I will use 3 examples to show that you can predict championship results as long as nobody disturbs any race result (as in Valencia 2006, Sepang 2015, Valencia 2015). Yes, I mentioned 2015 but keep your opinions to yourselves, nobody cares about what happened anymore. I will be using 2006, 2013 and 2015 for championship results. I picked these seasons because of the following reasons:
- 2006 had a race result disturbed by a crash in the last race
- 2013 was close in the points but the title was predictable
- 2015 was close in points and the title was predictable as long as there was no disturbance in the fight
Measures
I made an observation that title contenders are not supposed to be outside the points and taking that absence of points to any equation does not reflect the reality. Also we should not believe standalone best and worst results in points, because shit happens and luck is a thing for some riders.
My equation
FMWC which stand for Fraud Magic Winning Coeficient = (total_points - best_result - worst_result_inside_points) / (races_finished - 2)
2006 before last race
- Hayden had a FMWC of 15.692
- Rossi had a FMWC of 24.11
2006 after the last race
- Hayden had a FMWC of 15.71
- Rossi had a FMWC of 18.33
Measurement
Using FMWC you can notice that Rossi/Hayden difference in performance after the last race was 14.3%. In the championship it was 1.98%. Yes, Hayden won but Rossi was the better rider during the championship, by a margin of 14.3%. Shit happens, so he lost.
2013 before the last race
- marquez had a FMWC of 19.79. I'm counting his DSQ race as a finished race because it was his fault to miss the correct lap for the bike swap, he didn't crash.
- Lorenzo had a FMWC of 19.36.
Lorenzo had 8 wins, marquez had 6 wins, both had one DNF/DNS. Even before the race you could predict that the race would be close, but marquez was ahead. At that moment before the race, their difference in FMWC was 2.17%. For Lorenzo to win the championship, he would need to perform >2.17% better than marquez in Valencia result, which means he would need to finish 1'11''056 ahead of marquez, if he couldn't do that, by marquez performance history in the season he would win the title. By the end of the race, he finished only 7 seconds behind Lorenzo. Yes, Lorenzo strategy to make traffic for marquez to get overtaken could work but it wasn't safe, and it didn't pay off. He fucked up, his best chance was to pull away and let marquez fight with other riders and luckily crash or waste his tires and fade away in the race. He picked the wrong strategy, because the safest strategy was that, pull away and let the guy fight with others, because he was able to run away instead of making traffic and letting the guy keep his tires good to follow him.
2013 after the last race
The difference was just 4 points, but the FMWC was 19.53 for marquez and 19.73 for Lorenzo. The reason marquez won was because he could keep his FMWC below what Lorenzo needed to win. Lorenzo needed to be 2.17% better, but marquez managed to keep him only 1.01% better, so he lost. Lorenzo finished the championship as the best performance, but if you go back and read the "before the race", you will see that marquez was already ahead.
2015 and why Lorenzo would've won just like he lost 2013, even the same fucking points
I will make 2 examples, one excluding Sepang results as god is my witness and I swear bottles will start flying if someone says it wasn't the right thing to do, and another one including the results.
2015 before Valencia, without Sepang
- Lorenzo had a FMWC of 19.15
- Rossi had a FMWC of 18.57
2015 before Valencia, including Sepang
- Lorenzo had a FMWC of 21.0
- Rossi had FMWC of 18.4
It looks like Sepang gave Lorenzo a big advantage but we know what happened, the race was disturbed, so at this moment you would think that Lorenzo was being 12.38% better but in fact he was just 3% better as you can see in the FMWC excluding Sepang. The same result is reflected in the points after Valencia, a difference of only 1.52% and not >10%.
Will marquez win 2019?
Yes, as long as he doesn't get injured and miss more than 3 races.
What else did I predict for 2019 with my equation?
- Championship will be marquez > Rins > Dovi >= Petrucci > Rossi > Quartararo
- Quartararo will be the top satellite and rookie of the year
- Pol will finish ahead of Nakagami in the championship
- Aleix will finish ahead of Lorenzo, Mir, Zarco, Iannone
- Oliveira won't beat Iannone in the standings
- Rabat will beat Bagnaia
- Lorenzo will be outside top 15 in Assen, he might even DNF
- Lorenzo best result of the season will be in Brno or Austria
Conclusion is that you can predict race results with high accuracy even with unpredictable individual results when you create a model that removes great and shitty results when it's not normal but using the same weight for all riders at the same time. After 2 months and a half I admit that it is safe to say that if you want to compare riders, you can pick only the races where all of your selected riders finished and look by the time and not position.
I will use 3 examples to show that you can predict championship results as long as nobody disturbs any race result (as in Valencia 2006, Sepang 2015, Valencia 2015). Yes, I mentioned 2015 but keep your opinions to yourselves, nobody cares about what happened anymore. I will be using 2006, 2013 and 2015 for championship results. I picked these seasons because of the following reasons:2006 had a race result disturbed by a crash in the last race2013 was close in the points but the title was predictable2015 was close in points and the title was predictable as long as there was no disturbance in the fightMeasuresI made an observation that title contenders are not supposed to be outside the points and taking that absence of points to any equation does not reflect the reality. Also we should not believe standalone best and worst results in points, because shit happens and luck is a thing for some riders.My equationFMWC which stand for Fraud Magic Winning Coeficient = (total_points - best_result - worst_result_inside_points) / (races_finished - 2)2006 before last raceHayden had a FMWC of 15.692Rossi had a FMWC of 24.112006 after the last raceHayden had a FMWC of 15.71Rossi had a FMWC of 18.33MeasurementUsing FMWC you can notice that Rossi/Hayden difference in performance after the last race was 14.3%. In the championship it was 1.98%. Yes, Hayden won but Rossi was the better rider during the championship, by a margin of 14.3%. Shit happens, so he lost.2013 before the last racemarquez had a FMWC of 19.79. I'm counting his DSQ race as a finished race because it was his fault to miss the correct lap for the bike swap, he didn't crash.Lorenzo had a FMWC of 19.36.Lorenzo had 8 wins, marquez had 6 wins, both had one DNF/DNS. Even before the race you could predict that the race would be close, but marquez was ahead. At that moment before the race, their difference in FMWC was 2.17%. For Lorenzo to win the championship, he would need to perform >2.17% better than marquez in Valencia result, which means he would need to finish 1'11''056 ahead of marquez, if he couldn't do that, by marquez performance history in the season he would win the title. By the end of the race, he finished only 7 seconds behind Lorenzo. Yes, Lorenzo strategy to make traffic for marquez to get overtaken could work but it wasn't safe, and it didn't pay off. He fucked up, his best chance was to pull away and let marquez fight with other riders and luckily crash or waste his tires and fade away in the race. He picked the wrong strategy, because the safest strategy was that, pull away and let the guy fight with others, because he was able to run away instead of making traffic and letting the guy keep his tires good to follow him.2013 after the last raceThe difference was just 4 points, but the FMWC was 19.53 for marquez and 19.73 for Lorenzo. The reason marquez won was because he could keep his FMWC below what Lorenzo needed to win. Lorenzo needed to be 2.17% better, but marquez managed to keep him only 1.01% better, so he lost. Lorenzo finished the championship as the best performance, but if you go back and read the "before the race", you will see that marquez was already ahead.2015 and why Lorenzo would've won just like he lost 2013, even the same fucking pointsI will make 2 examples, one excluding Sepang results as god is my witness and I swear bottles will start flying if someone says it wasn't the right thing to do, and another one including the results.2015 before Valencia, without SepangLorenzo had a FMWC of 19.15Rossi had a FMWC of 18.572015 before Valencia, including SepangLorenzo had a FMWC of 21.0Rossi had FMWC of 18.4It looks like Sepang gave Lorenzo a big advantage but we know what happened, the race was disturbed, so at this moment you would think that Lorenzo was being 12.38% better but in fact he was just 3% better as you can see in the FMWC excluding Sepang. The same result is reflected in the points after Valencia, a difference of only 1.52% and not >10%.Will marquez win 2019?Yes, as long as he doesn't get injured and miss more than 3 races.What else did I predict for 2019 with my equation?Championship will be marquez > Rins > Dovi >= Petrucci > Rossi > QuartararoQuartararo will be the top satellite and rookie of the yearPol will finish ahead of Nakagami in the championshipAleix will finish ahead of Lorenzo, Mir, Zarco, IannoneOliveira won't beat Iannone in the standingsRabat will beat BagnaiaLorenzo will be outside top 15 in Assen, he might even DNFLorenzo best result of the season will be in Brno or AustriaConclusion is that you can predict race results with high accuracy even with unpredictable individual results when you create a model that removes great and shitty results when it's not normal but using the same weight for all riders at the same time. After 2 months and a half I admit that it is safe to say that if you want to compare riders, you can pick only the races where all of your selected riders finished and look by the time and not position. http://ifttt.com/images/no_image_card.png
No comments:
Post a Comment