Pace calculators don't work

same thoughts wrote:

they all assume ideal mileage. what works for marathon doesn't work for 5k.

too many variables. i think comparing WR is the best way to get equivalents.

Some of you have bad reading comprehension. At least one of the guys has said that mileage is irrelevant in his/her calculator, and none of the calculators being discussed are for equivalents.

You may be right, perhaps the 1500m will end up being my best distance.

To Ventolin, if you take runner X who has run 50 for the 400 and 1:50 for the 800, this is what yours and stats gangstas calculators give:

Ventolin:

1500 3:43.3

3000 8:04.0

5000 14:12.7

Stats Gangsta:

1500 3:45.2

3000 8:12.8

5000 14:34.1

So, you can't both be right.

which one did you have to pay $1.61 for ?

what is likely better : pay or free ?

now try both with Paula's 58.5

( albeit unknown date )

& same 2'02.8 for the calculator i use & from other guy

what does his/hers predict for 10k off

58.5 / 2'02.8 ?

now look up Paula '02...

Jundo wrote:

You may be right, perhaps the 1500m will end up being my best distance.

To Ventolin, if you take runner X who has run 50 for the 400 and 1:50 for the 800, this is what yours and stats gangstas calculators give:

Ventolin:

1500 3:43.3

3000 8:04.0

5000 14:12.7

Stats Gangsta:

1500 3:45.2

3000 8:12.8

5000 14:34.1

So, you can't both be right.

What's your point? People have opinions, they will always differ.

ventolin^3 wrote:

which one did you have to pay $1.61 for ?

what is likely better : pay or free ?

now try both with Paula's 58.5

( albeit unknown date )

& same 2'02.8 for the calculator i use & from other guy

what does his/hers predict for 10k off

58.5 / 2'02.8 ?

now look up Paula '02...

Sounds like this calculator is just built off of Paula's times???? Not a good indicator.

Malmo, that was my point exactly, they are different opinions, not facts. They are based on anecdotal evidence not science.

Also, Ventolin previously said that both his and Stats Gangsta's calculators would work for the same person, given 'correct user input'. I was proving that to be false.

Jundo wrote:

Malmo, that was my point exactly, they are different opinions, not facts. They are based on anecdotal evidence not science.

.

Jundo, where are you getting that idea from? I ask because my calculator is indeed based on actual science and physiology. From what I can see, Daniels' is based on statistics, and Ventolin's is based on some mix of statistics and personal opinion. So where did you get the belief that these are all based on "anecdotal evidence"?

I did not know your calculator is based on science. Most are not

Jundo wrote:

I did not know your calculator is based on science. Most are not

How do you know what most of the calculators are based on? Are you just making things up?

stats.gangsta_the_real_1 wrote:

Jundo, where are you getting that idea from? I ask because my calculator is indeed based on actual science and physiology.

Exactly HOW is your calculator based on "science" and "physiology"? This will be interesting.....

I just don't know of any experiment-based pace calculators.

Ones where somebody has taken maximum efforts by athletes and mapped them to find a consistent pattern across many different runners.

Jundo wrote:

I just don't know of any experiment-based pace calculators.

Ones where somebody has taken maximum efforts by athletes and mapped them to find a consistent pattern across many different runners.

That says more about your ignorance of the subject matter than the lack of experiments and pattern finding.

Not sure if we should talk about "opinions" or "facts" or "anecdotal evidence" or "science". When discussing "calculators" we are talking about theoretical models, based on some assumptions. How well they map to the real world, depends on how valid the assumptions are, at the time of mapping.Looking at the "Jundo Predictor", it is based on one simple assumption: that 3 race times (two inputs, and one output) can be fully characterized by 2 "constant" parameters: I'll call them "speed" and "resistance (or endurance)". All other factors are negligible. What can possibly go wrong? It's not a bad basis, but:1) When you train, and how you train, changes your personal fitness, including speed and resistance. These factors are not "constant" over time. Therefore, you need to chose your inputs wisely.2) Things like "heat", or "race tactics", or "running 3 laps in lane 2", or "drafting", etc., can impact your time independently of your current fitness. In order to make the calculator useful with these times, you need to compensate, adjusting the real times to eliminate these biases.Giving "correct user input" means attempting to minimize, or better yet, compensate for all of these variables, so that all three times are the result of the expression of the same speed and resistance. (This is where "ventolin" spends a lot of his energy -- quantifying the cost of heat, drafting, running in lane 2, uneven splits, etc.)I can't comment about "stats.gangsta", but I will briefly mention "competing" models to highlight an important difference:1) Purdy -- based on speeds from 40m to 100km, compiled in 1936 from real performances (Portuguese scoring tables)2) Daniels' VDOT table -- based on O2 measurements and performances from many athletes, and number crunchedPurdy is better at comparing two specialists in their event, like Bolt to Bekele, while Daniels is better for an athlete who wants to predict a 10K race time from a previous 5K race time. While both might be used for the same purpose, they have different strengths and weaknesses. They are not the same tool.Just like you don't use a hammer to turn a screw, you need to pick the right tool for what you want to do, and you need to understand its limitations.Every model has errors. (This applies to everything -- not just "pace calculators"). It would be lazy to assume that "pace calculators" fail, because the error is always 0. The usefulness of any pace calculator depends on the user's ability to model, and compensate for the errors, rather than assuming there are none. You say "ventolin" and "stats.gangsta" can't both be right. In fact, in the real world, they will both usually be wrong. How wrong depends on the source and magnitude of errors in the inputs, and how well the assumptions of their models hold up when going from inputs to output.

Jundo wrote:

Malmo, that was my point exactly, they are different opinions, not facts. They are based on anecdotal evidence not science.

Also, Ventolin previously said that both his and Stats Gangsta's calculators would work for the same person, given 'correct user input'. I was proving that to be false.

Not sure about needing that :23.7 to run 1:57.9 - I ran 1:57.5 for 800m, and couldn't have broken :24+ to save my life, I fact I'd struggle to break 25. (I could only run :12.0 for 100m with a flying start).

I was pretty much like the OP - best at 800m and 1500m with drop offs either side.

If you are a 800m/1500m type I don't see why as say a 51.0 400m runner, you couldn't go :54/57 for a 1:51.00.

I think it's just a profile of a strength type 800m runner. Relatively slow sprinting speed, but not enough endurance for a good 5000m.

A thought out a well-presented post.

'400 0:48.76

800 1:50.5

1000 2:23.81

1500 3:51.3

1-mile 4:11.01

3000 8:36.51

2-mile 9:19.95

5000 15:27.92

10000 34:00.12

half-mara 78:49.24

marathon 170:59.76'

Read more:

http://www.letsrun.com/forum/flat_read.php?thread=6748286#ixzz3mK4Vovyz

That just shows one thing. Stick with the 400m!

Jundo wrote:

I did not know your calculator is based on science. Most are not

Those that claim to be don't reveal their deductions anyhow so you can disregard their claims.

You can't convert between sprints, MD and LD. The energy systems used are too different and there is nothing in race times to indicate how they're being used in a particular performance. In 20/20 hindsight I deduced your 51 second ceiling from your 55/56 split, comparing you to the 800/400 ratio of similarly low-differential splitters, but that was just luck. Someone else with the same splits might easily run a 48 off of a superior drive phase and top speed which they don't use in the 800.

As I wrote earlier, from 600 to mile, times increase over distance on an exponent of ~1.135. That's from comparisons of MD runners "equally trained" at various distances. It's not precise, but it's more precise than the coach's cliche "add 4 and double it," and more transparent than the oracles that give you no formula at all.

what was Makloufi's last 1000, about 2:13... I'd estimate the equivalent 1500 as

133 x (1500/1000)^1.135 = 210.7

3:30.7 is about as fast as he's run outside Monaco, and even 2 seconds off of his Monaco PR ain't bad.

1:50.5 is better than 48.7

I think that statsgangsta is the only person who understands this: different people are better at different events, and putting in two values is enough to get that "b" exponent in the riegel formula.

b = ln(d1/d2) / ln(t1/t2)

If one has two times, you can find the ratio between paces as the distance of the races changes. McMillan, jackD, and all of those calculators assume that everyone is the same. We aren't. I actually made an excel sheet that incorporates age, training, and "best event", but I found that most people have raced over two distances within reasonable timeframes, thus allowing the two-point approximation.