Last Sunday on the weekly LSR, I got to talk with a follow runner that was about to run his first Marathon in 2 months time. The subject was how to find the correct Marathon pace, and it turned out that most of us on the group had wildly different rules as to how to find the correct or optimum pace.

We came up with the following list of methods and I wonder if any of you have anything specific to say about them. (Yes, I do know that is not a proper Q&A, but I have no ideas where else to ask the question...) Assume that you have run a half-marathon at 1:45:00 (=4:57 min/km)...

  • We had "double the distance, means add 13 sec/km or 20 sec/mile" (no official reference, but see Runners World), which gives 5:10 min/km (=3:38:00).
  • We had "2*HM+20 minutes", which gives 3:50:00 (=5:27 min/km).
  • Run for 60% HRR, which for me would give 5:25 min/km (=3:48:00)
  • We had "age-grade is constant", which would give 3:39:00 (=5:12 min/km) (see this calculator for numbers). The idea here is that if you run at the x% of the speed of the fastest HM runner at your age, then you can do the same for the fastest Marathon runner also your age.
  • We had Pete Riegels formulars, which gives 3:38:00 (=5:10 min/km). I have used these via SportTracks for some years, and testify that they are somewhat optimistic unless you really are doing an all-out effort.
  • We had "VO2max is constant" method based on work by Daniels and Gilbert, which gives 3:38:00 (=5:10 min/km).

Although the methods are rather different (at least as I seem them), they tend to give almost the same results: either 3:38:00 (=5:10 min/km) or 3:50:00 (=5:27 min/km). I have no doubts that the later speed will give a more comfortable and sure run with little risk of injury whereas the former will probably give some sore legs for the next week.

How would you rate these methods? Should we tell our fellow runner to go for the fast pace (he is young and confident of himself) or the slower pace?

4 Answers 4


Interesting conversation and results.

I like to use what is known as the Yasso 800 method named after Bart Yasso.

Instead of using your half marathon time, base it off of repeated 800 meter efforts.

To test, run a series of 800 meter controlled efforts with ~400 meter jog in between. The average time in minutes and seconds of your 800s will be your marathon time in hours and minutes. For instance, if you can run 6-10 800 meter repeats at an average of 3 minutes 20 seconds, your marathon time will be around 3 hours 20 minutes. From this you can estimate the pace and effort that you should target for the marathon.

In my experience, it's best to build during a marathon at least in effort if not pace. So again, using the 3:20 800s which equates to about a 7:38 pace, you may want to start out closer to 8:00 miles and build to a 7:30 pace or effort. Your results may vary.

  • Whether you should have a constant pace or increase your pace underway, seems to be a longstanding debate with almost religious undertones :-) I prefer to keep my pace throughout the race where possible. And speed up the 2-3 kms if I have the energy for it. Apr 9, 2013 at 17:16
  • no doubt. i think both approaches are reasonable depending on the athlete. and there may be other approaches that work too, for instance running 2 miles, walking 1 mile, etc. Apr 9, 2013 at 17:49
  • Love Yasso 800s!
    – user5324
    Apr 9, 2013 at 20:44
  • I find it amusing that I am a total outlier on Yasso 800's. I have run fairly consistent 800's in training, over 8 years or so, around 3:30's. But my best Marathon time ever is 4:07. But then I accept I am an outlier on many things. (Big and heavy).
    – geoffc
    Apr 10, 2013 at 17:04
  • 1
    @geoffc i'm sure there is a reason for that, finding it may be difficult (fitness? muscular endurance? form? etc). i've been there too and was able to improve my endurance by working with a great coach - and am still improving. best of luck to you. Apr 11, 2013 at 18:02

The answer largely depends on his training. How has he felt after his long runs and longer tempo runs and at what pace were they? I don't like to compare half times unless the half was run on some part of the same course as the marathon you're trying to estimate.

[Warning: non-scientific advice ahead]

All that being said, for a first marathon I'd suggest running the first 16 miles consistently at the easier of the two paces to gauge how he's holding up. If he feels strong at 16 pick up the pace and evaluate again at 20, then again at 23. In my marathons I've found that both physically and mentally it's easier for me to increase pace if I have targeted distance intervals. When I've gone out too fast, I've failed to maintain.

I use 16 because that's a distance I'm very comfortable running and I can accurately tell how I feel at that point. Plus, there are only 10.2 miles to go :)

I check again at 20 (the 10k left mark) and again at 23 (roughly a 5k) because those are easy distances to mentally "swallow" and I know how my body responds over them. They are also a nice way to mentally divide the remainder of the race.

  • I know many run with negative splits, but I have also seen some very strong advice against this strategy (missing a reference right now). As far as I can see the common advice for non-competitive runner is to use even splits. Apr 10, 2013 at 7:55
  • I can see where even splits might work best for someone who just wants to get to the finish in one piece but for the example given (based on the goal times) I made the assumption that this runner was interested in doing a little better than just finishing.
    – user5324
    Apr 10, 2013 at 12:04

I did my first marathon 5 days ago. I wrestled with the question of target pace. The tool at McMillan predicts a 3:22:25 for me based on my half marathon. That feels very aggressive to me.

I ran 3:38:11 on the day and I'm happy with that. I think if I'd started much faster I would have been reduced to walking/staggering at the end and would have ended up overall slower.

There's a new tool at the Fetch Everyone website that gives a much closer prediction for me: 3:36:58 - see http://www.fetcheveryone.com/training-calculators-improvedriegel.php

There's an article at http://www.fetcheveryone.com/eblasts/article37/ explaining the maths and research behind it. In short, they analysed the Fetch Everyone database of times for runners who've done 5 halves and marathons to see what happens in the "real world".

Try the tool - for first timers I'm confident it will give a better estimate than the original Riegels formula.

For reference, my best times for shorter distances are: 5k=20:10, 10k=42:50, 13.1m=96:11 I've run 30-33k 5 times in the last year. I ran 1850 miles in 2012.

  • Can you please explain how they calculate in that tool? It would be appreciated.
    – Freakyuser
    May 3, 2013 at 9:34
  • Interesting. The prediction is the 4:01:00 (=5:51 min/km) for our test person above, which seems very slow to me, but... May 3, 2013 at 12:06
  • Freakyuser, I don't know any more details than in the article37 that I linked to. I assume he kept the "shape" of the Riegel's formula and used some form of "best fit" to vary the exponent to find the best one for the runners in his database. I also found someone else who was inspired by his analysis to do some of his own. See runningahead.com/forums/post/984253efa2a040318cd931cd38f9d581 There is also a discussion thread you could join... May 6, 2013 at 18:48

For me, effort is the best "pace" with easy quick cadence upto 12-15 miles, medium effort upto ~22 miles, and then hard effort (with ape-like curled lips).

In general it is definitely dependent on your fitness, attitude, and the race atmosphere. This answer is opinion but McMillian does state in YOU (only faster) that their predictions depend on proper fitness. I think this may be true in general for the other methods

If you are well trained or a veteran marathoner then all those methods mentioned will work for you. The main difference from my point of view is that you chose one that will impact your training the least, and this I think is one reason for Yasso 800 being so popular (plus it easily to push hard for a shorter amount of time).

I dont like any of these predictors much because I'm not well trained for a marathon distance, and every attempt to meet those times is a nightmare for a marathon compared to a half marathon when I can meet the McMillian calculator prediction.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.