Just chating with somebody on another forum about this and it gave me an idea. Im sure there would be one or two problems to overcome but I wonder if this is possible?
Originally Posted by Gary5k View Post
Its quite interesting, though tbh i just look up previous years results & find similar runners to me and see what time they do via some po10 stalking. That usually gives me an idea what the course is like. And is more accurate than some points system magic.
Good idea!
It makes me wonder if you could input all of that data into a computer and create an algorithm to calculate your aproximate time based on all of the data on the Power Of 10?
It could be suprisingly accurate if possible! I think it could even be the most accurate race time predictor ever as its based on a number of human beings therefore it would give you an average of all of the pro's and con's of human beings plus conditions to give you a very close approximation!
You just need a huge database of times and courses and you could then use every runner within say 10 seconds of your given time and distance within say 2 weeks of running the course. (optimal
times, time frames and previous courses could be looked at and developed)The computer could then calculate an average. You could adjust the parameters if your search was lacking in numbers of hits. You could develop an optimal accuracy through trial and error?
Possible? Surely!
Ed could do this easily!
Thats something for him to look at when he get the time to fit it in!
Well I thought it was an interesting idea anyway. If it would work in practice or not is another question entirely!
Edited by user 08 December 2016 15:53:03(UTC)
| Reason: Not specified