VeloPower

Fact based wielertraining

[VeloPower @VeloPower_NL]

Let’s see how far this will go: https://t.co/9bGI7brnHR

[VeloPower @VeloPower_NL]

@svpino More resources on this topic: https://t.co/ejrNYlDbe7

[VeloPower @VeloPower_NL]

If you like (Dutch) cycling podcasts maybe my series will be for you: https://t.co/dVtI7QgoKy

[VeloPower @VeloPower_NL]

Good vibe today. ChatGPT discovered my better self. Would have signed for this. Circuit des Mines 🤔

[VeloPower @VeloPower_NL]

@ChristophMolnar pip install deepmultilingualpunctuation

[VeloPower @VeloPower_NL]

@jem_arnold Very nice reading list thank you for sharing

[VeloPower @VeloPower_NL]

Playing with whoop data. Next up merge alle other data sources

[VeloPower @VeloPower_NL]

@JeroenSwart Do you differentiate between rider types (depending on their role in the race/ contender type) on the Fatigued MMP values and are they also extracted from index sessions you describe?

[VeloPower @VeloPower_NL]

@SeanSeale @EliasLehtonen Thank you for answering my question in your podcast! Very nice it got mentioned. I should have added context though “age grouper, gran fondo cycling events (duration 4-6h)” Sorry for that

[VeloPower @VeloPower_NL]

@SeanSeale @EliasLehtonen What would be good training interventions for people who’s LT1 is relatively close to LT2 (LT1 power is at 90% of LT2 power) and power around LT2 is at 89% of 6 min MMP (“VO2max power”)? So the range “on the right” seems “small”? Note LT1 is already rather well developed

[VeloPower @VeloPower_NL]

@mervenoyann @christopher Maybe create something like I did so it will solve itself 😂 https://t.co/HaARGl90zO

[VeloPower @VeloPower_NL]

@EliasLehtonen Nice to see the alignment. One question. Isn’t a timpoint of approx 20 min to estimate CP a bit too long? I see it is rowing (I come from cycling), so don’t know if that’s common

[VeloPower @VeloPower_NL]

@altini_marco @ouraring Good to read!! However, everytime I read the word “epoch” I relate it too neural nets. I guess it’s “vakdeformatie” 😉

[VeloPower @VeloPower_NL]

@tunguz I have the same when using XGBoost

[VeloPower @VeloPower_NL]

@timpodlogar @Spragg_Perform @peter__leo @UBSportExR @spragg247 @japplphysiol CP and W’ yes. Think this also resonates more into rider type typologies (sprinter, etc.) / fiber type relations which I find effects of training modalities to be very interesting (coz they usually lead to different effects).

[VeloPower @VeloPower_NL]

@edarchimbaud Good overview. Maybe also include Prodigy from @explosion_ai ?

[VeloPower @VeloPower_NL]

@akreutzer82 My 2 cents. When looking at modelling these kinds of relations you can also look into Structural Equation models. The age, sex kind of variables would according to me be more different versions for the same structural model but with different parameter values.

[VeloPower @VeloPower_NL]

@MaartenvSmeden Polychoric and tetrachoric correlation

[VeloPower @VeloPower_NL]

@abhi1thakur Today. It’s almost inevitable if you want to keep progress on a project since most people see it as a burden. To me it’s valueable and in the end rewarding if you see it bumps performance more than tuning parameters on a smaller dataset

[VeloPower @VeloPower_NL]

@Spragg_Perform Heel goed 😉