Firstbeat Athlete

Stan · Post by **Stan** » Sat Jan 07, 2012 2:33 pm

Thanks for the reply Lee.
I am now beginning to think it could be the battery though. The display is fading and today although movescount recorded all the average data, it did not display the graph. FBA again failed to analyse. I have ordered replacement batteries and see if that helps - it wont hurt anyway!

Doc RowSlo · Post by **Doc RowSlo** » Sat Jan 07, 2012 2:56 pm

Stan wrote: I have ordered replacement batteries and see if that helps - it wont hurt anyway!

They're just CR2032's. Why not pop down to Tesco?!

If the display is fading (turning on the backlight is a good way to tell) then this is probably the culprit.

Stan · Post by **Stan** » Sat Jan 07, 2012 3:06 pm

I disabled the backlight to save battery power

It is definitley fading so hopefully the new battery will solve the issue - I ordered a dozen from amazon for 2.60 so that should keep it going for a few years

mgiven · Post by **mgiven** » Thu Jan 12, 2012 4:59 pm

Please excuse this posting that is not rowing related (gasp!) and from a Garmin and not Suunto user (double-gasp!!) The reason I'm posting HERE is that the Garmin Forerunner 610 recently became the first device in their lineup that is compatible with Firstbeat Athlete and lots of us Garmin users (mostly runners and some cyclists) are learning the ropes of FBA together. There doesn't seem to be an internet Firstbeat Athlete message board, forum or user community of any kind and any search for people talking a bout FBA on the net leads right to this forum, so you're the premiere net FBA experts (at least easy ones to find) if that was your intent or not

My question is to whether there are other groups (Suunto?) where others discuss FBA and all the finer details. Thanks again and pardon the non-rowing intrusion!

Michael

Stan · Post by **Stan** » Thu Jan 12, 2012 8:23 pm

Hi Michael, welcome to the forum.
Just getting over the shock of the idea that there are sports other than rowing

I am not aware of any suunto discussion forums regarding firstbeat athlete. There are a few of us here who use it though and I am sure if you have any questions we can at least make some comments. I believe most of us are using suuntos but I understand that there are a few polar watches that are also compatible. I saw on the recent fba update that garmin had jumped onboard too - I dont know if any of us are using Garmins. I think Paul Buchannan (Butch91) on the C2 UK forum uses both garmin and suunto and sometimes posts the outputs from both in his blog (you would need to join the C2 UK website to read his blog).

Huw Thomas · Post by **Huw Thomas** » Thu Jan 12, 2012 9:14 pm

mgiven wrote:Please excuse this posting that is not rowing related (gasp!) and from a Garmin and not Suunto user (double-gasp!!) The reason I'm posting HERE is that the Garmin Forerunner 610 recently became the first device in their lineup that is compatible with Firstbeat Athlete and lots of us Garmin users (mostly runners and some cyclists) are learning the ropes of FBA together. There doesn't seem to be an internet Firstbeat Athlete message board, forum or user community of any kind and any search for people talking a bout FBA on the net leads right to this forum, so you're the premiere net FBA experts (at least easy ones to find) if that was your intent or not My question is to whether there are other groups (Suunto?) where others discuss FBA and all the finer details. Thanks again and pardon the non-rowing intrusion!

Michael

Hi - I am now a proud owner of a 610 as well - its lovely - and the garmin agent uploads the r-r file directly into FBA
What questions You got??

It arrived yesterday! and first session today. Its the same price as the Suunto T6d, has built in GPS for the bike and stores r-r data
The "in watch " TE analysis doesn't seem to ask about activity level - but they get imported into the FBA automatically and you then analyse them according to your current profile parameters. The lap markers 1-6 are shown clearly (@ each 2k of the 10k today)

: fba.jpg (59.1 KiB) Viewed 5330 times

mgiven · Post by **mgiven** » Fri Jan 13, 2012 2:22 am

Thanks all and great that there is at least one other proud 610 owner here Huw

I still feel bad for hitting up a rowing group for all my Firstbeat questions, but I'm shocked there doesn't seem to be a Firstbeat user group that spans all devices. I've had most of the Garmin Forerunners and a few others and the 610 is my favorite. Still looks like a normal watch, tons of functions, but what makes it for me is the option to keep the night light on all the time (I run a lot in the dark), four data point screens, complete screen customization, vibrating alerts and now Firstbeat Athlete compatibility.

You already raised on point in that the watch itself doesn't have activity level on it anywhere, but your FBA profile grabbed it from Garmin Connect profile. In digging around it seems the algorithm that Garmin uses for "in watch" TE computation doesn't use activity level at all, but workout history with more of a focus on the workout at hand to spit out a TE. Garmin and Firstbeat both mentioned it would take the watch a few workout to "learn you" which I'm guessing means have some history data instead of sounding like the watch is alive

The FBA software instead seems to put a high value on activity level and I'm struggling through getting that right. It took the initial level set to it from Garmin Connect (7 in my case) and has already wanted to bump me up 3 times in the two week trial (they now want me at 8.5).

I think my question and the discusion currently going on in the Garmin forums is why the two (watch and FBA) don't match EXACTLY for TE. You know most detail oriented folks that like posting to internet message forums need everything to match and be in order, or at least know why. My guess for why the watch and FBA don't match exactly is that 1) The watch has far less computing power than the software and 2) They use different algorithms where the watch focuses more on current workout and other workouts stored, where FBA uses perceived activity level + other workouts stored in FBA. They will be close, but it seems logical that they may not match exactly.

With that said, what do you trust more? Some of the newer Garmin users are tinkering around with the FBA settings (activity level, max HR, V02max, MET) to try to get FBA to match the watch exactly since the watch is what is being worn DURING a workout and gives you a live status of TE. However I can see where this thought process might be backwards. Shouldn't we instead trust FBA more since it has more advanced computing power and the exact product of Firstbeat and not a Garmin tweak of the purchased algorithm? The value on the watch is close enough, but if I want them to match I'd instead ask why the watch isn't closer to FBA than vice-versa.

Does all this make sense or am I just rambling? I told you we had lots of questions. THANKS AGAIN!

badocter · Post by **badocter** » Fri Jan 13, 2012 12:35 pm

You are quite welcome here, but if you feel that guilty then i suggest you consider adding indoor rowing to your repertoire

You just need to get your hands on a Concept 2 indoor rower and then you can join us in the real fun

http://www.youtube.com/watch?v=E-6Risb44Xo

It is very important that it be a Concept 2 rower because C2 is the only brand for which there is an active community of competitors and fan support (like this forum) !!! It is also the only brand I am aware of where you can row with and race other people in real time online with the help of some third party software from a compancy called digital rowing. The machines are self calibrating, so everyone can compare with everyone else and know they are comparing apples to apples. The US vendor links are here, but C2 has websites in other countries as well if you are not in the US:

https://www2.concept2.com/VIA/index.jsp?RefAgent=IND
http://www.digitalrowing.com/Products/index.htm

FBA does tend to be manic depressive about wanting to change one's activity level and I think the conclusion that those of us in the thread have ome to is to ignore its suggestions once you get to around levels 8-8.5. Otherwise it often does not make any sense what it wants to do because the Coach function will be telling you to take gobs of rest days, but if you do take the suggessted rests it will then start wanting to downgrade your AC. Somewhere in FBA is a table of descriptions for how many hours a week and type of training you do and what AC level it corresponds to - the table is a better guideline than the automatic upgrade/downgrade algorithm methinks.

Doc RowSlo · Post by **Doc RowSlo** » Fri Jan 13, 2012 12:49 pm

Michael, you are quite right in your assumptions about the different TE's between FBA and the wristop computer (whichever one is used). The difference is simply because FBA can employ far greater computing power than the tiny watch and so is able to analyse the data with a more complex algorithmn (or to use the same algorithmn far more extensively).
As others have said, it's probably best to leave the AC at whatever level you think is correct and not to let FBA change it automatically.

There are forums out there that discuss all this stuff and I'm sure that Steve Wiser will be along to point you towards them shortly!
Try searching for "Suunto Training Manager" which was the forerunner (pardon the pun) to the vastly inferior Movescount. Both FBA and STraM are written by FirstBeat Technologies and use essentially the same algorithmn.

Huw Thomas · Post by **Huw Thomas** » Fri Jan 13, 2012 1:17 pm

I concentrate on epoc rather than TE - and use FBA to compare identical sessions. Changing activity level alters TE and so most of us stick to one activity level to make comparisons easier

I love the ability to compare identical sessions

From my blog wrote:Benchmark row WP L4 180x2 184x2 40 mins 10054m 1:59.7 (WP ref pace 1:39)
Because this is as reproducible row with standard strokes and paces it makes a great benchmark for fitness. I have done it several times after several different training weeks.

Second lowest epoc today (black line) My lowest EPOc for the session was a day after a hardish cycle (but not Dartmoor epic!). Similar training week came around again and lo and behold another extremely low epoc and HR for the session. Significantly better than red line 3 days ago. Green line is 2 days after an "empty the tanks ride" and was definitely overtraining.

Guess this means I'm not overtraining at present and the bike/row mix is good.
Quite surprised me but pleased that I am training at a maintainable level.

Some nice articles here

http://www.firstbeat.net/files/white_paper_epoc.pdf
http://www.firstbeat.net/files/white_pa ... effect.pdf

I use FBA and activity class 8.5-9 having trained x6/week for 3 years or so

Watch is a convenient way of recording r-r for analysis in FBA

Regards

Huw

slwiser · Post by **slwiser** » Fri Jan 13, 2012 1:44 pm

Huw

Looks like you are working with FBA very well and understanding how to use it to better train. It is great to see how your EPOC curves relate to your training cycle.

Your explanation of how you are using the watch, FBA and the EPOC results are very good. I like to review TE over a longer period of time just to see if my workload average has been. I am in a period where I need to back off load and duration for a month or so. Last month was a normal workout month but the day-job stresses added to my fatigue following a sever cold and my fitness level dropped quite a bit. Following a cold you would see your EPOC curve do the same as when you were over-training. Combinations of body stresses (sickness and day job stresses) and workouts can result in high EPOCs for the same workout as well.

badocter · Post by **badocter** » Sat Jan 14, 2012 1:15 am

mgiven wrote:it seems the algorithm that Garmin uses for "in watch" TE computation doesn't use activity level at all, but workout history with more of a focus on the workout at hand to spit out a TE. Garmin and Firstbeat both mentioned it would take the watch a few workout to "learn you" which I'm guessing means have some history data

The watch is probably using the "automatic" algorithm that FBA uses to suggest changing activity level, the watch simply isn't reporting the value and doesn't give the option to override. If the watch can display EPOC, you can work out ahead of time in FBA the EPOC needed to hit your target TE, and then monitor the EPOC on the watch.

mgiven wrote: I think my question and the discusion currently going on in the Garmin forums is why the two (watch and FBA) don't match EXACTLY for TE. You know most detail oriented folks that like posting to internet message forums need everything to match and be in order, or at least know why. My guess for why the watch and FBA don't match exactly is that 1) The watch has far less computing power than the software and 2) They use different algorithms where the watch focuses more on current workout and other workouts stored, where FBA uses perceived activity level + other workouts stored in FBA. They will be close, but it seems logical that they may not match exactly.

My guess would be that the activity level does not match is the primary cause of the discrepancy. If the watch can report the EPOC, compare that with the EPOC in FBA. If both EPOCs are the same, then the TE are different because of the activity level. If the EPOC are slightly different, I expect that FBA has greater capability to denoise the data (artifact correction). If it is due to artifact correction, if the error percentage in the data is high, then the algo will have to deal with more uncertainty and the difference in estimated TE can be greater. FBA reports the error percentage in the session summary. Chestbelt fit and contact as well as reciever placement (in your case the watch, in my case the pc-pod) can both impact the error percentage. A few pages back in this thread there is some discussion about my attempts to reduce the error percentage I was seeing. I now get 0%-1% for almost every workout where in the past I was anywhere from 4%-50%. Placement and distance to the reciever seemed to make the most difference. I am not sure if the electorde gel helped or not, but I still use it.

mgiven wrote: With that said, what do you trust more? Some of the newer Garmin users are tinkering around with the FBA settings (activity level, max HR, V02max, MET) to try to get FBA to match the watch exactly since the watch is what is being worn DURING a workout and gives you a live status of TE. However I can see where this thought process might be backwards. Shouldn't we instead trust FBA more since it has more advanced computing power and the exact product of Firstbeat and not a Garmin tweak of the purchased algorithm? The value on the watch is close enough, but if I want them to match I'd instead ask why the watch isn't closer to FBA than vice-versa.

Even FBA monitoring in real time calculates a slightly different TE than the one FBA calculates after it analyzes the session at the end. I typically see a drop of 0.2 between what the FBA real time monitor tells me I hit during the workout compared to what it comes up with in the post analysis (ie, the FBA real time monitor shows I hit 4.3, but the FBA post report may give a 4.1). For each new data point the real time monitor looks at, it can compare it only to the past data points, whereas in the post workout analysis for each data point, the analysis can also compare each data point to the future data points. The post report thus has a huge advantage on denoising the data because it has twice as much information available to assess each data point, past+future, whereas the real time monitor can only look at the past. The watch likely has the equivalent of the FBA real time monitor because the FBA post-analysis method probably requires more computing power than the watch is capable of. If the TE estimate is only 0.2 or so off it is not a big deal for assessing if you are overtraining or not. if you are overtraining you will be getting TE values of 4.5 on a session that is normally a 3.0.

At the end of the day either method is probably close enough to guide you, assuming you have input a reasonable activity level in FBA, and that the watch has learned you enough to have also arrived at a reasonable activity level. What will throw off both the watch and FBA are short interval workouts for which BOTH will report a TE that is too low ....the algo is only really valid for medium to long steady pace work. An 8x500m rest 3.5 minutes interval session will inevitably get a TE of 3 or so for me even though it feels like a TE 4+ and requires the recovery of a TE 4+.

badocter · Post by **badocter** » Sat Jan 14, 2012 3:10 am

To demonstrate the difference that even exists between the FBA real time monitor and the FBA post analysis report, below are the two reports as collected on a 45 minute steady recovery row I was doing tonight, 10323m@2:10.8/500m, 25 strokes per minute, which you will usually see reported elsewhere in "rowing shorthand" as 10323m@2:10.8R25. I was aiming for middle of UT2 HR range, and told the RTM to use smart training to pace me to hit a 2.5 TE. As you can see in the first screen shot, FBA-RTM told me I hit my target TE of 2.5 which should be EPOC 46 (with about 3 minutes left to go in the session), however, the post analysis report shows only a 2.1 with an EPOC of 31 (you will have to scroll down in the picture window to see it)

Both values are generated within the same piece of software with the same input data (min HR, maxHR, Activity level, etc....), and yet the TE differ by -0.4. And this a report whose data had an error percentage of 0%.

My experience is that the delta is smaller between the two estimates for higher TE workouts but it is still there, and it seems to always be in the same direction (RTM is always higher than post analysis). Since it is repeatable, you will get used to the pattern and eventually come to the same conclusion I did....don't get hung up picking apart differences of 0.1-0.2 on your measurements. Whether the workout was a TE 2.1 or a 2.5, at the end of the day , it was still a recovery row, and I have confirmation from the fact that my HR was in the UT2 band the entire session that it was a recovery row.

My day job involves mathematical modeling of dynamic processes, and I can tell you that the following statement seems to apply universally: "All models are wrong, some are useful"

: 2012jan13 badocter 45 min steady UT2 RTM; 2012jan13 45min steady RTM.gif (28.19 KiB) Viewed 5296 times

: 2012jan13 badocter 45 min steady UT2 post report; 2012jan13 45min steady post Report.gif (63.56 KiB) Viewed 5296 times

: 2012jan13 badocter 45 min steady UT2 post report 2; 2012jan13 45min steady post Report page 2.gif (41.48 KiB) Viewed 5296 times

badocter · Post by **badocter** » Sun Jan 15, 2012 4:40 am

I just posted a similar TE comparison for a speed interval session in this thread:

viewtopic.php?f=7&t=661&p=97513#p97513

In this case FBA real time monitor reported a 4.0 and the post report showed a 4.1. I think that is a first for me to see an RTM value that is smaller than the post report. It is also probably the first speed interval session for me to get a TE of 4 or more, but the rests were very short (only 80 seconds). And it felt like a TE 4+ session

badocter · Post by **badocter** » Sun Jan 15, 2012 12:51 pm

I stumbled on this paper by Firstbeat while googling. The list of references may be handy for someone trying to dig deeper.

http://www.pulseperf.com/pulseperf/publ ... _paper.pdf

Some of the text also leads me to believe that the developers think the method is valid for interval sessions. If that is true, then it might indicate that we are making the rests too long for some of our intervals sessions and that is why the TE is so low. Let's take the example of the 5x1500m rest 5 minutes. All of the 5x1500m rest 5' sessions I have data for have TE between 3.3-3.6. I have heard others say that the impact of 4 minute rest versus 5 minute rests is not very much since recovery hits a point of diminishing returns as the rest gets long enough. Longer rests might increase ability to pull a slightly faster split, however, if the Firstbeat Athlete training effect is valid for interval sessions, the extra minute of rest allows the EPOC to plummet longer which then leads to the TE of the session being lower. If the TE method is valid for intervals, then with the shorter rests like Plummy is doing for his 5x1500m rest 4', the TE for the session will be higher and have better cardiovascular training value.

I am beginning to think that the low TE for some of our speed sessions is that the rest times we are using are too long. We don't expect the impact of a 5' rest versus 4' rest on pace to be very much, but it sure allows the EPOC to drop quite a bit which we then have to raise back up on the next rep.

It may very well be that the optimal is variable rests where the next rep starts when the EPOC just begins to go back down so that we have no ground to make up (during the inital part of the rests, EPOC normally keeps going up because the HR is stilll elevated...at about the halfway point of the rests in some of the interval sessions it starts going back down.) It is worth noting that on the 20x100m rest 80" session I did yesterday, the EPOC never went down during the rests because the HR was still suffciently elevated even at the end of the rest. I reached a of TE 4.1 on the session because of this even though it was a speed session (and a very fast paced speed session at that relative to the standard Pete Plan speed sessions which usually get me a TE of 2.9-3.3 despite feeling horrendously hard). Pete himself set his 2k PB during the period where he was heavily leveraging sprint training to prepare for the MAD team 100k record. Is there a connection to his performance improvement and the fact that he was doing speed sessions that had a higher TE

Mat · Post by **Mat** » Sun Jan 15, 2012 2:42 pm

Interesting comments on rest Bill, and I can honestly say that I have long questioned in my own mind the long rest periods prescribed in plans such as the PP. I generally follow the IP where intervals are time rather than distance based, but have found from my experience that the following rest periods have produced optimal results;

Sprint training up to 3' per Interval = like for like rest (e.g. 2' interval = 2' rest)
Intervals from 3' - 10' = 3' rest per interval
Intervals over 10' = 2' rest per interval

The last one may raise queries, but it is my strong contention that these longer intervals are designed to improve aerobic fitness, and are important in training for longer pieces such as 1k - 60' TTs. As such, it is important to keep the HR up, and the rest period serves simply to have a quick drink and towel off before starting the next interval. Ideally, these could be cut down to 1', but I tend to find that 2' allows me to keep a faster pace up throughout

Doc RowSlo · Post by **Doc RowSlo** » Sun Jan 15, 2012 10:59 pm

Interesting thoughts and comments from Bill and Mat

As far as the reliability of TE during interval sessions is concerned, then I think that it's only really valid where the rest period is very short compared with the active rep eg 3x6k/2'r. Eddie Fletcher says as much in his Suunto rowing manual, I seem to remember. For something like 5x1,500m/5'r (or 4'r) where the rest period is pretty much the same duration as the rep, then the whole thing falls down. The TE might be of some use when comparing it to the next 5x1,500m session but little else, I suspect. As Bill suggests, the whole EPOC / TE from HRVA concept is based upon continuous exercise at a fairly steady workload.

I'd also agree with Mat in that short rest periods, that just allow a drink and towelling off, are generally more beneficial for the reasons that he gives. Of course, there are certainly other points of view (Mike Caveston for one!)

Doc RowSlo · Post by **Doc RowSlo** » Sun Jan 15, 2012 11:06 pm

badocter wrote:
In this case FBA real time monitor reported a 4.0 and the post report showed a 4.1. I think that is a first for me to see an RTM value that is smaller than the post report.

I've seen that quite often Bill.
My anecdotal impression is that for long aerobic sessions the EPOC / TE is overestimated in real time and that for long AT and harder ones, it's underestimated.
I can't really comment about interval sessions as I hardly ever do them! (Well, apart from 3x6k/2'r, but I don't think that that really counts as an "interval" session)

stumpy · Post by **stumpy** » Mon Jan 16, 2012 3:06 pm

Hi guys I seem to remember reading somwhere that the epoc curve should remain level in the rests to get the best training effect. So if the active rest is to long or to slow the curve drops. I used to adjust my pace to hit the predicted distance required in the active rest to get this to balance out. It didn't always work out though as more often than not I just had to stop for a bit of passive rest.

L2 4x2 r1500m 1:55.7

: data.JPG (63.76 KiB) Viewed 5244 times

Mat · Post by **Mat** » Mon Jan 16, 2012 9:52 pm

Thought it may be of interest to post my session from this evening. The session was 4 x 5' r2', with slow active rest. The warm up and cool down are also shown below, but the 4 "working intervals" are fairly clear to see;

: 4x5'r2'.jpg (83.61 KiB) Viewed 5234 times

Clearly the EPOC does drop a little between intervals, though not too much. I'll post tomorrow's once completed too, as these will be longer intervals (either 2x5k or 3x12'), but still with 2' rest

Huw Thomas · Post by **Huw Thomas** » Tue Jan 17, 2012 6:51 pm

Here's one of mine 9x750 on 1 min rest - the epoc levels nicely but doesn't fall
For me anything more than 1r on shorter intervals and it slips down

Stan · Post by **Stan** » Tue Jan 17, 2012 7:57 pm

Similar thing here. This is last weeks ctc (7x777). Notice the slight fall in EPOC between interval 6 and 7 when I had a 2 min rest compared with the previous ones with only 1 minute

: Image 1.png (6.73 KiB) Viewed 5229 times

Mat · Post by **Mat** » Tue Jan 17, 2012 9:44 pm

Mine below for the 3x12'r2' as promised. Huw and Stan's demonstrate well how steady the EPOC climbs through intervals with 1' rest where there's just enough rest to regroup before the next onslaught

: 3x12'r2'.jpg (58.16 KiB) Viewed 5221 times

mgiven · Post by **mgiven** » Sat Jan 21, 2012 6:55 pm

Hi all... back for another "newbie" Firstbeat question that I'm sure you can all help me with.

Just finished a run (gasp!) where I ran for 41 minutes (in the snow!) and then took a 15 break and then went out again for 41 minutes with my running group. Should I leave these logged as two activities or will that screw up the EPOC SUM calculations? My guess is that I should merge them. Keeping them as two would log one activity at 41 minutes for an EPOC peak of 212 and a second activity for 41 minutes for an EPOC peak of 138. Together it would be one 82 minute activity for an EPOC peak of 212. Woudln't keeping them as two activities make the "coach" calculate one extra activity, when it was really just one long one with a break AND sneak an extra 138 into my weekly EPOC SUM total?

Thanks!

Stan · Post by **Stan** » Sun Jan 22, 2012 10:18 am

I am now sure how much difference it would make to record them as 2x41 or 1x82. I usually make each training session one entry on fba. The exception would be mixed sessions where I row and then cycle where each session would be separated. It makes no difference to the heart of course, but I like to have an accurate record of what I was doing.