Run-Down Forums

Dan · Posted: Tue Nov 06, 2001 11:55 pm Post subject:

Something new for you to play with:

http://run-down.com/statistics/calc.php

Gotta round out the offerings on the new Stats section. Smile

That page was inspired by the work I did for the Fantasy T&F League with calculating Purdy points for performances. I thought it would be good to extend it to predictions for different distances based on several models.

One model that I would definitely like to add to the page is the Hungarian Points (basically the IAAF Rankings system). It seems to be the closest to the Purdy Points in terms of accuracy across a wide range of distances (most of the models are inaccurate below 800m or a mile), but it is basically a bunch of lookup tables that are updated from time to time based on annual top performances, making it difficult to calculate. The only reference I have seen as to how to calculate the Hungarian points is on Patrick Hoffman's site.

For the curious types, take a look at the explanations page ("Click here for explanations and credits") on the new calculator. The VO2 Max predictions were a real doozy!

Dan

Micah Ward · Posted: Wed Nov 07, 2001 4:49 pm Post subject:

OK Dan, the simpleton is confused. I tried the calculator useing 5K in a time of 20:00. My VO2 Max showed a time of 19:36, what exactly does that mean? Also, my coresponding time for 10K is 39:12 (my pr by the way). How can a 5K time of 20:00 equate to a 10K time of 39:12?

Dan · Posted: Wed Nov 07, 2001 4:56 pm Post subject:

Micah Ward · Posted: Wed Nov 07, 2001 6:45 pm Post subject:

I tried again and it worked. The first time I entered 5,000 meters and this time I entered 5 kilometers. It seems to make a difference. It is interesting to see the wide variation of predicted performances when you carry it out to the 100K distance.

Dan · Posted: Wed Nov 07, 2001 9:14 pm Post subject:

Oh sure, cheat and blame it on the system! Oh wait, that's a different thread... Wink

Note that technically, "5,000" is not a number, as a computer program sees it, but "5000" is. I suppose I could place a note about that on the page...

And yes, the 100k numbers certainly do diverge. Really, it's the ends of the distance scale that separate out the models. If I were to put in more distances between the marathon (42k) and 100k, we might see a more gradual divergence similar to the obvious inaccuracies of some of the models below a mile or so.

That's why relying on just one prediction model doesn't work all that well. Smile

Dan

[ This Message was edited by: Dan on 2001-11-08 00:18 ]

Micah Ward · Posted: Thu Nov 08, 2001 4:59 am Post subject:

Is it cheating if you are too dumb to know what you are doing?

Dan · Posted: Thu Nov 08, 2001 9:11 am Post subject:

Taken in a different context, that's actually a fairly deep question...

Dan

Dan · Posted: Thu Nov 08, 2001 9:14 am Post subject:

Ok, what the heck. Instead of being stubborn and requiring valid input regardless of the sensibility of the "error," I made a quick change to allow distances to be input with or without commas. Minimal difference in processing time (a few nanoseconds kinda thing) and a bit more idiot proof (sorry Micah Smile

).

Dan

Micah Ward · Posted: Fri Nov 09, 2001 6:39 am Post subject:

If the shoe fits I'll wear it.

Discussion Forum

Share the Love!

What's New