factors, chief among them being "Net Connection". It doesnt matter if u have a bad net connection in Online games, like it does in Live- Blitz /Bullet games. The other reason is the lack of "Obvious" and "Stupid" blunders in Online games (like say losing a piece/queen or even Mate!).This obviously is due to the fact that u have a lot of time to Think,Evaluate and Analyse the position unlike Live Blitz/Bullet games. The third reason is, in Online Chess, we spend a good amount of time Analysing/Evaluating a position, trying out different variations/double checkig our plans etc, thereby playing far better quality chess than Blitz/Bullet games. the last reason cud be the lack of stupid things happening in Online Chess, like say a Phone call,or a door bell etc...

Online Chess starts at around 1200. It takes some time to get down to 500 or so but much less time to get up to 1900...

Games won by much thinking or lost to blunders, phone calls and door bells are lost/won by someone else. This is a zero sum game so this explains nothing as far as the averages are concerned ( which are the embodiment of inflation/deflation ).

oohh...maybe i got it wrong...i mean i was talking about my personal ratigs difference. seems like the topic is entirely different Lol... :P

May be I got it wrong. But my main interest is the deviation on the level of averages which is much harder to explain than personal deviations.

For me it's 2 things:

I use the opening explorer - in Live games I make more mistakes in the opening

I use the analyze tool - in Live games I have to do it by head so I can calculate less far ahead