matsalo
Runalyzer
Posts: 46
Joined: Mon 4. Jun 2018, 11:30

### Marathon shape variables/parameters

Hi,

Just a quick question in order to understand the marathon shape function. As I understand it, it has a goal of 70km/week in general and 28km/week for long runs, and the user controlled parameters in settings controls the time span to evaluate at and the definition length of a long run.

Therefor I wondered if these values for weekly goals also could be tweaked with, or are these static and locked parameters?

Don't know how senseable it is to tweak with them, but the reason to ask was more of curiosity to understand the function.
0 x

laufhannes
Core developer
Posts: 777
Joined: Mon 29. Jul 2013, 20:59

### Re: Marathon shape variables/parameters

They depend on the optimal marathon prognosis (based on the current athlete's vo2max shape) and are derived roughly from common marathon training plans.

The exact calculus:
weekly mileage = pow(EffectiveVO2max, 1.135)
long run = log(EffectiveVO2max / 4.0) * 12.0

log is the natural logarithm here
This results in the following (vo2max / mileage / long run):
30: 47 km / 24 km
40: 66 km / 28 km
50: 85 km / 30 km
60: 104 km / 32 km
70: 124 km / 34 km

This is not adjustable by the user so far.
0 x
User help -- short questions via Twitter, Facebook

matsalo
Runalyzer
Posts: 46
Joined: Mon 4. Jun 2018, 11:30

### Re: Marathon shape variables/parameters

It makes perfectly sense to keep it the way it is.
0 x

torbich
Neuling
Posts: 2
Joined: Wed 19. Dec 2018, 00:10

### Re: Marathon shape variables/parameters

Hi everybody!

Looks like there is a problem with the number of points given for a long run for Marathon shape. The calculator states that

The points for your long runs are weighted by time and quadratic in distance. That means, a long jog yesterday gives more points than a long jog two weeks ago and a 30k-jog gives more points than two 20k-jogs.

But the actual points looks more like to behave as distance^4. For example, 31km run gives me exactly 2 points. So, 16km run should give about 0.5 points, but actually it is even not shows up meaning it has less than 0.2 points. 21km gives only 0.4 points, etc. Some basic estimate suggests power low ^4 instead of ^2, or simply double squating in the code.

Is it designed so or there is an error in the code?
0 x

torbich
Neuling
Posts: 2
Joined: Wed 19. Dec 2018, 00:10

### Re: Marathon shape variables/parameters

PS: looked at source code and find the following line:
(2 - (2/'.\$this->DAYS_FOR_LONGJOGS.') * FLOOR( ('.\$timestamp.' - `time`) / '.DAY_IN_S.' ) )
* POW((`distance`-'.\$this->MIN_KM_FOR_LONGJOG.')/'.\$this->getTargetLongjogKmPerWeek().',2)

now it is clear: only those kms which are above 13 are counted!
0 x

JoelH
Runalyzer
Posts: 66
Joined: Fri 30. Jun 2017, 15:33

### Re: Marathon shape variables/parameters

torbich wrote:
Wed 19. Dec 2018, 00:45
now it is clear: only those kms which are above 13 are counted!
You can change that minimal distance in your general settings =>
condif.jpg (96.52 KiB) Viewed 1103 times
0 x