anandrulez, nobody agrees about what "inflation" is. in a <relative> system like the rating system, the concept is pretty much meaningless: ratings can't be used to compare chess strength more than maximum 5-10 years apart in any near-meaningful way; i'd set the cut-off closer to 5 years rather than 10. thinking that one can "adjust for inflation" and thereby increase the scope for which players' skills can be compared using ratings, is imho the final evidence of utter confusion wrt (chess) ratings.

equally bad (or worse?): most people don't understand how to use ratings to compare <current> elite players. so why bother discussing "inflation"?

Yes if you are talking about middlegames. Most chess games are decided in the middlegame. Indeed, they had longer time controls in the past. However, they had to start playing over the board much earlier because novelties then occurred earlier too. Nowadays, many GMs begin OTB analysis at move 15 to 20; they simply follow a previously played game in under 15 minutes before that. I think this is what Grischuk refers to when he said in an interview that today's players actually have a time advantage over past players. (I may have misinterpreted this interview of course, correct me if I am wrong.)

Regarding endings, I would agree that today's top GM endings are worse than past endings because adjournments have disappeared.

But that is only one factor why GMs today seem to be playing poorer quality endings. Even with adjournments, notice that after just about 5 moves after adjournment, the past chess players had to start playing over the board once more. Regarding players being more tired today by the time they get to endings, it often happened in the past that a player had to play not only one, but several adjourned games in a single day. That must have been tiring too.

The most funny thing is that when games are entering into the most 'friendly', for spectators(amateurs), part of the game(endgame), they are often decide by blunders. So maybe today players aren't weaker but have simply less time in the endgame.

Couldn't agree more, while I think it would be difficult to give a precise measure of how much rating increase is due to inflation I think generally we can get a pretty good grasp of about where it is.

I also think the play has improved to some degree and many more players are playing so the talent pool is a lot broader, however a bit shallower post Kasparov.

vbd, why would anyone generally think that someone rated 2780 would "get crunched" by someone rated 2800? they're rated the same.

anyway, i'm totally convinced that kramnik, anand, carlsen and aronian all have equal or better chess skills compared to fischer in 1972. period. and that's not based on ratings at all.

I took that to mean players in the 2800's not exactly 2800 so perhaps Magnus with his peak of 2826 vs Fischer with 2780 peak. Now would Magnus really perform 46 elo higher than Fischer at his peak? I doubt he would even have come out ahead in a match up and I think that was VBD's point.

On another note a 2780 vs 2800 are nearly equal and yes you wouldn't expect the 2780 to get crushed.

FWIW, I think Fischer was better. During his epic win streak many of the games he played became known as all-time classics; bishop and knight vs Taimanov, the Alekhine against Spassky, the QGD against Spassky, and others. Chess moves and chess games are always the "tip of the iceberg". What we see is a small fraction of what was calculated and what was understood. I think that to accumulate so many amazing wins in such short time, Fischer's "iceberg" must have been incredible.

I discussed old versus new with a friend of mine who's a 2350 player. His opinion was that (for example) Anand 2011 is not necessarily a stronger middlegame and endgame player than Karpov 1982. Karpov is fiddlesticks today, but he doesn't even try anymore; he remained near the top of the world until he semi-retired in his late 40's. Ivanchuk and Gelfand are doing well today, and Lasker held his own in the 1930's.

1) Clock simuls. Compare the giver's thinking time to the total time of his adversaries. Compare his elo performance to his rating. This should give a vague idea how much relative time deficit affects performance elo-wise.

2) Compare the absolute level of play of similarlily rated players in games played at various time controls, given the difficulty of positions and the practicality is about the same. For example take a handful games from each time group: bullet 1+1; blitz 5+5; rapid 25+10; standard control 90min per 40 moves + 30s and put results on a chart. I'd expect a loghartihmic correlation; the longer TC-s, the less increase in strength.

So, as long as we're restricted to games not containing adjournaments, it's pretty easy.

But if the samples include adjournments, one could use penalty time additions, 1 or 2hrs for example.

I believe Fischer's level of play in both his 6-0 matches would be adequate to perform consistently 2800+ TPR even in today's elite tournaments.

Zatopek and Haile obviously were better runners than Nurmi on absolute yardstick, and their relative strength is a matter of speculation, methodology of comparison and personal taste. But that's not what's important. Chess, unlike athletic events, has suffered from the fact that relative comparisons using ratings have become possible only quite recently and there's still no systematic attempt nor scientifically 100% trustable way to measure chess strength directly from moves played. This is why computer-aided attempts to get an objective yardstick exist.

It depends on a person.
I wouldn't claim anything about the strength of past chess players, if not backed up by computer analysis.