P in the Pool

Kurt Godel

Sat Nothing, the day got away from me.

Sun 29k bike; 2hr turbo

After a LOT of coaxing the will-we/won't-we kids decided they would run the Tinahely duathlon, allowing me to do the bike leg. A good day, the rain held off, easy to get to and hassle free. Kid A ran the 5k, setting off too fast before twisting an ankle (his shoes weren't great), but finished the 5k in 32min, pleased as punch. There was only one bike left racked when I set off, so I had targets all the way around. I decided to do the route in a higher cadence, and see what happened. Felt good climbing to 5k, but descended nervously on the slick roads and got the anticipated speed wobble... managed to control it but time lost here. I had fiddled with my aero position beforehand, and it felt a lot more comfortable. Again in the small rings when climbing to Coolattin, and again a nervous descent. I tried to push hard over the last flat 5k, but missed out on my hoped for sub 35 (in 35:10). Tagged Kid B who ran the final 2.8k with gusto in 15:36. A good family race, and good fun. All were happy and talking about it when we went for a lunch treat.

I expected to go a minute faster, if I'm honest. Its a couple of seconds slower than a TT I did last month, which I can only put down to the cadence (69rpm then, 83 today). I worked harder last month (295W then vs. 274 today), but there wasn't much difference in overall speed (33.6kph then vs. 33.5 today). This suggests to me that I was more aero today, and that I should probably target a cadence in the mid 70's, for greater power.

Later I did an easy 2 hr on the turbo (164W, 67rpm) to round out the day.

career move

You bribed them with food .... cunning

Kurt Godel

career move wrote: »

You bribed them with food .... cunning

Kid A: was on €5 for every minute under 35mins,
Kid B: was on €5 for every minute under 18mins.

Expensive duathlon for me!

mossym

Kurt Godel wrote: »

I expected to go a minute faster, if I'm honest. Its a couple of seconds slower than a TT I did last month, which I can only put down to the cadence (69rpm then, 83 today). I worked harder last month (295W then vs. 274 today), but there wasn't much difference in overall speed (33.6kph then vs. 33.5 today). This suggests to me that I was more aero today, and that I should probably target a cadence in the mid 70's, for greater power.

.

you sure it's the cadence kurt? course today, same profile, same headwinds as your previous tt you are comparing to?

tunney

Kurt Godel wrote: »

I expected to go a minute faster, if I'm honest. Its a couple of seconds slower than a TT I did last month, which I can only put down to the cadence (69rpm then, 83 today). I worked harder last month (295W then vs. 274 today), but there wasn't much difference in overall speed (33.6kph then vs. 33.5 today). This suggests to me that I was more aero today, and that I should probably target a cadence in the mid 70's, for greater power.

so you looked at the different cadence and power figures. And then decided it was the higher cadence that caused the slower bike time and not the lower power. ...........

Of course the fact you didn't wear a chest strap today makes definitive statements difficult but. ... could the slower time be linked to the lower power? Which could be caused by lead motivation and not working as hard?

I do get your point though. You are not yet ready to generate race wattage at a more optimal cadence.

Two choices :
1) treat each leg individually and worry about the 10-30 seconds on the bike lost
2) consider it as a swim bike ruin and realise that you could save minutes pb the run. ....

Kurt Godel

mossym wrote: »

you sure it's the cadence kurt? course today, same profile, same headwinds as your previous tt you are comparing to?

Same course, similar winds, different slickness on the roads. But, no I'm not sure its cadence, just a hunch as its the biggest difference.

tunney wrote: »

so you looked at the different cadence and power figures. And then decided it was the higher cadence that caused the slower bike time and not the lower power. ...........

I'm presuming the higher cadence caused lower power. It certainly felt as though I was working very hard, so I was surprised to see lower power today. I was aero a lot more of the time which I'd attribute to similar overall speeds.

tunney wrote: »

Of course the fact you didn't wear a chest strap today makes definitive statements difficult but. ... could the slower time be linked to the lower power? Which could be caused by lead motivation and not working as hard?

I should have worn the HR monitor, there's a general cold doing the rounds in the house and I didn't sleep well. But yes, motivation wasn't great, it didn't feel like a race.

tunney wrote: »

I do get your point though. You are not yet ready to generate race wattage at a more optimal cadence.

Two choices :
1) treat each leg individually and worry about the 10-30 seconds on the bike lost
2) consider it as a swim bike ruin and realise that you could save minutes pb the run. ....

I take the point that grinding at 50rpm may lead to high wattage, but tired legs for the run, and I'm more than willing to give the higher cadence a go. Look forward to being able to test the run theory!

Kurt Godel

Mon 10min run; 2,600m Masters swim; 1 hr turbo

Run:
First run in ages, 10min at 5:16 pace. Very nervous that something would happen, so I can't say for sure that the heel felt good- hyper aware of every footstep. Anyway, got through the run without issue, heel feels good now (feels looser than before).

Swim:
200m warm up
6*50 (25 kick, 25 drill, 10s)
200m easy BE3
Main set:
2*200 on 3:30 (in 2:55)
90s rest
4*100 on 1:45 (in 1:25)
90s rest
8*50 alternate on 40s, 60s
90s rest
16*25 alternate on 20s, 30s
90s rest
350 warm down

Really enjoyed this session, we managed to get a lane at noon, first time going short and fast in a while. I hit all the targets, felt good in the water.

Turbo:
Easy hour to end off the day. Didn't push the cadence and just kept to what felt comfortable. (157W, 62rpm)

Kurt Godel

Tue 60 min turbo

20 easy
15x(1 spin, 1 easy)
10 easy
spin is > 95rpm

The 15*1min came in between 99-103rpm, and 306-372W. The 372W was in a higher gear, and was a bit of a stretch. Otherwise they were all solid enough, I'm getting a better understanding of my own personal relationship between power and cadence.

I actually forgot I had a bad achilles today, which is a very very good sign.

Dory Dory

Being a mathematician, I'd think the correlation between power and cadence would be a kick for you.

tunney

Dory Dory wrote: »

Being a mathematician, I'd think the correlation between power and cadence would be a kick for you.

Wait until he sees scatter graphing for power analysis......

Dory Dory

tunney wrote: »

Wait until he sees scatter graphing for power analysis......

Oh god. That might be "X" rated!!

Kurt Godel

three variables I'd imagine for any correlation analysis; power, cadence, and HR

tunney

Kurt Godel wrote: »

three variables I'd imagine for any correlation analysis; power, cadence, and HR

If including HR then you need temperature and a whole lot more!

Oryx

This is why I dont use power stats*. Toooo manyyy numberrrs...




*its probably also why I am broken.

Kurt Godel

tunney wrote: »

If including HR then you need temperature and a whole lot more!

*Geek alert*

Probably only Tunney and Dory are still reading at this point, so...

I have one exam left for my Math degree, and am studying for it over the next month. Its Linear Statistical Modelling, which is involved in attempting to create models which help explain or predict relationships between variables. The response variable in our proposed Model is Power- its the goal, its the variable which is dependent on one or more explanatory variables. In effect, what causes more or less power? Effort (HR), cadence, sex, age, BMI, temp, crank rings, etc, etc. The more data you have for the this, the better you can test for response in Power. And then you can test what happens when you change one explanatory variable, while keeping the rest the same (ie what happens when all other explanatories are kept the same, but cadence is changed?). And then you can discard some of the explanatories and create an optimal scientific model.

I'll probably go nowhere further with this, but it may help my studies to think of a Linear Statistical Model in terms of something personally interesting to me, so perhaps we can tease something out over the next few weeks.

Dory Dory

Kurt Godel wrote: »

*Geek alert*

Probably only Tunney and Dory are still reading at this point, so...

I have one exam left for my Math degree, and am studying for it over the next month. Its Linear Statistical Modelling, which is involved in attempting to create models which help explain or predict relationships between variables. The response variable in our proposed Model is Power- its the goal, its the variable which is dependent on one or more explanatory variables. In effect, what causes more or less power? Effort (HR), cadence, sex, age, BMI, temp, crank rings, etc, etc. The more data you have for the this, the better you can test for response in Power. And then you can test what happens when you change one explanatory variable, while keeping the rest the same (ie what happens when all other explanatories are kept the same, but cadence is changed?). And then you can discard some of the explanatories and create an optimal scientific model.

I'll probably go nowhere further with this, but it may help my studies to think of a Linear Statistical Model in terms of something personally interesting to me, so perhaps we can tease something out over the next few weeks.

You're not asking us to do your homework for you, are you? :)

pgibbo

Are you considering/tracking gearing at all when you're thinking about your cadence? I'm assuming based on the numbers you're achieving the higher cadence by being in a lower gear and it's the spinning this gear that's leading to a higher RPE.

Kurt Godel

Wed 3,000m swim

200sw,200p,100k
20x100sw on 1:55 (all the 100s +/- 3 seconds)
4x50 fast on 60
4x25 fast on 40
200 choice easy

I wasn't sure what sort of session this was supposed to be, hard or easy? Was told it was my call so decided it was an exercise in pacing, the important part was the +/- 3 seconds. They felt very solid, well rested meant they were all in between 1:37-1:42, and I found I could control even that variance very well.

the 50's in 39-41, the 25's all under 20.

Kurt Godel

pgibbo wrote: »

Are you considering/tracking gearing at all when you're thinking about your cadence? I'm assuming based on the numbers you're achieving the higher cadence by being in a lower gear and it's the spinning this gear that's leading to a higher RPE.

Yes I hit 372W for a minute at 100rpm in a lower gear, compared to 310W for a minute at 100rpm one rear cog up. Wasn't wearing a HRM, but perceived effort was much higher for the first.

Gearing would be an important variable, although I'd imagine it would have a direct correlation with HR (and thus could be substituted by HR).

tunney

Kurt Godel wrote: »

*Geek alert*

Probably only Tunney and Dory are still reading at this point, so...

I have one exam left for my Math degree, and am studying for it over the next month. Its Linear Statistical Modelling, which is involved in attempting to create models which help explain or predict relationships between variables. The response variable in our proposed Model is Power- its the goal, its the variable which is dependent on one or more explanatory variables. In effect, what causes more or less power? Effort (HR), cadence, sex, age, BMI, temp, crank rings, etc, etc. The more data you have for the this, the better you can test for response in Power. And then you can test what happens when you change one explanatory variable, while keeping the rest the same (ie what happens when all other explanatories are kept the same, but cadence is changed?). And then you can discard some of the explanatories and create an optimal scientific model.

I'll probably go nowhere further with this, but it may help my studies to think of a Linear Statistical Model in terms of something personally interesting to me, so perhaps we can tease something out over the next few weeks.

And therein is your problem Effort != HR, and HR to power is not linear.

So that Linear Statistical model is about as useful as an inflatable dart board.

tunney

Also look into power to heart rate decoupling and how that can be used to indicate suitability at present fitness levels for durations and intensities.

Kurt Godel

tunney wrote: »

And therein is your problem Effort != HR, and HR to power is not linear.

So that Linear Statistical model is about as useful as an inflatable dart board.

The terminology used in Stats can be very misleading- Linear Statistical Models are not necessarily Linear, in fact they are usually distinctly non-linear.

tunney

Kurt Godel wrote: »

The terminology used in Stats can be very misleading- Linear Statistical Models are not necessarily Linear, in fact they are usually distinctly non-linear.

Still effort != HR

Kurt Godel

tunney wrote: »

Still effort != HR

That would be the point of the Model- theoretically you could work out which combination of effort, cadence, plus other variables, equal optimal power output. You also get to quantify just how relevant the variables are (including factors like temp, gearing, weight, etc.), and their importance or otherwise to the model.

The proposal would be there is an optimal combination of cadence and effort which produces the best power output. Obviously time/fatigue/course would also be extenuating factors.

pgibbo

Kurt Godel wrote: »

The proposal would be there is an optimal combination of cadence and effort which produces the best power output. Obviously time/fatigue/course would also be extenuating factors.

Bike FIT would be key to this too

Kurt Godel

pm 2.4k run

Another short test, and again I was very nervous setting off. There was zero pain, zero stiffness, zero indication that I even had an injury... that's all good, but it can make you feel superhuman to be running again, and I had to work pretty hard to constantly keep the pace down under 5min/km. So far so good, but I'll need to keep the reigns on for a while.

career move

The reins you need have no 'g' in them

Bambaata

So did you negative spilt this run too ;-) ??

Hope it goes well Kurt but please take it slowly this time!! Give yourself a chance!! You've made so many gains that it would be a shame for you not to be able to realise them in a race setting!

ToTriOrNot

Kurt Godel wrote: »

*Geek alert*

Probably only Tunney and Dory are still reading at this point, so...

I have one exam left for my Math degree, and am studying for it over the next month. Its Linear Statistical Modelling, which is involved in attempting to create models which help explain or predict relationships between variables. The response variable in our proposed Model is Power- its the goal, its the variable which is dependent on one or more explanatory variables. In effect, what causes more or less power? Effort (HR), cadence, sex, age, BMI, temp, crank rings, etc, etc. The more data you have for the this, the better you can test for response in Power. And then you can test what happens when you change one explanatory variable, while keeping the rest the same (ie what happens when all other explanatories are kept the same, but cadence is changed?). And then you can discard some of the explanatories and create an optimal scientific model.

I'll probably go nowhere further with this, but it may help my studies to think of a Linear Statistical Model in terms of something personally interesting to me, so perhaps we can tease something out over the next few weeks.

Some of those variable might not have an effect at all on power, some you might need to transform (age is usually squared in lots of models if the relationship is not very linear!)
It sounds like an interesting project and if it helps you learn about it, go for it!!
Where are you hoping to get data from??

Kurt Godel

ToTriOrNot wrote: »

Some of those variable might not have an effect at all on power, some you might need to transform (age is usually squared in lots of models if the relationship is not very linear!)
It sounds like an interesting project and if it helps you learn about it, go for it!!
Where are you hoping to get data from??

The "Linear" in the title leads people to (reasonably) assume only linear relationships are used... quadratic functions are also used, the data is often transformed, or the mean in a sample set used.

I'm not sure where I'll get the data (could make it myself I suppose). Won't do anything for a month anyway as I'm too busy but right at the moment it helps me to formulate a model using Cadence/Power as Explanatory/Response, rather than the more academic variables in my textbooks.