SportTracks or WKO+

GavH

I've been using SportTracks to keep a log of my training and so far really like it. I'm particularly fond of the plugins such as Garmin Fitness that allows me to format bespoke sessions for upload to my 405 and 705, the Training Load plugin that allows me to measure TSS,TSB, CTL and ATL, even if I do not fully understand it as well as the plugin that estimates power for a given ride (which I definately don't understand but assume it's just a rough estimate) and the map of my route.

I've recently been trialing WKO+ as it seems to be generally considered the gold standard in analysis software. So far, there's a lot to like, especially the plethora of charts and data to analyse although it seems I won't get the most from it until I have a power meter. The only 'dissapointment' I have with it is that it won't give me a TSS value for a bike ride without power data but yet it can do it for a run. Can anyone explain this? The ST package manages it so why not WKO+? The fact that it cannot receive files direct from a Garmin 705 or a 405 also seems a bit weird, meaning I still need to open up Garmin Training Centre to then export the file across. No big deal and still quicker than ST, but nevertheless, for £60 worth of software I'd expect more.

Which brings me onto my question, is there (without having a power meter) any real value in 'upgrading' from ST to WKO+?

Are there any users of both packages that can share their experiences with the two?

Grateful, as always, for your thoughts.

Anonymous

No, the only reason I use WKO+ is the power function. If Sporttracks did power, I'd use that...

chrisw12

'tracks is great and when using hr it does a job. The trimp and other calculations seemed good to get an idea of where you're going if you can rely on hr.

For wko in the other hand you must have a power meter. You can manually enter power figures but if you haven't got a pm, how are you going to know the figures with any degree of accuracy.

My advice, use 'tracks, get an idea of what the metrics you mentioned do, then when/if you get into it get a power meter then get wko. Remember a powertap will work with your 705 so going down the power route will be slightly cheaper for you in the future.

APIII

Not sure what you mean about not being able to get 705 files directly into WKO+. It's a simple case of drag and drop.
Agree with the others, if you haven't got a powermeter you'd be better off with something else.

jibberjim

Even with a powermeter, I believe the PMC in training load plugin of SportTracks to be better. Despite being HR based, as long as you get good zones defined the number is always very close to a good WKO file, but WKO often produces odd results, because of the way NP and therefore TSS is calculated from the power file. Handling of stopped time being the big problem (although I'm also susceptible to producing NP busters too) The 11 hours of riding around I did watching IM LP had WKO giving my over 800 TSS (about what I'd get actually competing!) despite in reality it was nothing more than two 25minute periods at threshold and a bit of bimbling about.

For regular rides the numbers come out pretty similar (within 5% TSS always)

The big reason for WKO over ST is the in ride analysis, it's a lot better than the power tools in ST at analysing a single ride, so forensic analysis of why you screwed up in a Road Race is better. But for fitness tracking ST is better.

For in ride tracking GoldenCheetah is almost as good - but more importantly it's free and I don't believe WKO is worth the difference.

Certainly without a power meter WKO is not very useful for cycling.

Also running is pretty poor with WKO, since it relies on accurate altitude information - which you wont get from a garmin, and also it makes no allowance for the ground surface. So a 8.5km race at 4:13/km pace would seem like a pretty easy run for someone with a threshold pace under 4:00 - but in reality it's an extremely tough cross country race. HR based TSS overcomes that.

HR based TSS is of course not without its own problems, but it's a more reliable compromise than WKO+.

chrisw12

Money says that Alex responds to the post above.

800tss for 11 hours will perhaps get a comment?

Alex_Simmons/RST

chrisw12 wrote:

Money says that Alex responds to the post above.

800tss for 11 hours will perhaps get a comment?

GIGO

Anyone using Garmin units for recording power should ensure that:
- smart recording is disabled as it totally screws with the NP algorithm
- their FTP is set correctly in WKO+

Real NP busters (i.e. with valid data, not data corrupted by Garmin's smart recording system and other post-hoc manipulations) are exceptionally rare.

If you are seeing very high IF values for a given ride duration, then it far far more likely that either the data is invalid and/or your FTP is incorrectly set in WKO+.

Pokerface

So what is "Power Agent" that came with my PowerTap? And how does it rate compared to these other programs?

jibberjim

Alex_Simmons/RST wrote:

Real NP busters (i.e. with valid data, not data corrupted by Garmin's smart recording system and other post-hoc manipulations) are exceptionally rare.

So is commonly asserted, but I don't really know what this is based on as I know of no wide ranging study, and there's a tendency to simply reject any file that is produced as being an NP buster for indeterminate reasons by the only person I know who actually makes any attempt to collect them.

GavH

Jim, very detailed response above, thank you.

NapD, what about the ST plugin to analyse power that's available?

Alex, I've detected from this and a few other posts by yourself that the 705 is not all it's cracked upto be as far as Power is concerned. What's the deal with this?

jibberjim

GavH wrote:

Alex, I've detected from this and a few other posts by yourself that the 705 is not all it's cracked upto be as far as Power is concerned. What's the deal with this?

There's a few reasons, one is that in early incarnations of the firmware, it dropped values - so you'd miss maybe 5% of readings, which obviously makes a big difference. I've not seen any power drops on the latest firmware though.

The other problem is "smart recording" The problem with this is that for power you simply don't want this, you want to record as much as possible raw data. This is made worse by WKO+ not being able to handle sampling rates that are not consistent (a bug that's been promised to be fixed in the next version) if it was simply the loss of accuracy of the sampling it would be no different to the PowerTap's yellow headunit when it's not sampling at its maximum resolution, then it wouldn't matter but WKO's calculations make it particularly bad as then all analysis there fails.

It can be trivially dealt with simply by not having smart recording turned on.

The only other problem with using the garmin rather than the yellow computer is the lack of access to the torque values for checking calibration. However you can deal with this by using your computer and an ANT+ stick as it's only something you want to do at home anyway.

jibberjim

APIII wrote:

Not sure what you mean about not being able to get 705 files directly into WKO+. It's a simple case of drag and drop.
Agree with the others, if you haven't got a powermeter you'd be better off with something else.

For the 705 this is indeed the case, for the 405 though, you do have to transfer the files first to Garmin Training Center and then export the TCX before you can load it into WKO+.

Alex_Simmons/RST

jibberjim wrote:

Alex_Simmons/RST wrote:

Real NP busters (i.e. with valid data, not data corrupted by Garmin's smart recording system and other post-hoc manipulations) are exceptionally rare.

So is commonly asserted, but I don't really know what this is based on as I know of no wide ranging study, and there's a tendency to simply reject any file that is produced as being an NP buster for indeterminate reasons by the only person I know who actually makes any attempt to collect them.

Nothing that can be validated is ever rejected, all that is asked is for some evidence of an NP buster, which requires:
- A validated FTP, with methodology outlined at around the time of the suspected NP buster ride. That usually means several data points to remove potential for "outliers".
- Validation that power meter is apporpriately calibrated and was zeroed for both FTP check and suspected NP buster rides
- That data is not from any "downsampled" Garmin file or other such file with reduced data integrity (even if WKO+ did manage to deal with variable recording intervals, the data is still downsampled and will not be reflective of what actually happened).

The NP algorithm does not apply if you don't feed it appropriate data.

Note: I'm not saying NP busters don't ever happen, but they are rare. I have perhaps one such file amongst scores of thousands of files.

And it took a special kind of race effort to generate.

jibberjim

Alex_Simmons/RST wrote:

- A validated FTP, with methodology outlined at around the time of the suspected NP buster ride. That usually means several data points to remove potential for "outliers".
- Validation that power meter is apporpriately calibrated and was zeroed for both FTP check and suspected NP buster rides

Could you explain how a power meter can be zeroed incorrectly such that AP can be reported plausibly, but NP not? If the meter was incorrect, then AP would be inaccurate.

I have those for the two extremely suspect files I've produced (one with an NP over FTP for a 2:20 minute ride, FTP validated both pre and post the race in Time Trial efforts, and PowerTap zero'd by policy, and PT accuracy validated by static torque test, although again any PT innacuracy would impact all results. The other is only a 50minute ride, so doesn't meet the requirements to be called one, but it was still 15% higher than any isopower effort for the same time period)

Alex_Simmons/RST wrote:

Note: I'm not saying NP busters don't ever happen, but they are rare. I have perhaps one such file amongst scores of thousands of files.

Except you're not looking for them. And the majority of those files will not be from situations where they can be identified - to identify an NP buster it has to be a close to maximal effort (since otherwise it would not be peculiar even if the pattern of power was sufficient to generate an NP buster) So the number of files actually available to be identified as an NP buster is a small subset of those thousands.

It's not too difficult to construct ride files which lead to NP busters, and whilst they're obviously tough rides (being near maximal efforts in themselves) and they're only likely to happen in very technical or hilly crits by people with certain power files.

However, that's all off topic for this thread, where I only talked about the 800 TSS ride file, which wasn't from an NP buster, but from the TSS algorithm (and therefore WKO's implementation of it) which is grossly inflated by periods of zero power. That's the problem, you shouldn't need to understand the vagueries of the underlying algorithms to be able to successfully use WKO+. These problems are known, the software should be warning the user about it.

I recently assisted someone on another forum who was doing 300 TSS/d because his short easy 40minute runs were producing a TSS of 450. Now we know why that was, but he didn't, he trusted the software - if the software isn't usable, it cannot be recommended.

Alex_Simmons/RST

I'm not looking for them? I check every file I'm sent by a client. I am well aware of what it takes to create an NP buster. The file I refer to is in fact one of my own and you are correct in the type of ride required:



What I'm saying is that it is possible for the meter to be zeroed on most rides but not on others, which might cause outlier data. Also it is possible for PT torque zero to drift during an effort.

Nevertheless, given it's a PT, the scale of power calculation error (due to an incorrect torque zero) is magnified by wheel speed, such that the calculation of AP & NP could in fact be impacted in different ways.

I'm at a loss as to how a 40-min file can produce such high TSS. Perhaps you can share such a file.

chrisw12

Great discussion guys, very informative.

Jim I'd also like to know how someone could produce such a high tss for one ride?

Also your 800 for 11 hours, is doable but I would have thought very difficult and near to a race effort? What was the IF for this ride?

It got me thinking that if I do a 12hr tt next year, what sort of tss would this generate.

DaveyL

It was a run, not a ride - if I read it correctly.

Pokerface

Pokerface wrote:

So what is "Power Agent" that came with my PowerTap? And how does it rate compared to these other programs?

Asking again what the difference between Power Agent software and WKO+ or Sport Tracks is.

jibberjim

DaveyL wrote:

It was a run, not a ride - if I read it correctly.

It was both - it was "watching an Ironman" the garmin was being used both to record the total time of the IM, and to record me cycling and running during the day, the majority of the 11 hours was of course just standing around. 0 watts. The problem is that the TSS algorithm doesn't like standing around, it gives you values for it, even though you weren't doing anything. It's not a real problem, since as long as you know that, you can simply remove the entries that weren't actual cycling (ie when you're pacing up and down recording speed but no power obviously)

Actual cycling TSS for the day would've been more like 150 if taken as a whole on a single ride without the huge gaps walking around. It's easily dealth with if you know, but if you don't know - and I don't see it discussed anywhere on WKO+ etc. or in the Coggan and Allen book - then you run into the problems.

The difference between Power Agent and WKO, is that Power Agent has no PMC and no NP calculations. However it does have a pretty good ride viewer otherwise. It's worth considering, but I would say Golden Cheetah is currently superior for free individual ride viewer. Costs nothing to try it for looking at a ride.

DaveyL

I was referring to this:

"I recently assisted someone on another forum who was doing 300 TSS/d because his short easy 40minute runs were producing a TSS of 450. Now we know why that was, but he didn't, he trusted the software - if the software isn't usable, it cannot be recommended."

Which Alex and chris were in turn referring to.

jibberjim

DaveyL wrote:

I was referring to this:

"I recently assisted someone on another forum who was doing 300 TSS/d because his short easy 40minute runs were producing a TSS of 450. Now we know why that was, but he didn't, he trusted the software - if the software isn't usable, it cannot be recommended."

Which Alex and chris were in turn referring to.

Ah yes, they were runs where the individual had somehow managed to lose the WKO threshold pace setting he entered and it returned to the rather poor default of 9min/mile threshold pace, which was slower than he ever ran.

chrisw12

Jim I'm still basically confused here in what you're saying. Are you saying wko gives poor results in a rather extreme and may I say unusual example, you would not recommend it?

I also don't understand how if you did x hours of cycling, you ended up with 11 hours of data. I use the yellow computer and if I don't ride it doesn't record and eventually knocks off so wko never sees that data. So tbh that sounds more a problem with the Garmin than with Wko.

I kind of agree with you that if you pay for something then it should <perhaps> work in 100% of occasions or should provide warnings when it will not work. Then again we are dealing here with quite a technical product and perhaps people using it should only use it if they have a degree of intelligence and thus shouldn't need a warning.

jibberjim

chrisw12 wrote:

Jim I'm still basically confused here in what you're saying. Are you saying wko gives poor results in a rather extreme and may I say unusual example, you would not recommend it?

I'm saying that you have to fully know and understand the underlying algorithms for calculating TSS, NP etc. otherwise you'll have issues. It is not simply enough to understand the methodology behind it. That's for cycling. For running (which is a bit off topic, but also asked about here) the problems are larger unless you only run on a single surface, as it doesn't even attempt to address that issue. You can actually work around it by changing your threshold pace every day depending on the surface you're running on assuming you only do one run per surface per day.

chrisw12 wrote:

I also don't understand how if you did x hours of cycling, you ended up with 11 hours of data. I use the yellow computer and if I don't ride it doesn't record and eventually knocks off so wko never sees that data. So tbh that sounds more a problem with the Garmin than with Wko.

The yellow computer records when you're pushing the bike - anytime there's speed data, it records, so it's really not different. Certainly though if you have the pausing when not moving feature of the yellow computer then you run into the other way WKO+ gets things wrong - all your mean maximal graphs ignore those pauses, so it doesn't matter that your peak hour involved 5 minutes stopped waiting for the ambulance to get out of the way and actually took 65 minutes of elapsed time.

All of it is okay, it's consistent and you can make it work and then it's a powerful tool. However you need to know all these features of it, so you can be consistent in how you ride and record data. It's only if you made changes to your behaviour that it would start to record differently - for example if you switched your hill repeat intervals to being stopped at the bottom of the hill for 30seconds rather than coasting around for that time, TSS can be 14% different for the same workout. (see the wattage list for my explanation of how).

chrisw12 wrote:

I kind of agree with you that if you pay for something then it should <perhaps> work in 100% of occasions or should provide warnings when it will not work. Then again we are dealing here with quite a technical product and perhaps people using it should only use it if they have a degree of intelligence and thus shouldn't need a warning.

For sure! Which is why I'm here suggesting people stick with SportTracks, despite the fact I use both, and WKO has a number of good features. It's just unfortunate that the user interface, and the consistency isn't there. As a tool for experts it's pretty good - although you can get many of the features elsewhere. As a tool for non-experts it's not there at all yet.

chrisw12

Thanks for the detailed response.

I suppose then it's a case of each to their own. I do now understand some of the things you are saying and on reflection none of them affect me. I try my best to find rides with no pauses and a lot of my riding is done with np close to ap. When I do intervals I try to keep the protocol similar so at least I've got consistency.

Alex_Simmons/RST

OK, so you've been talking about rTSS, and not TSS (which is a bike power related metric).

I have no interest in commenting on running measures. This is a bike training forum and that's OT.

Secondly, if you stop for an extended period, then it's two rides (or 3 or whatever). So create a separate ride file for each ride in WKO+. That will deal with any residual RATSS* phenomenon. It will also deal with the Mean Maximal Power chart issue.


* Recovery Augmented TSS

Anonymous

Colin - Poweragent is just a basic program, it doesn't do things like TSS or NP...

I need to have a look at Sporttracks with the power plugin...

Anonymous

Tried it with Sporttracks, far from user friendly!

Alex_Simmons/RST

NapoleonD wrote:

Colin - Poweragent is just a basic program, it doesn't do things like TSS or NP...

Not yet anyway.

When the Joule bike CPU is released, Saris have advised they are also updating Poweragent to be able to provide the same metrics (NP, TSS) as will be available on the Joule's CPU display. They will be calculated metrics, meaning it will calculate these numbers from any power data fed into Poweragent.

I can't recall however if they are also implementing a Performance Manager (or equivalent) in the new version.

musto_skiff

How do I change the HR zone settings in ST?

I hae them set in my 705 but ST dosn't seem to be picking up the set zones in the Garmin?

liversedge

Alex_Simmons/RST wrote:

Secondly, if you stop for an extended period, then it's two rides (or 3 or whatever). So create a separate ride file for each ride in WKO+. That will deal with any residual RATSS* phenomenon. It will also deal with the Mean Maximal Power chart issue.
* Recovery Augmented TSS

Define "extended" ? 5s or 30 mins? I have seen odd artefacts in their algorithm for stops as short as a couple of seconds.

If you look really closely at the data, esp. with very short intervals, there are some real hidden oddities in there.

Alex_Simmons/RST

liversedge wrote:

Alex_Simmons/RST wrote:

Secondly, if you stop for an extended period, then it's two rides (or 3 or whatever). So create a separate ride file for each ride in WKO+. That will deal with any residual RATSS* phenomenon. It will also deal with the Mean Maximal Power chart issue.
* Recovery Augmented TSS

Define "extended" ? 5s or 30 mins? I have seen odd artefacts in their algorithm for stops as short as a couple of seconds.

If you look really closely at the data, esp. with very short intervals, there are some real hidden oddities in there.

Well it ain't my quote, but "long enough for a burrito and a beer".

yes, there are some oddities, and it is affected by the software used to suck out the data. even some of the native power meter software masks some things in the data (e.g. SRM Win post hoc modifies or hides some things that WKO shows verbatim, liked dropped data).

the NP algorithm has been modified between 2.2 and 3.0. i don't know in what way although they suggest it handles stops better.