- 16 May 10, 07:01#199419
For the past 12 months I have been working on a system of evaluating F1 driver performance.
I have worked for many years within a government statistical department gathering data, ensuring its fitness for purpose, publishing output and interpreting the direction of various industries for both UK and EU government.
The F1 model developed provides a statistical reference point for the level of evidence relating to an individual driver’s performance. Unlike human evaluation it will not make rash predictions relating to driver standards unless there is sufficient evidence to support a claim. i.e. it will not predict a Michael Schumacher or Ayrton Senna after only one race. On the opposite side of the coin it will look at all recent performances and give a rating based on a driver’s more recent results and also a record of the peak rating a driver achieves throughout his/her career. Of course the skills necessary may differ from era to era but each driver performs according to the rules and technology that prevail at the time, and the rating is a measure of how he/she performed against his or her peers given the current technology and rules.
For those of you that understand how an ELO rating system works the output from this methodology behaves in a very similar manner with some exceptions, although I cannot stress highly enough that these are not ELO ratings. E.g ELO ratings do not decline due to inactivity on the part of the rated subject but these ratings would reflect such inactivity. See http://en.wikipedia.org/wiki/Elo_rating ... al_details for an explanation of ELO rating methodology.
Formula 1 is my passion and I am content that the level of accuracy I have now achieved is as high as possible given the limited time and resources available to me as a personal hobby.
I am looking for one of four things.
a) Someone who is willing to invest in further refinement in return for a proportion of rights to future outputs from the system.
b) Someone who can advise me on marketing such data and how I might achieve a reasonable return on the time I have invested to date.
c) An established publisher or author in motorsport who would be interesting in publishing my outputs either in their raw form or as a collaboration in book form.
d) An individual or organisation who is interested in purchasing the exclusive rights to any current or future outputs that are generated from the methodology developed.
I do not wish to go into greater detail here regarding the outputs but to give some idea of the quality of those data being produced the following are some examples:
a) In the first year of formula 1, 1950, with no reference to preceding or subsequent GP data. the system evaluated those results achieved by Juan Manuel Fangio as being superior to that achieved by Nino Farina, despite Farina winning the title.
b) By 1958 Stirling Moss was rated as the top driver at 2686, while Mike Hawthorn in winning the title was at 2638. Most experts would of course agree with this evaluation of each driver given the full access to the history of each driver.
c) In 1962 Graham Hill took the World Title and at that point was rated at 2638. Jim Clark on the other hand was already rated as the best driver in the world at 2670 despite having not yet won either of his titles.
d) The same could be said for Alain Prost in 1984. When Niki Lauda won his 3rd World Title with a rating of 2636, Alain Prost was already achieving the current top rating of 2672, again before he had ever won the title. It is interesting to note that the rating being achieved by Lauda at this point in his career was much lower than his peak in the mid 1970s.
e) Lewis Hamilton's rating stood at 2648 at the end of 2008, his championship winning year. At the end of 2009 in a car that was far less competitive his rating remained at 2646, a difference of only 2 points demonstrating the reduction in the competiveness of his car did not deflate his rating.
Now some people may well say that we know all of the above already and many knowledgeable pundits would have said the same as that being identified by the rating methodology in the above cases. However, I would ask you to consider one point, and that is the opinion of experts is just that. Opinion. The above ratings have been calculated and not influenced by opinion and can be said to be truly objective as the same methodology for evaluating performance has been applied to every driver on the grid at the point in time the rating was being calculated. It is the fact that many of the ratings that are calculated coincide with expert opinion despite their objectivity and methodology that strengthens the case for the validity of the output.
In support of providing further evidence and to achieve one of my four objectives highlighted above, I am here making a limted level of output available to potential investors via Ebay in the form of laminated summary sheets at a minimal charge. These are being made available under the heading of ‘Formula 1 – Statistical Summary Sheets’ and are being sold by the seller <!-- Edited for Content -->. I hope you will all understand that I do not wish to make public all my findings here as to do so would destroy the value of my work. These sheets will be subject to copyright
1. They will identify all drivers since the inception of F1 achieving in excess of a 2600 rating and showing their respective peaks and when achieved. 2600 is the mean rating of all drivers ranked 10th in the world throughout the history of F1, while 2700 is the mean rating of drivers ranked at #1.
2. It will also include current performance ratings of drivers in excess of 2600 who are still active in F1 up to including the most recent GP at time of purchase.
3. It will also include more basic data like total # of F1 GPs won, total # of Fastest Laps and Pole Positions for each driver shown.
4. Finally it will show ‘GP years’ which is a crude measure of a driver’s success assuming that the # of GPs per annum had not been subject to the increase we have witnessed over the years. Each output would come with a page of explanatory notes so that data may be examined and interpreted correctly.
5. The world title in F1 is not a perfect barometer of talent and each sheet will be colour coordinated showing the drivers success in world title terms compared with his rating success.
Should anyone wish to discuss any of the four objectives identified above or wish to ask questions with investment in mind I would only be too happy to hear from you at <-- Edited for Content -->. However, please do not expect a quick response as I am currently working very long days and I am unable to check my emails on a daily basis.
Regards
Tim
I have worked for many years within a government statistical department gathering data, ensuring its fitness for purpose, publishing output and interpreting the direction of various industries for both UK and EU government.
The F1 model developed provides a statistical reference point for the level of evidence relating to an individual driver’s performance. Unlike human evaluation it will not make rash predictions relating to driver standards unless there is sufficient evidence to support a claim. i.e. it will not predict a Michael Schumacher or Ayrton Senna after only one race. On the opposite side of the coin it will look at all recent performances and give a rating based on a driver’s more recent results and also a record of the peak rating a driver achieves throughout his/her career. Of course the skills necessary may differ from era to era but each driver performs according to the rules and technology that prevail at the time, and the rating is a measure of how he/she performed against his or her peers given the current technology and rules.
For those of you that understand how an ELO rating system works the output from this methodology behaves in a very similar manner with some exceptions, although I cannot stress highly enough that these are not ELO ratings. E.g ELO ratings do not decline due to inactivity on the part of the rated subject but these ratings would reflect such inactivity. See http://en.wikipedia.org/wiki/Elo_rating ... al_details for an explanation of ELO rating methodology.
Formula 1 is my passion and I am content that the level of accuracy I have now achieved is as high as possible given the limited time and resources available to me as a personal hobby.
I am looking for one of four things.
a) Someone who is willing to invest in further refinement in return for a proportion of rights to future outputs from the system.
b) Someone who can advise me on marketing such data and how I might achieve a reasonable return on the time I have invested to date.
c) An established publisher or author in motorsport who would be interesting in publishing my outputs either in their raw form or as a collaboration in book form.
d) An individual or organisation who is interested in purchasing the exclusive rights to any current or future outputs that are generated from the methodology developed.
I do not wish to go into greater detail here regarding the outputs but to give some idea of the quality of those data being produced the following are some examples:
a) In the first year of formula 1, 1950, with no reference to preceding or subsequent GP data. the system evaluated those results achieved by Juan Manuel Fangio as being superior to that achieved by Nino Farina, despite Farina winning the title.
b) By 1958 Stirling Moss was rated as the top driver at 2686, while Mike Hawthorn in winning the title was at 2638. Most experts would of course agree with this evaluation of each driver given the full access to the history of each driver.
c) In 1962 Graham Hill took the World Title and at that point was rated at 2638. Jim Clark on the other hand was already rated as the best driver in the world at 2670 despite having not yet won either of his titles.
d) The same could be said for Alain Prost in 1984. When Niki Lauda won his 3rd World Title with a rating of 2636, Alain Prost was already achieving the current top rating of 2672, again before he had ever won the title. It is interesting to note that the rating being achieved by Lauda at this point in his career was much lower than his peak in the mid 1970s.
e) Lewis Hamilton's rating stood at 2648 at the end of 2008, his championship winning year. At the end of 2009 in a car that was far less competitive his rating remained at 2646, a difference of only 2 points demonstrating the reduction in the competiveness of his car did not deflate his rating.
Now some people may well say that we know all of the above already and many knowledgeable pundits would have said the same as that being identified by the rating methodology in the above cases. However, I would ask you to consider one point, and that is the opinion of experts is just that. Opinion. The above ratings have been calculated and not influenced by opinion and can be said to be truly objective as the same methodology for evaluating performance has been applied to every driver on the grid at the point in time the rating was being calculated. It is the fact that many of the ratings that are calculated coincide with expert opinion despite their objectivity and methodology that strengthens the case for the validity of the output.
In support of providing further evidence and to achieve one of my four objectives highlighted above, I am here making a limted level of output available to potential investors via Ebay in the form of laminated summary sheets at a minimal charge. These are being made available under the heading of ‘Formula 1 – Statistical Summary Sheets’ and are being sold by the seller <!-- Edited for Content -->. I hope you will all understand that I do not wish to make public all my findings here as to do so would destroy the value of my work. These sheets will be subject to copyright
1. They will identify all drivers since the inception of F1 achieving in excess of a 2600 rating and showing their respective peaks and when achieved. 2600 is the mean rating of all drivers ranked 10th in the world throughout the history of F1, while 2700 is the mean rating of drivers ranked at #1.
2. It will also include current performance ratings of drivers in excess of 2600 who are still active in F1 up to including the most recent GP at time of purchase.
3. It will also include more basic data like total # of F1 GPs won, total # of Fastest Laps and Pole Positions for each driver shown.
4. Finally it will show ‘GP years’ which is a crude measure of a driver’s success assuming that the # of GPs per annum had not been subject to the increase we have witnessed over the years. Each output would come with a page of explanatory notes so that data may be examined and interpreted correctly.
5. The world title in F1 is not a perfect barometer of talent and each sheet will be colour coordinated showing the drivers success in world title terms compared with his rating success.
Should anyone wish to discuss any of the four objectives identified above or wish to ask questions with investment in mind I would only be too happy to hear from you at <-- Edited for Content -->. However, please do not expect a quick response as I am currently working very long days and I am unable to check my emails on a daily basis.
Regards
Tim
Last edited by myownalias on 16 May 10, 07:29, edited 2 times in total.Reason: Removed Email address; will pass this post onto Admins for further evaluation.
