The dubious K-factor of 40
What da hell? I thought this was a chess blog? Your header sounds like math!
While I do like the occasional math problem this is indeed a chess website/blog! But bear with me. First a little explaining for those who don’t understand. First we will explain roughly how chess FIDE rating are calculated.
Lets take an example where the difference in rating is a 100 points, so lets say a 2000 rated player vs a 1900 rated player.
First we need to look up in this table (don’t worry, tournament organisers and programs do this for you):
From this table you see that for a difference of 99-106 (which fits our 100 point difference) the player with the 2000 rating is expected to score about .64 (64%) from this matchup. If said player wins, he scored 1-0.64 = .36 over expected score. Same thing applies for a draw 0.5-0.64 = -.14 and finally if he loses it’s 0-0.64 = -.64
Now we come to how the rating is calculated. Based on this score from the game, the player K-factor is used to multiply this number. Now for most players this K-factor is 20 (recently changed from 15). So if the 2000 player wins, he gains 20×0.36 = 7.2 points from this game and is now rated 2007,2 (always rounded up to 2007).
For players over 2400, this K-factor is 10 which means lower swings and finally this K-factor is 40 for new players for their first 30 games and for players under the age of 18 while their rating is under 2300.
The change from 15 to 20 already allows for bigger swings for most players (it’s a 33% increase after all!) but the K-factor of 40 for the young players is proving to be somewhat controversial.
These changes in K-factor are rather recent and my opinion is that they will have to review it a bit before too long. There used to be a time when the starting FIDE rating was 2000. This has been gradually decreased and now everybody can more or less get a FIDE rating and I believe it goes down to about 1000 these days. This has pros and cons. Of course it’s great that players can start playing and most likely immediately get an offical FIDE rating after their first tournament. In the past this could however provide a little bit of a hurdle for younger players (who remember in the past would receive their first rating over 2000 when ready and go from there) who would get a starting rating in the 1300-1500 range and would then spend years and years waiting for their rating to catch up to their strength.
I believe the 40 K-factor for players U-18 years old was put in for precisely this reason. This however results in huge swings…and not only upswings! One of my friends for instance had been steadyily building his rating but then he plays a bunch of underrated players in an international youth tournament and loses over 100 rating points without having a huge disaster tournament!
Another recent case of the opposite is almost ridicilous. Have a look at this young talent, Parvis Gasimov:
Obviously this is a young and talented player BUT you don’t go from 1949 in October 2014 to 2517 on January 1st!! That’s beyond ridicilous!
If you take a look at his tournament results, sure they are impressive but not this!
He starts of on the November list with two tournaments rougly on par (1949) then he has a fantastic European U-14 with 7/9 and gains 211 rating points! His opponents average 2206 rating points.
On the December list there is another jump. But note how convenient it is! Our hero is now rated 2157 and remember that the K-factor for under 18 year old players only applies while their rating is under 2300.
On the December list he first gain 98 points from some Russian tournament. Then either he plays a rather shady tournament or somebody either organised him really well or gave him good advice. This player now with a rougly 2200 rating (and probably his rating hasn’t caught up to his strengh if judged by European Ch.) plays a 9 round tournament with very weak players where beats an 1800 a 1900 and then a bunch of players who’s rating floor needs to be raised.
The rating floor being raised means that these players are rated more than 400 rating points below the stronger player. So in this case their rating is 1757* when calculated. For all we know these could be players that barely know how to play but I only bothered looking up his first round opponent which was a 7 year old girl rated 1515 😉 ….so most likely a very well calculated move by his trainer or somebody close to him. From this very weak field and for his 9/9 he picked up 40 rating points. For comparison when I scored an IM norm with a rating still under 2300 I don’t think I even gained 40 points!!
Ok, back to our hero, Parviz Gasimov. He is now strategically well placed with a 2295 rating only 2 months away from a 1949 rating!! This means that he has another month of play to use his K-factor of 40 to gain more rating. This he did with another very strong performance of 12.5/15 in a tournament in Aleksin in Russia. That tournament even started on the 27th of November so he gained his rating from that tournament from his previous 2157 rating (….this is all getting a bit complicated if you are not familar with ratings!!). Although 12.5/15 was impressive, he didn’t beat any titled players and nobody over 2300! He did draw against a 2350 rated player.
Also during this rating period somehow a tournament that started in August was snuck into the January calculations (either handed in too late or last round was played in December). From that tournament he gained another 40+ points.
Then finally he played one tournament on his current rating (2295) and FINALLY managed to beat a player rated over 2400 when he beat Indian IM Das Sayantan. From this tournament he gained only 7 rating points.
So after only 3 months, he goes from 1949 to 2517. And note during this time (and a LOT of games) he beats only ONE player rated over 2300!! Does it feel normal to you that a player that has beaten ONE player over 2300 in his career is now rated 2517?
Now…I am in no way saying that anybody was cheating or anything like that. I don’t know this player and I believe he is a talented one for sure as Azerbaijan is serious about their chess! But I don’t believe he deserves to be rated 2517 at this moment in time and I also think that somebody took care to utilize the current flaw in the rating system to it’s maximum by strategically selecting tournaments and strategically making sure certain tournaments were not sent in on time.
I remember that there used to be a CAP on how much you could increase your rating over a certain period (3 months if remember correctly). I believe rules such as this should be put in place….or at least something to prevent and extreme case like this. Over 500 points in 3 months is NEVER normal!!
As you can see there is plenty of scope to bend these rules and something needs to be done. In the past there have been some shady tournaments here and there around the world where results are only on paper to get a rating for somebody that needs it. This K-factor of 40 certainly invites something like this to happen. If you have a chess community somewhere and you get a very young promising talent you could now utilize this player to hugely benefit the community and both him and other players to achieve their goals.
Now how could they use this young player? Well they would simply strategically arrange results in tournaments over a short period of time losing a lot to the youngster until he is just under 2300 (this is assuming a bunch of players rated 2100-2300). Then for the next month they would again arrange a tournament or two such that this player would go over 2500. Going over 2500 is a HUGE benefit for a young talent as now he never needs to worry about his rating while chasing GM norms to get his title. One only needs to get over 2500 once to get the GM title if you attain the 3 required GM norms.
Once this player has reached 2500 and his K-factor is now 10, he can start to strategically lose a bunch of points to his opponents who earlier donated their rating. Note though that they donated with a K-factor of 20 for their minus points while the plus points used a K-factor of 40. Now in our fake future scenario the young player is donating back with K-factor 10 on the minus points and the others receiving points with K-factor 20. As you can see this has created a huge inflation within this chess community (or country or however we envision it). So now the 2300 players can try to get to 2400 courtesy of this young talent and everybody has gained. You have been living under a rock if you don’t think something similar hasn’t happened in the past (hint: it has!)
Please leave a comment be registering on the page or discuss on Facebook here: