Wednesday, 30 March 2011

Predict your Ultra Performance?

There are plenty of time comparators around for conventional race distances, which enable you to estimate for example, if I can do a 10k in 40 minutes then what should my marathon target be? I've always thought that it would be interesting to construct a similar comparator for ultra races so that you know what is a realistic time to be shooting for, but there is a bit of  a problem here.

Conventional races are run under standard conditions. OK, temperature, humidity, wind and so on are variable, and road race courses are not all the same but even in a marathon a world class time won't vary by more than 5 minutes or so, say 5% of the race time. Ultras vary in so many aspects such as amount of ascent, underfoot conditions, how much you are obliged to carry, how well you know the course, etc that the basic length of the course has limited meaning on its own. To take a couple of examples a bit removed from home ground, in the US where 100 mile ultras are more common than in Europe, the course record for the Rocky Racoon 100 in Texas is 12 hours 44 minutes (yes, that's 7:38 mile pace!) whereas the course record for the Hardrock 100 in Colorado is 23 hours and 23 minutes, almost double the time.

Ultras are so different that it is impossible to use any "scientific" parameters as a basis for comparison. Even trying to factor in a multiple of physical parameters such as length, height gain, ground underfoot, etc is a pretty hopeless task too because many of these are so variable in themselves  -  does 500ft of height gained over 3 or 4 miles on a smooth jeep track represent the same effort as the same 500ft gained in a quarter of a mile up a boulder and root-strewn hillside?

I've come to the conclusion that the only logical basis for comparison is the actual performances of runners in real races. Now from here on you can treat this as a proposal to be shot at or improved upon, or just a bit of fun, because I'm going to start making assumptions and judgements and it's unlikely that you'll agree with all (or any) of them but anyway here goes.

I've looked at the finishing times in a number of races that I'm interested in, and in most of which I've competed (though not always finished!). I decided that just taking the winning times was too variable because at the top end of the field there are a limited number of really talented athletes, and you don't get a consistent number of these turning up at every race. Similarly I didn't want to include a lot of athletes at the back end of the field because many people here are not wanting to optimise their time, they just want to complete the course and this gives rise to different strategies and performances. I decided in the end to concentrate on the top 20% of the finishers in each race, although I admit I have weighted the results depending on my interpretation of the overall quality of the field, how many results are representational, and so on. I could go on a bit about the details but I'll leave it there for now and just go on to the results.

The races I chose to compare were

D33  -  33 miles  -  flat (?)
The Wuthering Hike  -  32 miles  -  1300m of ascent
Round Rotherham  -  50 miles  -  800m
The Highland Fling  -  53 miles  -  2600m
Hardmoors 55  -  54 miles  -  2700m
West Highland Way  -  95 miles  -  4500m
Lakeland 100  -  104 miles  -  6300m
UTMB  -  102 miles  -  8900m

I took results from 3 years if possible, 2 years minimum, and only from where the race was substantially the same (eg UTMB results prior to the Flegere inclusion would skew the times downwards quite a bit). I would have liked to include the Hardmoors 110 but I don't think there have been enough finishers yet to give meaningful results.

So what you see below is what I think is a reasonable and simple comparator for athletes targeting a race they haven't run, using information from one that they have. The first column gives the time ratios (which if you consider directly in hours gives a nice range of times for a typical mid-pack performer) and the second for interest gives my best times in the races I have completed.

Event                           Time Ratio           AJC pb

D33                               5
Wuthering Hike              6                        5-45
Round Rotherham         9                        9-04
Highland Fling              11                       10-10
Hardmoors 55             12                       11-21
West Highland Way     24                       23-35
Lakeland 100               35             
UTMB                           40

So to use this (should you want to!), say you have a Highland Fling time of 12 hours and want to predict your West Highland Way time, the calculation would be 12 divided by 11 multiplied by 24, giving a target of 26.18 hours or around 26 hours 11 minutes.

The proviso of course is that you have to be trained for the distance. If you use a conventional predictor and put in say you can do 40 minutes for a 10k, you find that your marathon prediction is around 3 hours 7 minutes, of course only achievable if you have put in the right training to run a marathon. Similarly here, an 11 hour Highland Fling won't get you a 35 hour Lakeland 100 unless you have trained for it because they are very different events.

My own times seem to stack up OK, showing that I'm better at 50 milers than 100's, which I know and that the Highland Fling is probably my best because I have run it several times and now know how to tackle the course to suit my own strengths and weaknesses.

As I said, just a bit of fun.

As a postscript, I took all the height gains for the longer races from the "official" qualifying race list on the UTMB website, which shows the UTMB itself in 2008 and 2009 as being 163km and 8900m ascent. Those who were there will know that the 2010 race was reduced to about half length because of the weather conditions, but the 2011 race is listed as being 166km and 9500m of ascent  -  they must have found another hill to climb this year!


John Kynaston said...

Interesting Andy.

If I put in my 2009 times ...

Fling 9.49 ....

9.8/11 x 24 = 21.38

So I should have run the whw in 21hrs 23secs and in fact I did it in 19hrs 52mins.

So either I ran well or your model (for me anyway) needs some adjusting????

The thing that I really want to be able to do is an age comparison.

If a 35 yr old runs the whw in 17.55 and a 52yr old runs it in 19.55 ... which is the better performance age graded wise?

Or if I want to compare my times as a 52yr with my times as a 62yr old.

As you say all good fun and no doubt only of interest to you and me!!

See you at the Fling .... but hopefully not in the forest above Crianlarich like last year!

graeme reid said...


Your model works well for me. 11 hour fling relates to a 24 hour WHW (actual 23.25)so not far off. In my mind I was hoping for a 40 hour UTMB - your model is bang on.

See you in 4 weeks