PDA

View Full Version : internet ratings: are they full of it?



qpawn
26-06-2006, 05:26 PM
Are internet ratings of any value? I am 1700 to 1750 ELO on the net and only about 1200 in ACF :D

In fact, I feel like giving up chess except for playing on the net since I feel that about 1700 is where I really stand. I play far more confidently over the net and my tactics are much better. I even beat players over 1800 ELO sometimes and I can never do that in person.

Why should I bother playing chess in person when I can be over 1700 on the net? I have even got draws in all the net simuls that have been done by titled players.

Rincewind
26-06-2006, 05:47 PM
My feeling is that ratings on the net are around 200-300 over the ACF rating in most cases (certainly ni mine also). This could be due to players of lower ability competing more on the net as compared to those who play competitive OTB chess.

It is important to remember that the number doesn't mean anything as an absolute measure. It is just a comparison which only has a meaning when you perform a difference between to ratings. So two players ACF rated at 1200 and 1500 or those which FICS ratings of 1500 and 1800, the difference is still 300 points which ever way you dice it.

I don't know if this is good or bad news for you. But I'll repeat the mantra of ignore ratings and enjoy chess. :)

Kevin Bonham
26-06-2006, 06:22 PM
There are a number of ways internet ratings can become out of whack. As well as the one Rincewind suggested, ratings also become inflated because players deliberately pick the time controls and opponents' rating range to which they are most suited. Some players also artificially inflate their ratings by working out how to beat computers on those servers and playing them repeatedly.

The 500 point difference in qpawn's case is a bit on the high side and suggests his comment about playing better on the net is likely to be accurate. Typically players who are most familiar with one way of playing will initially perform worse when switching to the other.

By "titled players" do you mean players with OTB titles?

qpawn
26-06-2006, 06:30 PM
Yes, there are players on playchess with bona fide IM, GM etc titles [ ones got via FIDE] who give the simuls. I have been in about 3 simuls given by IMS and have drawn in every case.

You might be right about the time controls. On the site "slow" is defined as 15 or more as long as there is an increment [any] I have played about 100 games there on the slow rating [ a good sample] and hover between 1700 and 1750. I can not do anything with bullet/blitz but that's largely due to my ancient PC and dial up.

Vlad
26-06-2006, 08:54 PM
500 certainly sounds like a very reasonable number to me. In fact, the current difference for me between ACF and ICC is exactly 500. :)

My son has a rating around 1100-1200 on the playchess. His junior NSW rating is gonna be around 350-400, that translates to approximately 700 ACF. Again the difference is almost 500. I can certainly see that playing over the internet for him is much easier: 1) there is no touch move rule; 2) He can place a piece on a squire and then realize that it is a bad move and put it back. Pretty much he can check all the squares around by moving one by one. 3) He is not intimidated by bigger boys. They can not start asking questions like "How old are you?" or "Which grade are you in?".:lol:

Vlad
26-06-2006, 09:17 PM
My feeling is that ratings on the net are around 200-300 over the ACF rating in most cases (certainly ni mine also). This could be due to players of lower ability competing more on the net as compared to those who play competitive OTB chess.



The phenomenon you have described above is called "adverse selection", for which Akerlof got a Nobel Prize in 2001. I wander if you knew about that or worked out yourself. :)

Desmond
26-06-2006, 09:36 PM
Are we comparing apples and oranges here? Most chess sites I know of are either rapid or correspondace, and I suspect that these are being compared to "Standard" ACF ratings.

Rincewind
26-06-2006, 09:42 PM
The phenomenon you have described above is called "adverse selection", for which Akerlof got a Nobel Prize in 2001. I wander if you knew about that or worked out yourself. :)

Seems obvious to me.

qpawn
26-06-2006, 09:50 PM
I find that I can only play well against players of 1900 ELO and above on playchess if there is a generous time limit like 60 mins each with 10 sec inc etc. The standard seems to rise quite suddenly at about 1900 onwards and after 2000 ELO I am really out of my depth; beyond 2000 it is quite clear that such players know more theory about chess than waht I do. From memory I think that the higest rated player I have beat was 1893 and I have got quite a few draws with players on 1950.

Vlad
26-06-2006, 09:56 PM
Seems obvious to me.

I like mathematicians using words "obvious" and trivial all the time. :)

Vlad
26-06-2006, 10:11 PM
Are we comparing apples and oranges here? Most chess sites I know of are either rapid or correspondace, and I suspect that these are being compared to "Standard" ACF ratings.

Even though you are right that the time control is different in the internet and OTB, clearly there is a common component.:)

Jesse Jager
26-06-2006, 10:27 PM
does it really matter what ur rating is, to enjoy chess. If so, i think u maybe playing chess for the wrong reasons.

Vlad
26-06-2006, 10:42 PM
does it really matter what ur rating is, to enjoy chess. If so, i think u maybe playing chess for the wrong reasons.

99% of people play chess for the wrong reason. You must belong to the lucky "right" 1%. :owned:

Jesse Jager
26-06-2006, 10:45 PM
I sure hope so, or else i would have wasted a hell of alot of time.

Desmond
26-06-2006, 10:46 PM
does it really matter what ur rating is, to enjoy chess. If so, i think u maybe playing chess for the wrong reasons.
It is a quite simple progression:
1. You play chess to enjoy it.
2. You find winning more enjoyable than losing.
3. The better you get, the more you win and therefore enjoy playing chess.
4. You associate getting better with an improving rating.
5. You enjoy watching your rating improve.

qpawn
27-06-2006, 12:17 PM
Well, I don't play chess FOR a rating. But it is a measure of how I am going.

The difference in my strength of play is so much higher on the net compared to OTB; it's like comparing the soccer skills of San Marino and Poland :D

Further, my net rating is improving and my oTB rating is going south.

piecelover
27-06-2006, 02:02 PM
I used to play online on chess21 and on playchess. The differences from fide elo rating and these online ratings are so huge that I doubt that they are calculated with similar formulas. For example I have a friend who is rated 100 points above my rating (FIDE rating) and online he's sometimes rated 500-800 points more. How can this be? I beat him OTB (not too often though:cool: ).

qpawn
27-06-2006, 02:27 PM
One reason for that can be that on line you can choose only to play people of a certain strength. Indeed, you could connive to get a high rating by only playing weak players. In the playchess ratings the average rating of your opponents is given; for my slow rating of 1757 I have played against an average level of 1695. In other words, I have got my rating against about the same level as the rating. But there are someplayers who are rated about 1900 who have played ,on average, against a 1550 level! Clearly, their ratings are far more doubtful than mine.

I don't know what formula is used by playchess but I think that it's the standard ELO . In contrast the formula in ICC is definitely non-standard; a custom formula involving USCF ratings .

piecelover
28-06-2006, 05:11 AM
I've been thinking about this some more. Maybe it is because you have the opportunity to play thousands of games, even tens of thousands. You don't have an opportunity to play that much games in real tournament play (or in your whole chess career). For example: If you win a OTB tournament with 9 out of 9 and you have a performance of 3000 then you will get e.g. 50 elo points. But next week you'll be in lousy form finishing at a last place with 3/9 and lose all those points. So if your current strenght is 3000 points performance you have an opportunity to reach that rating in on-line blitz play. You play some 100 games and you are on the top! But if your performance for that thay is 2200 because you did not drink your morning coffee then if you insist on playing you will probably reach that rating quickly.
What do you think?