VO2max and Race Prediction over-estimation

VO2Max is a metric for your peak effort over a short time. There is not a fixed correlation between your peak effort and your marathon effort.

If you are a sprinter type, your marathon effort will probably be lower (time is longer) than for the average runner with a VO2Max equal to yours.

And if you are an endurance type, your marathon effort will probably be higher.

There is a race prediction formula by Pete Riegel which describes the relationship:
T[SUB]2[/SUB] = T[SUB]1[/SUB] x (D[SUB]2[/SUB] / D[SUB]1[/SUB])[SUP]C[/SUP] Where T1 and T2 are your race times, D1 and D2 are your race distances, and C is a coefficient which can vary between 1.06 and 1.10.

The C coefficient is the key to what I wrote above. Different people will have a different value of C. A person with a high C coefficient will see a larger pace penalty when going at longer distances.

So it is possible that you simply have a high personal value of C in the Riegel equation.

It is of course also possible that Garmin's marathon prediction is off for the average runner. I wouldn't know. I am engineer, not a running coach. I think I have read somewhere that Garmin's race prediction comes from the Cooper Institute (probably the same Cooper who came up with the Cooper test?). So you might be able to find more about the statistical background for Garmin's predictions if you look into what they have published.

The main thing to keep in mind with the race predictions is that it's just a lookup table based on your estimated Vo2Max. There's no analysis of your training or anything else like that.

Also your Vo2Max estimate is only going to be as accurate as your HRMax setting is. If your HRMax setting is too high so will your Vo2Max estimate which in turn means you will get unrealistic race time predictions.

Also wish this was a bit more exact. 30:15 10k estimate, ran 31:26 last week. Even on a track, that's off by a considerable amount. I think it may get confused during interval workouts where the overall pace is reasonably fast but varied, and the HR very up-and-down. All of that said... if I can drop the 67 minute half it predicts next week, I'll send my race winnings straight to Olathe.

Also having issues with the "Training Status" indication; basically I'm only ever "Productive" or "Peaking," with no real reason to be one or the other. Seems like it fluctuates daily.

Also your Vo2Max estimate is only going to be as accurate as your HRMax setting is. If your HRMax setting is too high so will your Vo2Max estimate which in turn means you will get unrealistic race time predictions.

Hm... I am confused. So, I can fool Firstbeat only by setting my HRmax too high? What does all that science behind their algorithms do, then?

On a side note, If that indeed is the case, that could explain the overestimation. Like I said, I had my HRmax set to 185 - and I used to approach that number regularly earlier this year. However, after starting a training program at the beginning of summer and increasing the volume by quite a bit, my average HR dropped significantly, but I also find it hard to get even to 180 even on hard runs or intevals.

Also wish this was a bit more exact. 30:15 10k estimate, ran 31:26 last week. Even on a track, that's off by a considerable amount.

Well, I would call it as precise as I would ever expect from a fancy piece of software like Firstbeat. In my case, the prediction is 36:45 for 10k, but I can only do something like 42:30, I'm afraid (did a time trial last night!).

What about the people who have a really high HRmax (like me)? Then VO2max and race prediction are way off? I did runs with HR 196 (strap used) and 197 (wrist HR). As I haven't been completely blown up I supposed my HRmax at 200 and selected that value. With the FR35 VO2max was calculated at 41, then it went down to 39. Well, there were several runs that were a bit hard although they were slow an not long. The cardiologists didn't really find anything. Meanwhile with the FR235 it is back at 43, even 44 a short time. The race prediction, well...

Hm... I am confused. So, I can fool Firstbeat only by setting my HRmax too high? What does all that science behind their algorithms do, then?

As the saying goes “garbage in, garbage out.” Any algorithm is only as good as the data that’s fed into it so it needs at least one solid non-estimated value to seed the formula. The Firstbeat analysis bases most of its results on your HRMax and if it’s set too high it won’t think you’re working very hard at a given pace/HR and give a high Vo2Max, conversely it your Max is set too low it’ll think you’re working harder at the same pace/HR and give a low Vo2Max.

The Firstbeat analysis bases most of its results on your HRMax and if it’s set too high it won’t think you’re working very hard at a given pace/HR and give a high Vo2Max, conversely it your Max is set too low it’ll think you’re working harder at the same pace/HR and give a low Vo2Max.

So, fixing the exaggerated VO2max should be as easy as setting a lower HRmax, right?

What would be the most reliable way to set a proper value:
- set it lower and let the watch figure it out via automatic calibration?
- perform a guided threshold test?
- perform a filed test as recommended here here https://connect.garmin.com/en-US/help/how-to/

Run to exhaustion over a period of 25 minutes. The first 5 mins at warm up increasing speed gradually. The last 10 minutes should be difficult. The last 5 minutes should be very hard. The last 2 minutes very very hard. You should feel ill at the end. MaxHR is the highest recorded rate.

The Threshold test doesn't achieve MaxHR and is not intended to.

I guess I was wrong to say that "most" people get overestimated results, because I based that on forum complaints and random runners I know IRL, because most people won't complain if something is working and many people won't set their max HR properly. I shouldn't have generalized my own experience to "most people".

So I'll just talk about what I've seen personally.

I use a max HR (195) that I've seen multiple times over a 2 year period, and I've always gotten ridiculously high VO2 Max/race predictor results during that time. Just recently I hit 192 in a stamina workout despite not going 100% and being held back by some minor overuse injuries. I've even gone higher than that (196-201) when I was detrained and/or running in hot weather, but I decided not to use those numbers because I considered them outliers. If I were to take my VO2 Max numbers seriously, then I should've been able to run a 1:21 half at a couple of points during the last two years, and a 1:28 half today, but my PB is only 1:34. Then again, maybe I'm always being held back by nagging overuse injuries.

OTOH, Garmin has also estimated my LT threshold at 4:21-ish a few times over the past few years, which gives a much more realistic prediction of 1:34:30-ish for a half (close to what I've done a couple of years ago) and 42:23 for a 10k (very close to what I've done a few months ago). Might be relevant that I've never seen a LT threshold higher than that, and for sure I've never been in better shape than that, unfortunately. So from my POV, unlike the VO2Max/race predictor numbers, the LT threshold gives me a realistic ceiling on what I can do.

I think that both algorithms rely on max HR, so I don't know if the discrepancy could be explained by having a too-high Max HR alone. Maybe injuries could explain it, but why is the lactate threshold so spot-on under ideal conditions? (When I'm not so injured.)

Another complaint I've seen on the forums is that the LT test doesn't match VO2Max/race predictions. Granted that neither of them appear to be using the other as a "sanity check", you would nonetheless expect matching race predictions if they were both producing sound results.

Anyway, I hope changing your max HR fixes things for you. I haven't had much luck and I'm not sure that changing my max HR from 195 to 192 would change my HM prediction from an unreasonable 1:21 to a more reasonable 1:34-1:38.