torsdag 8. august 2013

How do I predict my marathon time?

I have a brother in law who is every bit as crazy as me when it comes to running. He is aiming to get a few more marathons under his belt before he retires, and he has begun to combine holidays abroad with running events: Provence Luberone Marathon, Prague Marathon and now he has signed up for his second Lake Garda Marathon in October. A few family members will join as support team (really just enjoying the holiday), but I decided to join him in the race as well.

This will be my first marathon, and although I ran my first mountain ultra race earlier this year it will be totally different to do 42k on asphalt. That aside, I have started to think about what kind of race time I should target.
There are a couple of race time calculators that let you estimate your race time based on another race time you have done for a different distance. I did my first real half marathon last year and was very pleased with my time: 1:28:29. Based on this RW’s race calculator (which more precisely should be called Pete Riegel’s race calculator since it is based on a formula by Pete Riegel) predicts that I should be able to do a marathon in 3:04:29. The Mackmillan running calculator is a bit more pessimistic (or realistic?) and predicts 3:06:13. Then I found another method called the Yasso 800’s. I went to my local track and ran 10 x 800 meter intervals as fast as I could. My worst 800 meter time was 2 minutes 56 seconds. According to the Yasso 800 theory this should indicate that I’m capable of running a marathon at 2 hours and 56 minutes: Just convert minutes to hours and seconds to minutes. According to Coach Jeff at RunnersConnect you should add approximately 5 minutes to this time, giving me a prediction of a 3:02:00 marathon – still a fairly optimistic estimate. (By the way, I include a blog reply that I posted on Coach Jeff’s article at RunnersConnect below).

I suspect that my biggest problem will be the unfamiliar hammering from running on asphalt. I’m not going to do any training on hard surface in advance, and I will quit the race if I start to feel serious pain in a knee or a hip. I’m not going to ruin myself on this event!  Apart from that, stamina should not be a problem based on my experience form the mountain ultra race. My personal estimate, based on my gut feeling, is that I could be able to run at best a 4:30/km pace, resulting in approximately 3:10:00 as my marathon finish time. Allowing for some unforseen problems and delays and the fact that it is my first marathon, I have decided to set a target for 3:15:00. Now it remains to see who has the best predictive powers; Pete Riegel, Mackmillan, Bart Yasso, Coach Jeff… or me!


Hi Jeff, thanks for an interesting article!
Your link seems to have 'expired' and I'm not able to find any (original) source saying that  Yasso 800's primary function is to act as a predictor for your marathon time. In what I believe must be the original article, dated 28 September 2001, Bart Yasso is quoted as saying;  "I've been doing this particular workout for about 15 years, and it always seems to work for me. If I can get my 800s down to 2 minutes 50 seconds, I'm in 2:50 marathon shape. If I can get down to 2:40, I can run a 2:40 marathon. I'm shooting for a 2:37 marathon right now, so I'm running my 800s in 2:37."  The article then goes on to describe the Yasso 800's as a work out method aimed at getting you up to speed for your target marathon pace.

I want to comment on the discussion about correlation as this is a difficult concept with some common misconceptions. First of all, the lemons/road accidents example is not relevant. It's a "fun fact" example of correlation without any causality - a coincident. You will easily be able to find such coincidental correlations between data sets if you have access to a lot of data sets and carefully select the variables and limit the data intervals you look at. The Yasso 800s and marathon finish times obviously have a correlation with some sort of causality involved, and it's not coincidental. If the best approximation is a factor of 60 (converting minutes to hours and seconds to minutes) is another story. And I haven't found any scientific research about just how strong the correlation is, but I'm sure we can all agree that a certain correlation would be measured if such research was carried out.
From Wikipedia: The conventional dictum that "correlation does not imply causation" means that correlation cannot be used to infer a causal relationship between the variables. This dictum should not be taken to mean that correlations cannot indicate the potential existence of causal relations.

However, the causes underlying the correlation, if any, may be of several different types - six to be precise (Jaccard, J., & Turrisi, R. (2003). Interaction Effects in Multiple Regression). In the case of Yasso 800s and marathon times I suspect we experience what is called a spurious relationship: "A spurious relationship is one in which X and Y are related, but only because of a common cause, Z.  There is no formal causal link between X and Y."  What is the Z in this case? VO2max, running efficiency, max speed, endurance etc. - such parameters will all influence both the Yasso 800 and the marathon time. Even if VO2max is more important for the Yasso 800 than for the marathon time (I take your word for that!), it will probably have some effect also on your ability to run a marathon. Similarly, your stamina is important for your ability to finish 10x800 meters even if it is more important during the marathon.
So to summarize, I believe there is no formal causality between Yasso 800 and marathon times, but there is probably a spurious relationship based on the effect from common physical parameters as VO2max, running efficiency, stamina etc. How strong the correlation is will require some formal research, and hence the predictive power of the Yasso 800s is, for now, based on somewhat anecdotal evidence. To what extent the Yasso 800 is a good workout strategy in order to improve your marathon time is something I leave for the experts to discuss. If you were my coach Jeff, I would surly listen to you :-)  For what it's worth, I believe modern research shows that interval training (like the Yasso 800s) is a great place to start for untrained beginners. For more experienced long distance runners they could probably still be good for speed training and increasing running economy, but I follow your reasoning that there is probably better ways to achieve this for a marathon runner.

Ingen kommentarer:

Legg inn en kommentar