Is this the best way to predict your marathon race time?

[bencrozier]

I remember reading about this several years ago and thought it was very interesting.... and very cool.  Watch the video! 


http://www.runnersworld.com/video/1,8052,s6-6-0-2,00.html?bcpid=2884340001&bclid=22921443001&bctid=23037459001

What do you guys think?

[Sasha Pachev]

Yasso 800s are almost as good at predicting your marathon time as running 100 meters all out. There is a measure of truth to both methods of estimating marathon performance. If you cannot run 12.5 in 100 meters, or if you cannot do 8x800 in 2:05, do not expect a 2:05 marathon. That is where the truth ends. If you can, you are likely still a far cry from a 2:05 marathon. You need to have a number of other things in order to realize your speed potential over the marathon, and it usually takes a long time and a lot of work to get there. The following needs to be there already or get developed:

- Slow-twitch fibers
- Solid aerobic support system (muscle fiber adaptations, heart, lungs, blood composition, etc)
- Healthy liver
- Healthy muscle glycogen storage/utilization system
- Impact-resistant muscles,ligaments,tendons, and bones

All of the above is not necessary to complete a successful Yasso session. So Yasso's workout predicts your marathon potential assuming all of the above works right, which would be the case for a well-trained marathon runner. But so would a 100 meter sprint, a 5 K race, an all out 800, and just about anything in that department. I would say a 5 K would be the best predictor in that list.

A more accurate prediction of what actually is going to happen in a marathon is how you feel at the end of a fast paced 20 mile run.

[Gary Culver]

No

[Steve P]

Any predictive model (theory) will have simplifying assumptions built into it. Yasso 800s assume that a person has enough fitness to complete the marathon distance. And it doesn't consider other factors that may play an important role in determining the actual time. However, simple models (such as this one) often perform as well as or even better than more complicated ones...or at least they perform adequately for the purposes for which they are intended.

In the video, Yasso claims that his model has achieved considerable success for individual runners. However, I don't know how scientifically he assessed the accuracy.

To know which is the "best" model for predicting marathon race times in general, I would argue that you'd need to do an in-depth statistical analysis that considers confounding effects, sample size, dependencies, etc. The originator of a theory is naturally going to believe his/her theory is valid and may focus on facts that support this theory and bias the analysis in favor of that theory.

Sasha believes his model is better than Yasso's model, but I have to believe this claim is based on intuition and anectdotal evidence, because he hasn't cited a reliable source. There's nothing wrong with using your intuition or anectdotes to devise a model, but showing that one is better than another requires solid statistical evidence. Sorry, pet peeve of mine.

I am currently reading a book called The Lore of Running by Noakes, which is excellent because it discusses decades of research. He has an in-depth discussion on this topic in Chapter 2. Essentially, he claims your 10-K race time is a simple, effective predictor of marathon time (but of course it also has simplifying assumptions). If you care enough, you might check it out.

The bottom line is that it depends on how precise an answer you want. If you are interested in a pretty good approximation, Yasso's method will probably work well enough for you.

[Jon Allen]

Good points, Steve- evidence is required for a true scientific study, not intuition.  Lore of Running has much science, and I like Ptfizinger's Advanced Marathoning for the same reason.

My personal favorite for predicting marathon times is to run a 1/2 marathon about a month before, and use one of the many online calculators that will give you equivalent marathon time (about 2.1-2.2 times your half marathon time).

[Sasha Pachev]

The ultimate predictor of your marathon time is go out and run the marathon. The closer you get to that, the more accurate that prediction will be. So 8x800 is better than 100 meter sprint, continuous 5 K is better than 8x800, 10 K is better than 5 K, half is better than 10 K, and 20 miles is better than half.

Regarding feel vs science. If you want to be a good runner or a good running coach you must master the art of feel. You must develop it to the point where you can trust it over science because no science can replace it. If you have that feel, you will do well even if you know little about science. If you do not, you will be in an endless cycle of pouring over latest research data, and having no clue what is relevant to your training and how to apply it effectively.

A good analogy would be driving. You do not pull out a calculator and to a statistical analysis to determine a safe driving speed in given conditions. You just pay attention to how far you can see, how straight the road is, if it is raining or snowing,  how well you are able to control the car, etc. Every single day most of us make driving decisions that could cost one his life if he is wrong, and most of those decisions are purely intuitive, not a whole lot of science behind them. In spite of the lack of science, we manage to do a pretty good job of coming home alive, in a reasonable amount of time, and with a non-crashed car on a daily basis.

In running it is very similar. Properly developed feel is your most trusted friend. Think of Frank Shorter running a 1200 meter repeat on target with the 0.1 second precision without getting any mid-interval splits, that is what I mean by mastering the art of feel. The best athletes are not produced in a lab. They often get access to a lab because they are the best and the scientists want to study them, but it is more of the scientists documenting what the athletes have figured out already rather than science leading the way. A good coach will look at a few numbers, ask the athlete how he feels, and he will know what to do. Very unscientific, but it is yet to be beat with a scientific approach.

[Dallen]

It's obviously got a lot of built in assumptions, but I think that the Yasso 800's are a fairly accurate predictor. Fairly accurate meaning probably within 5-10 minutes. I think it has some value to people who have never run a marathon and just want to know a good starting time to shoot for.

I prefer half marathon time X 2 + 10 minutes.

[Jon Allen]

Sasha- agreed the best predictor of marathon time is to run a marathon

Regarding science versus feel- I think you're talking about a slightly different thing than Steve and I are:  If an engineer wanted to design a fast car, he would run analysis and extensive testing on different designs, engines, etc to come up with the optimal design.  Now, to drive it the fastest, you would need an experienced driver who could drive by "feel", like knowing when to accelerate and slow down, how to turn, etc.  Similarly, to get a faster athlete, you can do analysis and testing on a large population to come to certain conclusions, such as doing XX will result in XX (for example, 400 m repeats can improve your 1600 m time).  Now, in applying it to individuals, and especially in how that individual runs on a given day or race, there is some degree of "feel".  But, the background science going into the training and nutrition can make a big difference.

but it is yet to be beat with a scientific approach

That's a pretty broad, brash statement.  You don't think the world class runners have lots of science behind their training?  I think science is a big reason WR's keep improving in all sports.  I don't think today's swimmers and runners are better than the ones 50 years ago.  I think they have better science (including training, nutrition, etc) that help them get faster and faster.  Ryan Hall with some science to optimize his running will be Ryan Hall without science.

[Sasha Pachev]

Runners today are faster than 30 years ago because Kenyans and Ethiopians are running in competition rather than just from village to school , not because of science. Ryan Hall does not live in a lab as far as I know, he just trains. He is beating the guys that live in a lab.

[dave rockness]

Yasso 800's were 15-20 minutes off for my Fall marathon (2:50-55 Yasso's w/3:10 marathon).  However, I was pretty close to the half marathon x 2 + 10minutes (1:31 half marathon)

[Jon Allen]

Runner's World interview with Terrence Mahon, Ryan's coach:
"I will create their training programs off of a scientific model that is event specific with a goal of being as economical as possible at their respective race speeds."

You don't have to live in a lab to benefit from science.  That's all I'm saying.

[Sasha Pachev]

I believe in attempting to benefit from science. I read scientific papers and other info quite frequently, ponder it, and try to figure out if it is relevant and how it applies. 

However, I've seen little evidence that science has offered a whole lot of benefit to runners. Interval training was invented by a runner (Emil Zatopek) not a scientist. The idea of base mileage for aerobic conditioning was invented by a runner and a coach (Arthur Lydiard), again not a scientist.  Granted they had a scientific mind, and did a lot of trial and error of their own, which you could call science. But that is the kind of science that it takes to figure out how to run faster, not your regression analysis, randomized study with controls, etc, in short doing a whole lot of work to prove the obvious or end up inconclusive because you lack that feel that I was talking about earlier.

On the other hand, I see a lot of runners that have become slaves to science. Or shall we say pseudo-science. Science is not be deemed true science just because it follows a recognized protocol. The ultimate test of a true science if it yields fruit. So if a science follows the protocol, but consistently fails to tell you what you need to do to reach your goal beyond the obvious, that is not science to me, that is a waste of resources masquerading as science. And that is, unfortunately, what many runners become a slave to.

[Steve P]

There should be a balance between science and intuition. Just as there should be a balance between the mind and heart in spiritual things.

An analogy is medical care. There's not a set code of correct/incorrect answers for treating all patients. The doctor must use her intuition to treat individual cases. But there are many guiding principles that are broadly applicable and that have been borne out by decades or centuries of research. I would rather have a doctor treat me who knows and generally follows such principles (but also relies on intuition to an extent) than a doctor who merely does what her intuition says and ignores the science. (I better stop or I'll go into a rant about people not vaccinating their children...

On the same note, I'd rather base my running theories on the best available research at the time and then use my own judgment and experience to tailor that to my specific preferences.

I will refer back to the original question, which was whether Yasso 800s are the "best" way to predict a person's marathon time. The only practical reason that I can think to do so would be to determine your target pace when you're not sure what it should be. The main point I was trying to make is that stating as fact that one approach is better than another without having evidence to back it up is not very credible. Sure, such evidence wouldn't prove the ultimate truth, but it has more credibility for me than an idea that has only been evaluated in someone's mind. That, of course, doesn't prevent us from expression our opinions.

Following a certain protocol doesn't turn a study into "science." But it can add credibility. There are many biases that can creep in, and such protocols help to avoid those. Otherwise, you could get people who do a study on SPAM consumption in runners and "prove" that higher consumption predicts better marathon performance because they only fed the SPAM to the runners who had performed faster previously and thus they have an advantage over the others. Silly example. I'm just saying that subtle biases can creep in.

Probably a lot more than Ben was asking for, but still an interesting discussion.

[bencrozier]

Everything that we know about running as fact, is part of the science of running.  Science (from the Latin scientia, meaning "knowledge") refers to any systematic knowledge-base or prescriptive practice that is capable of resulting in a prediction or predictable type of outcome. In this sense, science may refer to a highly skilled technique or practice.

In its more restricted contemporary sense, science refers to a system of acquiring knowledge based on scientific method, and to the organized body of knowledge gained through such research.

Science is a continuing effort to discover and increase human knowledge and understanding through disciplined research. Using controlled methods, scientists collect observable evidence of natural or social phenomena, record measurable data relating to the observations, and analyze this information to construct theoretical explanations of how things work. The methods of scientific research include the generation of hypotheses about how phenomena work, and experimentation that tests these hypotheses under controlled conditions. Scientists are also expected to publish their information so other scientists can do similar experiments to double-check their conclusions. The results of this process enable better understanding of past events, and better ability to predict future events of the same kind as those that have been tested.

Personally, I would prefer to know as much as I possibly can about running, aka, the science of running so that I train as efficiently as possible.  Intuition should play a role in developing a hypothesis that I must prove or disprove using the scientific method.

[Steve P]

Here's a study by Noakes (the author of Lore of Running) and Myburgh on predicting marathon time:

http://www.ncbi.nlm.nih.gov/pubmed/2359150

I don't have access to the full article, but in his book, Noakes summarizes as follows:

Performance in shorter-distance races, particularly at 10 or 21 km [half marathon], remains the best predictor. In our studies, we found that a marathon time could be predicted from either 10-km or 21-km time, using the following equations:

• time for 42 km (min) = 5.48 x 10-km time (min) - 28.00 (min)
• time for 42 km (min) = 2.11 x 21-km time (min)

The equation using the 21-km time was found to be the more accurate predictor.

That finding is what many would have expected, I'm sure, but it's an illustration of using scientific methods to evaluate a hypothesis that seemed intuitive to the authors. Part of their reason for doing the study was to see whether this method (which they believe takes into account VO2-max and "running economy") was better than other methods that only use VO2-max to predict performance.