STRAVA as an cycling anti - doping tool and STRAVA as an global cycling talent identification tool

Hi STRAVA users (best application in the world :-))  

I have two ideas worth spreading. 

First - statistical evaluation of longitudinal STRAVA cycling power data, like blood and steroid data of profesional athletes in Athlete's biological passports designed by WADA, could detect and reveal suspicious (=not the results of training and growth, but the consequence of use of prohibited substances or methods) increases in cycling power data of recreational and elite athletes. 

Statistical evaluation of big data of anonimous cycling power data of global cycling comunity could reveal what is "normal" and what "abnormal" longitudinal cycling power progresion due to training and growing or aging. Longitudinal evaluation of cycling power could also detect disease or overtraining.  

Second - statistical evaluation of longitudinal STRAVA cycling power data could reveal individuals with above average cycling power progression and in this way identify and inform global talents (new category beside PB and KOM also GT18 - global first hundret talent list under 18 years old...) to persue sport in which they could have a bright future. Do not let any talented kid be left behind...

What is your opinion?

By Mitja 








  • It would be cool if you could use Strava to track things like this, but the simple fact is, Strava can't even figure out how to identify an e-bike ride or someone driving in a car compared to a regular bike ride.  If their algorithms can't tell the difference between someone riding a bike at 10mph up an incline and another person driving it at 60+mph, there is no way in heck they will be able to identify minute differences in average power of a rider. 

    評論操作 永久連結
  • Dear Mr. Jason, 

    While I agree that Strava algorithms still have room for impruvement, their tehnology already today is not completely useless. The doping case of french recreational cyclists Mikael Gallego with 367 KOMs told me that. Why?

    One rider wrote:"I'm the same weight as Micka and yet he's just put 2'15" into me on the Montaud hill climb, that's normal, Strava shows he's putting out 45 watts more than me but I'd really like to know what his training is because our CP20 on the Col de Porte in mid May was the same as mine and I've only managed to scrape an extra 5 watts over the summer". 

    This anonimous rider coment shows, that power comperison from the same starting poin (= same CP20 in mid May) of two riders on the same hill in the same period of the season revealed 45 W progresion of one rider over the summer. When routine control by the French Anti Doping agency revealed a cocktail of drugs in the urine of the cyclist Mikael Gallego, including corticoïdes, morphine and testosterone, this shows us how cycling power data could be used to distinguish clean from doped cyclists... 

    If a model of critical power (CP), (for additional reading please go to: on STRAVA big data  was developed and used, the sensitivity and accurency of model could be validated on historical longitudinal cyling power data from past riders cought doping through other methods... In this mener, with the use of maschine learning tehniks, such model could be greately impruved.   

    Last but not least, STRAVA data can also be used as a location proof (profile of the terain can be verified) for comparing locations of training with location data stated in profesional riders ADAMS file. Remember this was the case of doping suspension of cyclist Michael Rasmussen when he was in yelow and leading the Tour de france race in year 2007...  (location in ADAMS file and true location of training before the Tour was not the same).

    We can go even further... big differences between power data achieved between training and competition could rise red flag of suspection... Antonie Vayer, former Festina trainer and coach of more than 500 athletes including three world champions, established a fixed trashold of cycling power, measured indirectly, for three week races, that is indicative of doping... that treshold power is 430 W for 78 kg standard weight of a cyclist (for additional explanation please see : Historical analysis of positive doping cases confirmed that his cut off value was in 88 % of the cases correct...(while sensitivity of his method is high, accurency of doping susiction is not known, or in other words, false positive ratio of his method is not known). Despite this fact, a cut off value of cycling power could be easily applied on live STRAVA cycling power data and the antidoping authorities could be instantly informed of suspicious recretational or professional rider(s) in a given state where such measurements were recorded...  Remeber, cyclists mus train a lot, so there is a lot of training data to analyse...   

    If enyone has additional comment, please join the discused topic.



    評論操作 永久連結