W/kg ranges corresponding to PPA road seeding

[s14phoenix]

Slow week sadly.

 

Maybe wouldn't be so slow if less watts and more aero????  

[pe3nguin]

Maybe wouldn't be so slow if less watts and more aero????  

Oi. Low blow

[s14phoenix]

Oi. Low blow

 

high socks... not low...  

[Pure Savage]

Slow week sadly.

Tell me about it

[Patchelicious]

http://i0.kym-cdn.com/entries/icons/facebook/000/003/285/ike_where_this_thread_is_going-vi.jpg

I bet if I say "that boat looks like a crappy Noah's Ark", that the thread will get closed....

[pe3nguin]

I bet if say that boat looks like a crappy Noah's Ark, that the thread will get closed.... 

And there goes my coffee and keyboard

[Dubber]

I think the hard thing is we are all using different weights to determine our FTP. Weight with the bike, weight without. Are all watts equal? How are we suppose to reach a consensus when the parameters aren't established 

And what if the rider was doing their test on an MTB with a vertical shock - surely that gives at least 7 watts extra?

[pe3nguin]

And what if the rider was doing their test on an MTB with a vertical shock - surely that gives at least 7 watts extra?

Don't know anyone who rides MTB, but sure I guess we could give them a few charity watts.

[Fat Boab]

And what if the rider was doing their test on an MTB with a vertical shock - surely that gives at least 7 watts extra?

 

Don't be absurd. But their mass might be a little less so W/kg definitely would increase......

[thisismyotherbike]

 

Plot of the usable data.

 

The fitted curve is y = 621.78*e^(-1.074x)

 

There are curves that produce a higher R², but I prefer this one because it provides me with more hope   

 

At the risk of setting off another rant from one of the FTP usefulness skeptics, it seems I  need to improve my W/kg by 30% if I'd like to get to A batch    

[Fat Boab]

 

Plot of the usable data.

 

The fitted curve is y = 621.78*e^(-1.074x)

 

There are curves that produce a higher R², but I prefer this one because it provides me with more hope   

 

At the risk of setting off another rant from one of the FTP usefulness skeptics, it seems I  need to improve my W/kg by 30% if I'd like to get to A batch    

 

Nice to see a scientific approach paying off! What's the R^2 value?

[Jehosefat]

 

Plot of the usable data.

 

The fitted curve is y = 621.78*e^(-1.074x)

 

There are curves that produce a higher R², but I prefer this one because it provides me with more hope   

 

At the risk of setting off another rant from one of the FTP usefulness skeptics, it seems I  need to improve my W/kg by 30% if I'd like to get to A batch    

 

I would argue that R² is not the correct target here. Should rather try to maximise rank ordering so use something like Kendall's-Tau?

[Fat Boab]

I would argue that R² is not the correct target here. Should rather try to maximise rank ordering so use something like Kendall's-Tau?

 

Do expand?

[Karman de Lange]

 

Plot of the usable data.

 

The fitted curve is y = 621.78*e^(-1.074x)

 

There are curves that produce a higher R², but I prefer this one because it provides me with more hope   

 

At the risk of setting off another rant from one of the FTP usefulness skeptics, it seems I  need to improve my W/kg by 30% if I'd like to get to A batch    

 

Just to skew your stats a bit, but its not that far off

 

Did FTP (#4dp) on Saturday, did one tonner on Sunday.  Results:

 

Index 19.85 :  2.7w/kg

 

Index before one tonner:  24.29 at 2.8w/kg 

[Jehosefat]

Do expand?

 

R² is a measure of total explained variance (iirc) whereas what we are particularly interested in here is how well the ranking of FTP corresponds to the ranking of seeding because the seeding index itself is already designed as a relative ranking system.

 

Alternatively we should use some sort of classifier algorithm if we want to solve the OPs question of "What range of FTP corresponds to each letter rank as per the PPA seeding?" (Although I don;t think we have anywhere near enough data for this to be a robust option at the moment)

[Fat Boab]

R² is a measure of total explained variance (iirc) whereas what we are particularly interested in here is how well the ranking of FTP corresponds to the ranking of seeding because the seeding index itself is already designed as a relative ranking system.

 

Alternatively we should use some sort of classifier algorithm if we want to solve the OPs question of "What range of FTP corresponds to each letter rank as per the PPA seeding?" (Although I don;t think we have anywhere near enough data for this to be a robust option at the moment)

 

Ta muchly. Everyday is a school day!