ELO rating of 2021 F1 drivers
As people familiarized with chess knows, ELO rating is a (scientifically based) system to rate and rank chess players according to their results. This system (or others similar to it) has subsequently been used in some other sports and, particularly, in online games.
-------------------------------------------
Warning: boring babbling jargon ahead
You can junp to the end of this section
-------------------------------------------
Basically, the ELO (after Dr. Arpad Elo) of a "player" is a number meaning the performance of that player. To put it in context, a 200 ELO points of advantage for A respecting B mean A has aprox a 75% probability of win against B (i.e. A will beat B aprox 75% of times), while and advantage of 100 ELO points mean aprox a 67% chance of winning.
One of the beauties of this system is it allows to calculate the probability of win for any player against any other player or estimate how many wins will have in a series of plays (races). Another use, in fact its main use, is to rank the players according to their results.
In the case of racing I will take each race as one "game" where driver A wins over driver B if A is better classified than B in that particular race. Thus, the winner of a race is seen as winning against all his (not her in this F1 case...) opponents. As with chess, it doesn't if the win is marginal or devastating, the result is win or lose (let's note there is not draws here).
---------------------------------------
End of babbling
---------------------------------------
By now, there has been 10 "games" (races) and the matrix with all the wins/loses of every driver against every other driver (only the first 8 in the championship) is as follows:
For example, Verstappen has won 8 times against Bottas, who only won 2 times against Verstappen.
The calculated ELO for all those 8 drivers, putting arbitrarily 2800 ELO for the best driver (because 2800 tipically is the ELO of a chess world champion type of chess player) are:
Not huge surprises there (there cannot be really any , isn't it?) but perhaps it is slightly surprising Lewis is above Max. It is due to Hamilton having slightly better results against Leclerc and Ricciardo.
According to those ratings, Norris has aprox a 75% winning chance against Ricciardo (that is a 3-1 winning ratio). The real data between them is even crueler: 9-1 . It shows the 9-1 score is a bit bad luck for Ricciardo in comparison to how both of them perform against other competitors. I.e. without bad luck, Ricciardo would have 2-8 or 3-7 versus Norris.
Rating difference between teammates is minimal in Ferrari but it is around 150-200 points in the other teams. It means typically the strongest teammate beats the weakest one 70%-75% of times.
-------------------------------------------
Warning: boring babbling jargon ahead
You can junp to the end of this section
-------------------------------------------
Basically, the ELO (after Dr. Arpad Elo) of a "player" is a number meaning the performance of that player. To put it in context, a 200 ELO points of advantage for A respecting B mean A has aprox a 75% probability of win against B (i.e. A will beat B aprox 75% of times), while and advantage of 100 ELO points mean aprox a 67% chance of winning.
One of the beauties of this system is it allows to calculate the probability of win for any player against any other player or estimate how many wins will have in a series of plays (races). Another use, in fact its main use, is to rank the players according to their results.
In the case of racing I will take each race as one "game" where driver A wins over driver B if A is better classified than B in that particular race. Thus, the winner of a race is seen as winning against all his (not her in this F1 case...) opponents. As with chess, it doesn't if the win is marginal or devastating, the result is win or lose (let's note there is not draws here).
---------------------------------------
End of babbling
---------------------------------------
By now, there has been 10 "games" (races) and the matrix with all the wins/loses of every driver against every other driver (only the first 8 in the championship) is as follows:
Code:
VER - 5 8 8 8 8 8 8
HAM 5 - 8 7 8 9 8 9
BOT 2 2 - 7 6 7 7 7
NOR 2 3 3 - 5 7 8 9
PER 2 2 4 5 - 7 6 8
LEC 2 1 3 3 3 - 6 7
SAI 2 2 3 2 4 4 - 5
RIC 2 1 3 1 2 3 5 -The calculated ELO for all those 8 drivers, putting arbitrarily 2800 ELO for the best driver (because 2800 tipically is the ELO of a chess world champion type of chess player) are:
Code:
HAM 2800
VER 2788
NOR 2643
BOT 2634
PER 2607
LEC 2525
SAI 2496
RIC 2443According to those ratings, Norris has aprox a 75% winning chance against Ricciardo (that is a 3-1 winning ratio). The real data between them is even crueler: 9-1 . It shows the 9-1 score is a bit bad luck for Ricciardo in comparison to how both of them perform against other competitors. I.e. without bad luck, Ricciardo would have 2-8 or 3-7 versus Norris.
Rating difference between teammates is minimal in Ferrari but it is around 150-200 points in the other teams. It means typically the strongest teammate beats the weakest one 70%-75% of times.