Xmas Present for Stats Guys - Possible Track Distance Times Equivalents

for women elites :

2'00.00 = 2'34.02 = 4'02.00

1'59.00 = 2'32.77 = 4'00.13

1'58.00 = 2'31.52 = 3'58.25

1'57.00 = 2'30.27 = 3'56.38

1'56.00 = 2'29.02 = 3'54.50

1'55.00 = 2'27.77 = 3'52.63

1'54.00 = 2'26.52 = 3'50.75

1'53.00 = 2'25.27 = 3'48.88

this just confirms to me how sh!!ty womens' 1500 has been for years

most seasons, you get a 1'55/1'56 lead ( 1'54-flat when pamela was around ) & = best 1500sb by someone shouda been 3'52/3'53

if anyone has doubts, the wr of 3'50 by chinese was run off an outrageous 1st lap & shouda been well below 3'50 at even-pace, somewhere in the 3'48s, so it is on a par with kratoch's 1'53.28wr

ventolin^3 wrote:

so here it is

1'45.00 = 2'15.27 = 3'33.88

1'44.00 = 2'14.02 = 3'32.00

1'43.00 = 2'12.77 = 3'30.13

1'42.00 = 2'11.52 = 3'28.25

1'41.00 = 2'10.27 = 3'26.38

1'40.00 = 2'09.02 = 3'24.50

Doubling the 800 times listed will yield:

3'30.00

3'28.00

3'26.00

3'24.00

3'22.00

3'20.00

Then, subtracting those values from the 1500 times listed yields:

3.88

4.00

4.13

4.25

4.38

4.50

Based on this pattern, an 800m runner will a 2'00 time should run a 4'00 1500m. I'm not sold.

Venty run this female runners numbers thru your magic twanger thingie.

1:58.7-800m

4:00.2-1500m

4:21.8-mile

8:53-3000m

ventolin^3 wrote:now, next step is to get the equivalent for 3k, but i'm back at one of 3 ( or 4 ) possibilities, again, one of which is correct, but even i have to admit, i'll need some input to nail the right one

i'll post the 3k possibilities later

you have to go to komen & 7'20.67 & examine the circumstances around it

he had run a "tired looking" 13-flat less than 2/7 before & in rieti he was offered a suicidal 57.5+ opener

fully rested & with sensible even-pace he wouda smashed 7'20, perhaps even as fast as 7'18-high, but is this = to a 3'26.38, hicham running 3'26.00wr just under 2y later ???

i still think whatever he couda run ideally was still fractionally inferior to anything = to 3'26.38

i have some possibilities, one of which, again, i think is the correct =

& i have the corresponding = for associated 3ks for 3'30.13 linked to each to help more in deciding :

3'26.38 / 3'30.13 =

7'20.06 / 7'27.56

7'19.21 / 7'26.71

7'18.36 / 7'25.86

7'17.51 / 7'25.01

one of these is the 100% correct equivalency IMO; the underlined is my gut-feeling, but i'd need some input

( worth noting saidi-sief was a 3'29.5 guy who ran 7'25-flat off a jog for 2k & then blasted a 54/55 last lap for it, so anything ~ 3'30 & 7'25 are well established bedfellows )

did you mean the following with 3's replacing 7's in the right hand column?

3'26.38 / 3'30.13 =

7'20.06 / 3'27.56

7'19.21 / 3'26.71

7'18.36 / 3'25.86

7'17.51 / 3'25.01

regardless, another very interesting ventolin thread.

I tend to think of ventolin's 3k equivalents as too fast, because the times offered at a frequently contested race have rarely been matched.

the next step is to find the 5k = to

1'41.00 / 3'26.38 / 7'18.36

this one is giving me helluva lotta trouble

i can't get any big gaps, because the method used now shortens the intervals to a paltry ~ 0.5s

it is one of

12'36.51

12'36.00

12'35.50

12'35.00

12'34.50

12'33.99

this one needs to consider merits of komen/geb/kennster

last 2 were certainly capable of 12'36.00 at even pace in their wrs, kennster maybe into 12'35-low

komen maybe only 12'36 at perfect-pace in '96 if that had been a 5k in rieti & he was probably better in '97

i'm going to narrow it down to

12'34.50 / 12'35.00

on basis best times/talent shoud eventually evolve to blue-riband event of 1500

but i will need to go look at iaaf tables here for some input

it coud be that the big-3 over 5k were better than 3'26.38 & it coud be one of the higher figures

this post i will need to come back to a few times...

Why are you having a conversation with yourself? It's clear from other posts that you and Ventolin ^3 are the same person.

jjjjjjjj wrote:I tend to think of ventolin's 3k equivalents as too fast, because the times offered at a frequently contested race have rarely been matched.

of all the < 7'27s ever run, only komen/hicham ever got a split under 4'56

they were given 4'53+, each in wild circumstances there

( i saw latter live tv, former unfortunately only later on weekend highlights )

here is a 7'26 off an appalling 2k in some of worst conditions ever :

i suggest you offer what a 4'49iwr holder couda run for 3k here with perfect weather/pacing off 4'53 split & who knows, 12'37/26'17 endurance to hang on, rather somewhat like mo off 2'33+ at the bell in monaco ???!!!

ventolin^3 wrote:but i will need to go look at iaaf tables here for some input

it coud be that the big-3 over 5k were better than 3'26.38 & it coud be one of the higher figures

this post i will need to come back to a few times...

had enough data points when including 1k/2k ( latter not mentioned ) into curve-fitting software

i was slightly out with the 5k, meaning kennster's 12'37+ wr with 57 finish is likely good/better than hicham's 3'26wr !!!

however, the 10kwr is weak

prelim final line ( i'm happy 100% upto 3k & i'll take the 5k/10k )

1'41.00

3'26.38

7'18.36

12'35.50

26'07.50

"Equivalent" in what way? In how "good" they are? The issue is this isn't a well-defined problem. I think in order to determine how "good" they really are you have to be crunching more statistics of fastest actual performances in history.

jjjjjjjj wrote:

I tend to think of ventolin's 3k equivalents as too fast, because the times offered at a frequently contested race have rarely been matched.

Yes, I've always thought his 3k equivalents have been too fast. I don't really think of the 3k being an infrequently run race (a 2k I would agree with that). I've also thought his 10k estimates were too generous.

Mr. Obvious wrote:

jjjjjjjj wrote:

I tend to think of ventolin's 3k equivalents as too fast, because the times offered at a frequently contested race have rarely been matched.

Yes, I've always thought his 3k equivalents have been too fast. I don't really think of the 3k being an infrequently run race (a 2k I would agree with that). I've also thought his 10k estimates were too generous.

His equivalents are based on fitness required to run that time at a perfectly maxed performance. I think the reason that 3k and 10k equivalents seem too fast is that these races are harder, mentally to max performance in (at least IMO). Staying extremely focused for 25 laps on a track is very very tough, and staying extremely focused for 7.5 laps at nearly 4 lap race pace is perhaps even tougher. If 3k and 10k were contested more frequently, people would get ever so slightly better at racing them, and the records probably would be slightly faster. Imagine if a 10k race had ever occurred in which 2 competitors ran the kind of race that El G and Lagat did in 2001. I don't think 5 seconds faster than 26:17 would be impossible. Of course, I'll never be anywhere near 7:20 or 26:17, so I might be making all of this up.

Raptured wrote:

His equivalents are based on fitness required to run that time at a perfectly maxed performance. I think the reason that 3k and 10k equivalents seem too fast is that these races are harder, mentally to max performance in (at least IMO).

I don't think his equivalents are based on fitness at all. A system based on fitness to run a race would be something like Jack Daniel's VDOT system--which ties performance to a measurable physiological parameter. Ventolin's tables (and this is not meant as a criticism) are based on formulating and equation and drawing a curve to approximate performance.

It is possible that the 3k and 10k are harder races mentally, in which case I would say that it points out the limits of such mathematical modeling (that it can't account for such factors). It may also be that they are less raced (certainly true of the 10k today). I've just seen Ventolin's calculator give credit for fitness between 7:18-7:7:24 and

This is a HORRIBLE display of statistics and is dribble math. Hardlooper had it right as well. You haven't even clearly defined the question and the goal of this exercise. Everything is clouded in vague words like "equivalent fitness". And to vVenties discredit he doesn't even use the word fitness in the original post, just equivalent. Equivalent in terms of WHAT?

ventolin^3 wrote:1'41.00

3'26.38

7'18.36

12'35.50

26'07.50

more relevant, as wrs are rare, woud be the equivalents for ~ 3'30, as that is what we normally associate nowdays as the barrier for an elite - medal challenging ability

the equivalents i get are :

1'43.00

2'12.77

3'30.13

4'47.73

7'25.86 !

12'48.00 !!

26'32.57 !!!

this was somewhat of a surprise, but not so much after reflection

the 10k is really a p!ss-weak event & even pretty much unknown guys like kemboi/dinkessa have gone 26'30

if the 10k had number of "3'30 talent" running 10k regularly ( instead of all on the roads ), the equivalent benchmark for it shoud not be even 27'00 or even 26'45, but a clocking much nearer 26'30

for a more common ~ 1'44 / 3'32 equivalents :

1'44.00

2'14.02

3'32.00

4'50.23

7'29.61

12'54.25

26'45.07

this just confirms how far behind 5k & especially 10k have been lagging behind 1500m

for 1'45 equivalents :

1'45.00

2'15.27

3'33.88

4'52.73

7'33.36

13'00.50

26'57.57

this looks good

we actually have some almost equivalency for oft-touted 13-flat = 27-flat

calculator indicates 13-flat is fractionally better than a 27-flat

still indicates how fortunate we are with quality of 1500 guys worldwide for decades & that the longer distances have a lotta catching up to do

Predicting running times is a simple regression problem. Take, say, the top 100 men for which we have recorded 800, 1500 and mile times. You need a 4th-degree polynomial to fit through the 3 data points of any 1 person's times. Thus you can use a 100x4 matrix to represent the coefficients of the polynomials that represent all 100 persons' times. Plug it into MATLAB's nonlinear regression function to get the coefficients for the master equation that predicts a person's 800 mile based on their 1500 and mile, or their 1500 based on their 800 and mile, or their mile based on their 800 and 1500.

unfortunately regressions don't factor in assumptions such as elite circuit holding same number of 1k or 2k or 10k races/y as 800 or 1500 & that the level of talent is same in each event & has specialised in each event, i e a 1k or 2k specialisation as much as an 800 or 1500 one, etc , etc