TRIMP version used for PMC(TRIMP)?

[Art]

I've only recently got a power meter on my road bike, but also ride my mountain bike just as much without power meter. I have loads of rides going back to 2014 with HR data but no power, and also will still complete rides without power on my mountain bike unless I find a cheap way of adding it the mountain bike!

So I've been looking at how to use PMC both historically with just HR data and going forward with both HR and power data depending on the bike.

Strava premium has a fitness and freshness graph which can use both (in fact you can choose - power only, hr only, or hr and power) - although there are few details on how it works. (see attached)

I understand that I could specify a TSS for rides with no power data from now on, but is there any plans (or a feature I've missed) where I can get a PMC curve using TSS if available, or, say, TRIMP Zonal if no power data - I've looked back and TRIMP zonal is pretty close to TSS.

Also I have seen the PMC (TRIMP) preset chart - which TRIMP score does this use?

Thanks!

[Ale Martinez]

Preset PMC (TRIMP) is based on TRIMP Points, to have a scale compatible with TSS you could use TRIMP (100) Points which assigns 100 points to 1hr at LTHR to build your PMC.

To have a "mixed" PMC it would use formulas s.t. sts(TSS>0 ? TSS : TRIMP(100)_Points) for ATL, lts(TSS>0 ? TSS : TRIMP(100)_Points) for CTL and sb(TSS>0 ? TSS : TRIMP(100)_Points) for TSB in a Metrics Trends chart 

[Nigel Van de Velde]

Hi,

You can also manually enter a value for TSS or BikeScore for the non-power based rides. It takes a while to complete all your previous rides though...

Kind regards,

Nigel

[Art]

Thanks. TRIMP zonal points are way closer to TSS for my rides (where I have power). So would prefer to use them. I don't understand your last line - are you saying you can get a graph that uses TRIMP zonal (say) if there's no TSS value?

That would be good - and must be what other programs are doing? There's no preset I take it - how would I enter the type of formula you've suggested?

Thanks.

[Art]

Thanks - although I don't think I would be bothered to do that long term ...

[Ale Martinez]

El lunes, 20 de junio de 2016, 6:56:39 (UTC-3), Art escribió:

Thanks. TRIMP zonal points are way closer to TSS for my rides (where I have power). So would prefer to use them.

You can build a PMC using Stress curves based on TRIMP Zonal Points adding a new Metrics Trends chart, see: https://github.com/GoldenCheetah/GoldenCheetah/wiki/UG_ChartTypes_Trends#metric-trends

I don't understand your last line - are you saying you can get a graph that uses TRIMP zonal (say) if there's no TSS value?

That would be good - and must be what other programs are doing? There's no preset I take it - how would I enter the type of formula you've suggested?

No, there is no preset doing that yet, it could be added for download with v4.0. In the meantime see the above link, in "Curve Details" options there is a new one: "Formula", there you can type the formula(s) replacing TRIMP(100)_Points by TRIMP_Zonal_Points

[Art]

That is brilliant - many thanks! You may have just saved me getting a power meter for the mountain bike ;)! The TRIMP zonal values are pretty close to actual TSS score so for the MTB rides this will keep the PMC fairly accurate.

I'm to also display training loads on the graph too but the formula is not working:

(TSS>0 ? TSS : TRIMP_Zonal_Points)

Thanks.

[Ale Martinez]

El martes, 21 de junio de 2016, 12:58:08 (UTC-3), Art escribió:

That is brilliant - many thanks!

To give credit where it is due, it's Mark's work 

I'm to also display training loads on the graph too but the formula is not working:

(TSS>0 ? TSS : TRIMP_Zonal_Points)

It should work, may be there is some subtle error, if you post a screen capture of the Curve Details may be we can find it or, better if you are using the last development build, use "Export Chart" and post the resulting .gchart file

[Art]

Thanks Ale, and Mark!

I changed the "average" to "peak" under aggregate, and it gives the correct figure - I'm not quite sure what it's averaging but hey ho.

As you seem to know about these things - do you know the calculation for hrTSS used by training peaks? I would like to see how this correlates.

Thanks.

[Mark Liversedge]

On Wednesday, 22 June 2016 09:02:56 UTC+1, Art wrote:

I changed the "average" to "peak" under aggregate, and it gives the correct figure - I'm not quite sure what it's averaging but hey ho.

When aggregating (e.g. all rides for a week, month or year) the setting you choose controls how it is calculated.

Total will just add the values, Average will take the average by time, peak will just retain the maximum value.

So if you have HR aggregated for two rides where ride 1 has 200bpm for 1hr and ride 2 has 150bpm for 4hrs as a total that would be 350, as an average it would be (200*1 + 150*4) / (1+4) =  160 BPM and as a peak it would be 200bpm.

So the reason it was averaging the result is because that is the setting you chose.

Mark

[Art]

Thanks Mark for explanation. But this is plotting a point for each day - I may be selecting a period to graph, but these TSS points are plotted for each day - so unless there's more than one ride in the day (which there isn't in my case), average and peak should be the same shouldn't they?