It’s not science to predict race results and a bit of guess work and luck is always involved to get it right.
As with the right training it is a puzzle. In particular when it comes to ultra's.
Why bother? you may think. Just run your race and at the end you have your result. Accurately and scientifically measured.
But the good reason behind the predicted end time is that it gives us the pace we need to run.
Why bother? you may ask again. Any pace will do. Just run as fast as you can and as long as you finish it does not matter.
I am however convinced (and this is not just me, it is a widely accepted “fact”) that running an even pace from the beginning to the finish will bring the best results.
There is a discussion about the “fine art” of pacing going on and there are voices who claim that a negative split produces even better results. That is you need to run the second half of the race faster than the first half. But I would not go that far. I think a negative split would only matter in tactical races. Therefore I accept to run the second half slower than the first but still try to run as “even” as possible.
But all that requires to know the pace I will be running from the start in advance. Roughly. Of course there are other factors which can spoil the race. Injury, stomach, weather etc. But let's not take them into account just now.
There are quite a few race pace predictors or prediction methods available which are based on exising results for shorter races. E.g. your Marathon time. Even more reliable are methods based on 50k races or 50 milers.
I have a Marathon PB from my last Marathon in Berlin 2009 which is 2:46 (6:20 minute/mile pace).
If I take that time and apply it to one of the methods suggested by Noakes (Davies and Thompson): I would be able to produce a 7:49 for the 100k. A 10k split of 47 minutes. 7:34 m/m pace.
There are more methods like the following four:
Age Grading: 7:53
VO2 max: 6:41 (not realistic)
Riegel Formula: 6:54
Cameron formula: 7:42
The average of the above four: 7:17 (7:03 minute/mile pace). That would still be a very risky pace I would think.
I googled around and found another opinion from a Shawn McDonald.
“I like a factor of 2.8 times the marathon time equals your 100k time.”
That would give me a 7:44 which is close to other predictions
Based on my 90k time 6:55 (2010) and using the four formulas above I would get a 7:49. Which is again about 7:34m/m pace.
The fact that I actually ran a 90k in 2010 makes the prediction very reliable. So if I started off going 7:34m/m pace I would be quite safe.
But that is not what I am going to do. I have not given that 90k (Glasgow-Edinburgh “Double” Marathon) my full attention in training. Therefore I hope I can do better than that. In fact what I currently have in mind is a 45minute 10k pace, 4:30/k (= 7:14 pace).
That would give me a 4:45 50k split and a 7:30 finish.
In the heat of the day that may change a little. I will probably start a bit faster and later slow down a little. But above all that there is one golden rule which I have learned to follow in all my races: Running in the first half of the race needs to feel easy, comfortably. If I have to work hard in the first half I am too fast (or 'you' are too fast).
But I am not quite done yet. It is still a few weeks before the race and the plan may change. So watch this space :-)