Training vs racing using a power meter..."

SolarEnergy said:

Aside from the unorganized vs organized aspect. What is the difference between TSS accumulated during a race, such as a 60min crit, and training session TSS ?

There's no real difference necessarily between racing and training, the difference lies in what the effort was composed of. For example, a hard 60 minute crit will have an IF of ~1, and garner ~100 tss points. It could be made up of an hour of jamming (similar to an FT workout) some hard laps, some easy (similar to a v02max workout) or huge attacks followed by easier periods (like an anaerobic capacity workout). In each case, the training stress is similar, but the response is not.

SolarEnergy said:

Nowdays, one can precisely log power and hr data every 5 seconds (or so), and sort all that in a computerized training log. It's then possible to summarize this data, and see how much L4, L5, L6 and such have been done this week, month, year.

Shouldn't this new reality motivate riders to participate more and more in "training" races?

Yes, I think it should, as long as the athlete will truly ride it for training, and the nature of the event is specific to the athletes goals.

SolarEnergy said:

Shouldn't this new reality motivate riders to participate more and more in "training" races?

A couple thoughts:
1) there's more to training than just generating TSS.
2) it does matter how those L4, L5, L6 minutes were generated, and that's much more controllable during a training ride than a training race. IOW, 20 separate 1-min bursts to FT do not provide the same training stimulus as a 20-min interval at FT (unless the recovery periods are very short).

SolarEnergy said:

I mean, back in the days when we couldn't (or simply wouldn't) carfully log data gathered during a race, we coaches could only attempt to guess, or estimate the stress or strain of a race in the overall training week. What was happening during a "training" race, was totally outside our control.

Nowdays, one can precisely log power and hr data every 5 seconds (or so), and sort all that in a computerized training log. It's then possible to summarize this data, and see how much L4, L5, L6 and such have been done this week, month, year.

Shouldn't this new reality motivate riders to participate more and more in "training" races?

I think I'm with you here, with two exceptions. One, I think the training rides should be "designed" to produce specific adaptations, which dictates the ride structure. This may not sound like it is much different from having a computerized training log, but I think the difference is huge. Second, most training races and group rides are actually inferior workouts from the perspective of goal #1: a targeted set of physiological adaptations. I believe in starting with the training goals (defined in adaptation terms) and then backing into the ride designs. And, I think the entire thing should be automated, not just the post-ride log.

frenchyge said:

A couple thoughts:
1) there's more to training than just generating TSS.
2) it does matter how those L4, L5, L6 minutes were generated, and that's much more controllable during a training ride than a training race. IOW, 20 separate 1-min bursts to FT do not provide the same training stimulus as a 20-min interval at FT (unless the recovery periods are very short).

Those are both excellent points.

In looking at a LOT of training race files (and flat early season race files as well), tons of time is spent in the Anaerobic Capacity bucket, yet the effect on AC isn't really all that great, as the effort are submaximal and done with incomplete recovery, so they really aren't that 'anaerobic' in nature....

FWIW, after having a levelling off last year of FTP and pVo2 (and after growing wearing of doing any more intervals), I substituted a Tues. and Thursday training race for my interval days, and did 1-2 hr. of low tempo on Wednesday. Thursday is a 40 second climb, 1:30 moderate (or hard if you drill it at the front), rinse, repeat 30-50 times...Tuesday is varied, but mosty a 3 mile circuit which may or may not have a .5 mi climb.

6 weeks later, I recorded my highest 20 and 5MP, after not doing a 5 or 20 minute interval for that whole duration. Don't get me wrong, I don't question my preceeding training protocol at all, as the structured intervals gave me a super solid 'base' (for lack of a better term), and doing some structured intervals one of the two race days may have even been better--but my motivation was waning.

My main point is that the dynamic demands of weekday races may have different benefits that one would think.

whoawhoa said:

Yes, I think it should, as long as the athlete will truly ride it for training, and the nature of the event is specific to the athletes goals.

Yeah I guess. And god knows how difficult it is to keep this in mind.

Not wanting the victory, but instead, exploring, trying things and so on.

frenchyge said:

IOW, 20 separate 1-min bursts to FT do not provide the same training stimulus as a 20-min interval at FT (unless the recovery periods are very short).

Same TSS, same amount of time spent in L4, different response of course.

In fact, same stress put on the body. But when poorly organized, the response won't be as good.

I'll ask you what may be a dumb question Frenchyge. Can CyclingPeak account for this? In other words, is it possible to analyse the data, by applying filters? Like : Show me all my L4 segments that are at least 10min long?

RapDaddyo said:

I think I'm with you here, with two exceptions. One, I think the training rides should be "designed" to produce specific adaptations, which dictates the ride structure. This may not sound like it is much different from having a computerized training log, but I think the difference is huge.

I see. Your training methodology, is close to real things, but is still organized (designed) to produce specific adaptations.

RapDaddyo said:

Second, most training races and group rides are actually inferior workouts from the perspective of goal #1: a targeted set of physiological adaptations.

For many (take for example the OP of the thread I quoted in my question), can a crit done in the pack be considered as a L3 workout? with some higher intensity bursts or course.

RapDaddyo said:

I believe in starting with the training goals (defined in adaptation terms) and then backing into the ride designs. And, I think the entire thing should be automated, not just the post-ride log.

I like that idea.

RipVanCommittee said:

FWIW, after having a levelling off last year of FTP and pVo2 (and after growing wearing of doing any more intervals), I substituted a Tues. and Thursday training race for my interval days, and did 1-2 hr. of low tempo on Wednesday. Thursday is a 40 second climb, 1:30 moderate (or hard if you drill it at the front), rinse, repeat 30-50 times...Tuesday is varied, but mosty a 3 mile circuit which may or may not have a .5 mi climb.

Sorry I hit a big langage barrier here. You're saying that you started to do races twice a week, instead of interval days twice a week, and it paid off well?

How do you race logs look like?

If someone could show, and comment on a criterium training race log. That would be cool. Although I am determined to see that for myself this summer, when 1) I get my powertap and 2) When I attempt to our local weekly crit

Thanks guys.

SolarEnergy said:

Can CyclingPeak account for this? In other words, is it possible to analyse the data, by applying filters? Like : Show me all my L4 segments that are at least 10min long?

No. I have actually suggested something exactly like this. Andy has hinted that a new version of CP is in the works, but I don't know if it has this feature. You could use my match analysis app for this, I suppose, although I have not run it against a ride file for this purpose. But, even that app is not designed to analyze multiple ride files and keep a running log. Having done the match analysis app, I can tell you it is not as simple as it sounds to identify the qualifying ride segments. It's not so hard to apply a set of rules (assuming they can be defined precisely), the hard part is to define the rules. The problem is the segments that are not continuously in a range.

SolarEnergy said:

I see. Your training methodology, is close to real things, but is still organized (designed) to produce specific adaptations.

My training rides have always been designed to produce specific adaptations. They're just not traditional interval repeats. So, I may get a total of 40 mins at L4, but it could be a 15min L4 segment at the 35min point of the ride and a 25min L4 segment at the 2hr point of the ride. In between the L4 segments could be a half-dozen other segments (e.g., L5 & L6) and recoveries.

SolarEnergy said:

For many (take for example the OP of the thread I quoted in my question), can a crit done in the pack be considered as a L3 workout? with some higher intensity bursts or course.

Sure.

RapDaddyo said:

Having done the match analysis app, I can tell you it is not as simple as it sounds to identify the qualifying ride segments. It's not so hard to apply a set of rules (assuming they can be defined precisely), the hard part is to define the rules. The problem is the segments that are not continuously in a range.

Told you so. 😉

SolarEnergy said:

Can CyclingPeak account for this? In other words, is it possible to analyse the data, by applying filters? Like : Show me all my L4 segments that are at least 10min long?

The "Fast Find" feature would allow you to do exactly this, provided that you could come up with criteria by which the start and end of each segment would always be identified to your satisfaction. As RPO pointed out, the latter is the hard part, as any such "rules" would really be rather arbitrary.

RapDaddyo said:

You could use my match analysis app for this, I suppose, although I have not run it against a ride file for this purpose.

I don't know what that is. Some sort of a CyclingPeak add on app?

RapDaddyo said:

It's not so hard to apply a set of rules (assuming they can be defined precisely), the hard part is to define the rules. The problem is the segments that are not continuously in a range.

I see what Andy and you mean. What if a 14min L4 segment is interupted by a 2 second long L6 right in the middle.

RapDaddyo said:

My training rides have always been designed to produce specific adaptations. They're just not traditional interval repeats. So, I may get a total of 40 mins at L4, but it could be a 15min L4 segment at the 35min point of the ride and a 25min L4 segment at the 2hr point of the ride. In between the L4 segments could be a half-dozen other segments (e.g., L5 & L6) and recoveries.

Do you sometimes participate to training races? Do you try to control your segments in a training race pretty much the same as you do in a ride?

Was your training methodology inspired by some sort of "racing logs" reverse planning?

acoggan said:

The "Fast Find" feature would allow you to do exactly this, provided that you could come up with criteria by which the start and end of each segment would always be identified to your satisfaction. As RPO pointed out, the latter is the hard part, as any such "rules" would really be rather arbitrary.

In other words, if you don't operate this feature properly, it may be difficult to clearly asses and log the segments that are likely to be be beneficial training wize.

Would cross referencing this data with hr data help ? The power fluctuation is instantaneous, but HR doesn't fluctuate that quickly. IOW, help retreiving the segments by HR, log them by both power and HR.

My godness. Coaching cyclists has really became something else. 🙄

Thanks Andy. Really appreciated.

SolarEnergy said:

I don't know what that is. Some sort of a CyclingPeak add on app?

No, it's a separate application that uses a CP CSV ride file for input. The app is designed to identify and quantify anaerobic segments and recoveries. The identification and quantification parts are working fine and have now been conformed to the Critical Power Model with inputs for AWC, CP and weight in kilograms. The recovery piece is still a bit dodgey and I think I need to break it into two parts (alactacid and lactacid) and have different recovery half-lives for each because one part (alactacid) recovers fairly quickly and one part (lactacid) recovers slowly if at all (in the course of a ride). At the moment, the recovery algorithm doesn't do this (split the components).

SolarEnergy said:

I see what Andy and you mean. What if a 14min L4 segment is interupted by a 2 second long L6 right in the middle.

Not exactly. If you closely examine a ride file from a road ride you'll find individual observations all over the place, even when the cyslist is attempting to ride at a steady power. That's easy enough to deal with with rolling averages. I'm talking about a segment within a ride that is long enough (e.g., 30 secs) to affect the physiological response in the middle of a segment at a different level (e.g., a 30s L5 segment in the middle of a 20min L4 interval). It's easy to figure out what to do when chatting over coffee. But, to embed the logic in a program, you have to define the rules very precisely. Even the leading edge/trailing edge methodology that Hunter refers to in his match article makes both type I and type II errors. I have solved that problem, but there are others.

SolarEnergy said:

Do you sometimes participate to training races?

Yes, as well as fast group rides.

SolarEnergy said:

Do you try to control your segments in a training race pretty much the same as you do in a ride?

No, unless I am strong enough relative to the field to go to the front for awhile.

SolarEnergy said:

Was your training methodology inspired by some sort of "racing logs" reverse planning?

My training methodology was inspired by the topology of my rides. Trying to do traditional interval repeats on a typical course here (rolling w/a few major climbs) was trying to force a square peg into a round hole. So, I adapted my workouts to the topology of my rides and now the peg and hole are the same.

SolarEnergy said:

Shouldn't this new reality motivate riders to participate more and more in "training" races?

That would depend on their training goals.

For most people, training races involve an extra investment in time to get there and warmup, commuting to the race site, paying for registration, and the risk of the idiot crash. Most training races don't have enough women to provide an adequate and fair field for women except in a couple of cycling hotbeds so women would have to race with men which defeats part of the purpose except for elite women.