How is overall SPM calculated?

[Gammmmo]

Probably been covered elsewhere but here goes...within say a 5K session you can have the effort divided up into 1K bits for your persual.

1. How is spm calculated at each 1K....is is rounded up or down?

2. How is the overall spm calculated?
-Does it take the rounded 1K spm averages and average them and then round?
-Is it more accurate and ignore the 1K averages, look at the underlying spms for the whole session, average them and then round that?

3. How does the rounding work? e.g. if spm=25.5 does it get rounded to 26 or truncated to 25? What about say 25.4? Or 25.6?

[jamesg]

Due to lack of places, spm numbers are truncated both ends. So if we row at 123.x strokes per minute, or 23.x, we see 23.

Averages seem to sum successive truncated data, divide and then truncate that too. So if we row one split at 123.x spm, and the next at 124.y, (or 23.x and 24.y) the average seen is 23.

Ergdata totals strokes in the entire workout, which might help.

[Citroen]

All split values in PM2/PM3/PM4/PM5 are instantaneous.

As the split time/distance is reached the PM records the current state of the counters.

No averaging, no rolling averages. All rounding is simplisticly rounded down. Not mathematic or algebraic.

This is why a restricted to 20 SPM 30 minute row is still valid at 629 strokes as far as the PM works.

[Gammmmo]

Due to lack of places, spm numbers are truncated both ends. So if we row at 123.x strokes per minute, or 23.x, we see 23.

Averages seem to sum successive truncated data, divide and then truncate that too. So if we row one split at 123.x spm, and the next at 124.y, (or 23.x and 24.y) the average seen is 23.

Ergdata totals strokes in the entire workout, which might help.

So, you could arrive at a more accurate overall average by averaging the split averages and leaving as say a decimal to one decimal place? This is what I sometimes do with my training notes.

[jamesg]

Yes, if not using ergdata, I just consider split average 20 to be 20.5 and so on. Not that it matters much.

If we set up one minute splits, the problem should disappear, as PM visibly (in ergdata) counts a stroke at the finish. Any incomplete stroke in one split should appear in the next. If so there can be no truncation error in a set of one-minute splits, since all stroke numbers must be whole already; nothing to truncate. Which seems logical, if we think there's no such thing as half a stroke.

[Carl Watts]

This has been covered in this forum for years.

Basically I made a stroke counter before ErgData even existed and came to the conclusion that the average SPM recorded in your top line of results in the PM3 and PM4 was WAY off. The monitor always rounded down and was not consistent.There was the occasional row that rounded down over 1spm

Now that ErgData is available you simply take your total stroke count and do your own math and you can record your average SPM to 1/10ths and the result is very accurate.

I guess many people will argue its not important, however in the case of the 30R20 its pretty important as the rating is an important part of the test and give or take a couple of strokes is all that is allowed really if you want to start comparing the results to others..

[Gammmmo]

This has been covered in this forum for years.

Yup, thought so as per opening post.

I guess many people will argue its not important, however in the case of the 30R20 its pretty important as the rating is an important part of the test and give or take a couple of strokes is all that is allowed really if you want to start comparing the results to others..

Well...I dunno. We are creatures of habit and it's very hard to change the way in which one rows a particular piece done frequently. As an example, when pushing my 5K PB, Henry told me to try to increase my spm. Now, of course that is only possibly if one gets more aerobicaly fit but it also takes a mental rewiring as to what feels right. I would argue that it's very hard to see significant changes in spm and so any accurate (possibly to one decimal place) change would be useful to see and not clouded by rounding/truncating.

[jackarabit]

This has been covered in this forum for years.

Yup, thought so as per opening post.

I guess many people will argue its not important, however in the case of the 30R20 its pretty important as the rating is an important part of the test and give or take a couple of strokes is all that is allowed really if you want to start comparing the results to others..

Well...I dunno. We are creatures of habit and it's very hard to change the way in which one rows a particular piece done frequently. As an example, when pushing my 5K PB, Henry told me to try to increase my spm. Now, of course that is only possibly if one gets more aerobicaly fit but it also takes a mental rewiring as to what feels right. I would argue that it's very hard to see significant changes in spm and so any accurate (possibly to one decimal place) change would be useful to see and not clouded by rounding/truncating.

Turning “would be useful” to is useful is as simple as getting ErgData and dividing stroke count by time of piece. This has to be done at the afterparty of course as converting mod 60 time to decimal is not something easily accomplished by my old head in medias res. Bit of a chore but gets easier once you accept that the head on shoulders can do things the PM can’t. Look at it as easier than cobbling together Carl’s optical beam counter (pun intended) alongside your rower seat.

HINT: Think of 6” as .1’ and 3” as .05’. EXAMPLE: 10:33.X becomes 10.55’. (I toss the itty bit to right of decimal before conversion.). From there, stroke total/10.55=spm. Rounding from two decimal places to one in the result is adequate precision imo.

[Carl Watts]

Ergdata is great, it just needs a couple of bux fixes which have been reported to Concept 2 for the next version release.

This coupled with RowPro covers pretty much all the data you would ever need. Temperature and humidity logging is still done manually and entered into the results but it doesn't look like mobile phones or the Concept 2 monitors will have the required sensors for this anytime soon.

Ergdata just needs the stroke count and the projected finish up on the main screen at the same time for me.

[jackarabit]

Calculated spm, dps, and W’/s to one decimal pace on Wolverine L4 work last yr:



Can’t say I learned anything earthshaking but maybe others will.