# Interesting article on the (in)accuracy of race time predictors (Read 525 times)

GC100k

Using world records, I get an exponent of 1.07 for 5k up to marathon and 1.08 if you drop it down to 800m.

Back to the author's original point, it makes sense that using 1.15 for half to full prediction may not predict what you could or maybe should do, but since that's about what most people did, it would, for most people, predict what you're likely to do.

I must be a bit of an oddball (unless I am reading things wrong -- highly likely!!!).

I ran 2:50 for a marathon in March (with a poor second half).  Based on the numbers they use, I would have to have a sub 1:18 half (which I have never run).  To add to this, the year I ran my fastest marathon (2:44:49) was six months after I ran my half marathon PR (1:18:02).  I guess I am more of a distance guy since my marathon PR points to a sub 1:16 half...

And you can quote me as saying I was mis-quoted. Groucho Marx

Rob

Not at all. Same guy, same week, similar weather conditions, but adequately trained for 5k, not adequately (or less adequately) trained for marathon. Even if you were to say you were equally inadequately trained for each distance, I still wouldn't expect the formula to work; it's predicated on you being adequately trained for both distances. Failure modes for 5k and for marathon are very different.

Hard to imagine anyone ever tries to extrapolate from 5k to marathon (or even half). There are certainly some people for whom 1.06 will work fine, but for many people the coefficient needs to be infinity because they are undertrained to the point of automatic DNF. And everything in between.

Dave

Yes. My point is, it's inappropriate to use the formula at all in such cases. There is relevant physiology that is captured to some extent, over some range, by the original formula. Tweaking the exponent to account for varying degrees of undertraining is trying to shoehorn in applications very artificially. IMO. Certainly, it makes little sense to do as the author proposes and simply modify the formula, keeping it solely based on input half time. That's throwing out important information -- how well trained you are.

Yes. My point is, it's inappropriate to use the formula at all in such cases. There is relevant physiology that is captured to some extent, over some range, by the original formula. Tweaking the exponent to account for varying degrees of undertraining is trying to shoehorn in applications very artificially. IMO. Certainly, it makes little sense to do as the author proposes and simply modify the formula, keeping it solely based on input half time. That's throwing out the most important information -- how well trained you are.

FYP.

Dave

Craig

Sure, if you're woefully under-trained then 1.15 is going to make a better prediction that 1.06. That should be fairly obvious.

+1. 1.15 is ridiculous for anyone who has trained appropriately. And the guy seems to want people to use his new, improved calculator. Honestly, I will be pretty annoyed if this is now widely used to justify crappy performances as equivalent.

+1.

Anecdotally, 1.06 was eerily accurate for me when well trained.

Incidentally, I ran a poorly trained half, followed 4 months later by a poorly trained full, and the 1.06 was also eerily accurate.

Dave

This thread reminds me of David Liu's discussion, in trying to set his goal marathon pace, based on calculators.  It has been interesting to follow.  3:30 seems a little slow, 3:20 seems a little fast.  One calculator says X, another says Y, what should the pace be?  But that is the whole point: estimates are better based on training, on the course, and on the weather, not on a calculator.  Anyone can come up with a bullshit calculator, even me, I could make a calculator that factors in seasonality and Zodiac sign.

What if I was born prematurely and feel stronger affinity for the sign under which I *should* have been born?

This thread reminds me of David Liu's discussion, in trying to set his goal marathon pace, based on calculators.  It has been interesting to follow.  3:30 seems a little slow, 3:20 seems a little fast.  One calculator says X, another says Y, what should the pace be?  But that is the whole point: estimates are better based on training, on the course, and on the weather, not on a calculator.  Anyone can come up with a bullshit calculator, even me, I could make a calculator that factors in seasonality and Zodiac sign.

"When a person trains once, nothing happens. When a person forces himself to do a thing a hundred or a thousand times, then he certainly has developed in more ways than physical. Is it raining? That doesn't matter. Am I tired? That doesn't matter, either. Then willpower will be no problem."
Emil Zatopek

What if I was born prematurely and feel stronger affinity for the sign under which I *should* have been born?

You would fall into the category of people born on the cusp, and would have an additional factor correcting that factor

Eric's calculator right here defaults to 1.08, but you can plug in whatever coefficient you want.  Make the method fit the data, as did the author of the article.

The point is, for most people the marathon is a whole different race from all shorter races, because it's the one where you hit the proverbial wall, when your body runs out of its stored fuel.  At that point, the formula may well fail, because a bunch of other variables suddenly come into play.  How well trained are you?  How much experience do you have?  How determined are you?  How bad is that blister hurting?

Well at least someone here is making relevance to the subject. - S.J.

scappodaqui

rather be sprinting

Wouldn't a more ideal race prediction system have to do with runners' knowledge of their specific thresholds?

I've always had a real problem with the idea that someone who runs a 25-minute 5k and someone who runs a 15-minute 5k are *really* using the same energy system.  One of them is a lot closer to VO2max pace than another!

Same for a 2:15 half marathon vs. a 1:15 half--one is what Daniels calls 'threshold pace' (aka the pace that feels like your legs are burning off for the whole race) and one is run at the speed at which the better runners/elites run a full marathon.

I guess what I'm saying is there really should be a physiological measure of performance instead of a purely numerical one.

Or at least 'how does my 25 minute race time predict my 2 hour race time'--going by effort over time vs. distance?

PRs: 5k 19:25, mile 5:38, HM 1:30:56

Lifting PRs: back squat 176 lb

Wouldn't a more ideal race prediction system have to do with runners' knowledge of their specific thresholds?

I've always had a real problem with the idea that someone who runs a 25-minute 5k and someone who runs a 15-minute 5k are *really* using the same energy system.  One of them is a lot closer to VO2max pace than another!

Same for a 2:15 half marathon vs. a 1:15 half--one is what Daniels calls 'threshold pace' (aka the pace that feels like your legs are burning off for the whole race) and one is run at the speed at which the better runners/elites run a full marathon.

I guess what I'm saying is there really should be a physiological measure of performance instead of a purely numerical one.

Or at least 'how does my 25 minute race time predict my 2 hour race time'--going by effort over time vs. distance?

Your question doesn't make sense on two levels:

1) Most importantly: a race is a measure of how fast a given human being can cover ground. It's not a contest of physiological systems. "How fast a runner can cover ground" in other words is no less precise of an indicator of fitness (and in many ways much more precise) than VO2max, LT, or any other physiological marker.

and

2) Slower 5kers will also be relatively slower marathoners -- so the relative duration of the event is actually a factor that falls out of the analysis when comparing runners of different capabilities.

Options,Account, Forums

From one point of view, a good prediction system is one that tends to predict numbers that are accurate for a lot of runners in some subset.

So, if one can agree on a subset (eg, those well-trained for the distance), then one could look at it as a curve-fitting exercise.

It's a 5k. It hurt like hell...then I tried to pick it up. The end.

GC100k

I've always had a real problem with the idea that someone who runs a 25-minute 5k and someone who runs a 15-minute 5k are *really* using the same energy system.  One of them is a lot closer to VO2max pace than another!

How do you know that?  That's a common misconception about slower runners that may have some truth but isn't the whole story.  Ed Eyestone never did get that someone could be running hard and be running more than 6 minutes per mile.

I will grant you that just about everyone running a 15 minute 5k is running really really hard.

But the 25 minute 5kers include some folks jogging at their regular pace and others going all-out.  When I run a 23 minute 5k I'm sprinting at "OH MY GOD I"M GOING TO PUKE, EXPLODE, AND DIE!!!" pace.  I'm at whatever threshold you could name.

So it's not automatic that slower runners are running in a different "system" than faster runners.

And different distances don't mean different systems. The world records from 1500m to marathon fit remarkably close to the 1.07 exponent.  The sprints (100m, 200m, 400m) fall a little below (faster) it and the 800m very slightly below.  The 20k, 25k, and 30k are a little above (slower), which you would expect because they're seldom raced.  The half-marathon and marathon records ratio has an exponent of 1.08.  Everything aerobic seems to extrapolate according to similar ratios.