What time can I run for my marathon?

What time can I run for my marathon?

This is the million-dollar question isn’t it? Whether you’re experienced or brand new to the marathon, it will likely play on your mind unless you are one of the lucky minority who sets this question aside and simply says ‘I’ll run to feel this time and see where I end up’. The latter can be a very good way to run a marathon but if you want a clearer idea, then read on.


My own experience

When I ran my first marathon in Dublin, the same question pierced my mind. Back then – like most ‘younger’ men – I craved certainty. Although my older self is more comfortable with uncertainty and ambiguity, I still understand and appreciate the need to have a general idea – a plan A and a plan B (the fallback).

Previous readers here will remember I was a lucky man for my first marathon because I had a physiologist-coach on the payroll (ok, so I wasn’t lucky – I had made good choices). Two weeks before the marathon she told me ‘you were aiming for 3:30 but based on these blood-lactate readings you’ll be closer to 3:20’. I ran 3:18:48 and the rest is history – clearly in this case science offered a very accurate prediction of the possible.

But what factors affect whether such a prediction will hold water? And if you do not have access, or budget or desire, to hire an exercise-physiologist are there other ways to get a proper marathon time goal? And why would you want it in the first place?


Why’d you want to know!

Let’s start at the last question – although a very experienced athlete can extract optimal timing almost without peering at their watch, this often is not workable for beginners (nor is it precise enough, it seems, for those trying to break the world record!) or people used to relying on their watch and ‘out of touch’ with their internal feedback systems.

What happens in these cases is normally the runner starting too hard and paying the price later in the form of a rapid slow-down (sometimes called ‘The Wall’ in its most dramatic form). Other times, the runner lacks confidence and starts so slow that they do not run to their full potential even when they dramatically speed up at the beginning. Statistics are simple: the most efficient marathon strategy is the most even (because uneven pace is costly) and to utilize this we need to have a good idea of our own capabilities.

So how can we gauge our level?


Recent race performances

Race performance on one distance has a strong predictive value for another distance. The more recent the race and the more similar in distance to the marathon, the stronger the predictive value. Simply enter your race result into any of many ‘VDOT’ or ‘VO2max’ calculators or Apps and you will get equivalent performances. It’s a good idea to ‘shop around’ and look at a few to get a range of targets as some calculators are noted to be ‘more optimistic’ than others. Once you have a few options in the same ball-park use your gut-feeling to hone in on the primary target.

When you work with us, we can help do this assessment based on our internal calculators, so you do not need to do it.

The drawbacks of this method: some calculators are overly optimistic, and they do not consider whether you did appropriate training for the marathon. If you have a strong 5 km time of 17:30 for instance you should be able to run low 2:50ies for the marathon. But if you did not do very marathon-specific training then this ‘speed’ may not transfer onto the longer distance. For this reason, a longer race (like the half-marathon or 10 mile) are better options to do in a marathon build-up although many coaches now believe they are detrimental because – energetically – they are more different from the marathon than you might expect.


Pick the right race to predict from!

Many runners who can comfortably run well under 90 minutes for the half-marathon cannot break 3 hours for the marathon because their training is not balanced correctly towards endurance. You can get away with a mistake like that much more easily on half the distance. If you believe your training plan has been heavily endurance based (such as our own 80/20 weighted plans) you can be more confident.

The second drawback is that you do need a recent race result – ideally within 6-12 weeks for it to have a good predictive value. A two-year-old result is as good as worthless. Ask yourself the question ‘how’s my fitness compared to when I ran that race?’. Since we often ‘peak’ for marathons a result in an earlier race may also be misleading because you did not train for it and did not peak for it (so a 5 km done during marathon training may give you TOO LOW an estimate).

Example: In 2012, my half-marathon race performance of 1:20:28 was the most recent relevant time and predicted a marathon time of just over 2:48. I ran 2:55:58 – something that can be accounted for due to the tough course and 25 degree Celsius conditions (we’ll see more about these ‘x factors’ in part 2 of this article.


Blood lactate levels

A lactate test conducted by an exercise physiologist will predict your marathon time under normal conditions with a high degree of accuracy. This is one route you can go down – but because such a test is hard work it should be done just before the standard two-week taper begins. Generally the speed where you maintain 2 mmol/l/kg of bodyweight blood lactate levels (called the Aerobic Threshold or AeT) will be what you can maintain for the marathon.

The downside: you’ll have to budget 100 to 150 euros and take a few hours out of your day to do it.

Example: This test predicted 3:20 for me in 2007 when I ran 3:18:48 and predicted 2:48:48 for me in 2012 when I ran 2:55:58 (but would have been predicted to run 2:48:10 going by my half-marathon time if not for adverse conditions). In both cases the test was very close to predicting exactly my time. My tests were performed first by the now defunct PeakHealth and later by UCD researcher Romain Denis.


Lactate threshold

Using the test above or the slightly less precise lactate estimates from gadgets such as the Garmin ForeRunner 630 (and later models) you can calculate your marathon pace based on the Lactate Threshold speed estimated by your watch. My ForeRunner predicted my Lactate Threshold to occur at 174 beats per minute heart rate and a speed of 4:47 min/km in the summer of 2016.

Marathon pace is estimated to be 95% of this speed. 4:47 min/km is 12.54 kph (just divide 1:00:0 in an Excel sheet with your pace or estimate it remembering that 5 min/km is 12 kph and 4 min/km is 15 kph etc.). 95% of 12.54 is 11.91 or 5:02 min/km pace (you get this by dividing 1:00:00 by 11.91) or a 3:32:29 marathon.

You really need to health-check these metrics. In my case I ran 1:06 for the 10 mile earlier this year which is essentially a bit slower than LT and this time equates to a much faster marathon time of 3:03:57 (Lactate Threshold occurs at your best 1 hour speed).

(as an anecdote this ties into our weight article nicely if I wanted to compare my fitness level over the years as I have aged, detrained etc. When I was fit enough to run 2:48ish I weighed 67 kilos. Now I weight 72 kilos. By losing those 4 kilos I am estimated to gain 24 minutes and to run about 3:08 using this calculator whereas using my own VO2max conversion I am estimated to run 3:26). Whether we believe 3:08 or 3:26 this would be the truer reflection of true fitness lose over time than if we do not account for body-weight).


Training pace

Your training paces offer another important clue to how well you can do – especially the pace of your long runs, medium long runs and steady runs (such as Out & Backs and other sub-threshold and threshold runs).

There is a correlation between marathon pace and easy run pace as long as you are certain you executed your long runs below what is called the ‘Aerobic Threshold’ (AeT) or ‘Max Aerobic Function’ (MAF). Very simply, if you run your long and medium runs at an intensity where you can easily converse and breathe through your nose if you had to, then you can probably trust the pace as a predictive tool.

If you train on extremely hard courses (i.e. ‘live in the Wicklow Mountains’!) or in very adverse weather conditions (say the hot weather of Dubai) then your training paces will probably underestimate what you can do.

But remember the trap: if you are constantly ‘straining’ (rather than training) then your paces will look more impressive for long runs, but they will not have a very good predictive value because you are not running these runs at ‘true easy’.


Example: When I ran 2:55:58 in Copenhagen 2012 my training pace was generally 4:50 to 4:40 min/km for all long and medium runs and that was on courses that were tougher than most marathon road races. The expected ‘easy’ pace for someone running 2:55:58 is 4:44 to 5:00 min/km. If you remember from the previous section my ‘real’ predicted time (if not for the adverse conditions) was 2:48. A 2:48 marathoner will do his long and easy runs in 4:30 to 4:50 min/km generally – so in this case there was another strong correlation.

If your easy runs are very erratic you can also use paces from sub-threshold runs such as Out & Back. We prefer our 1 hour Out and Back format – taken from Lydiard – because it happens at sub-threshold pace when done correctly. In my case I ran most of my workouts in 4:20 eventually down to just under 4 minutes/km at this intensity. The expected for my predictions: 4:16-3:56 min/km pace. Again a very solid predictor using a ‘moderate 1 hour run’ as your gauge.


In part 2: you set a good target, but you didn’t hit it!

Recommended reads

Lore of Running (Dr Timothy Noakes)

Healthy Intelligent Training (Dr Keith Livingstone)

IAF – Marathon Training – a scientific approach (Arcelli/Canova) (only orderable directly from the IAAF)

The following two tabs change content below.


Director and coach at Borg Coaching Services
Rene Borg is the head coach of Glendalough AC and a passionate runner competing over all distances and terrains.

Latest posts by René (see all)

3 Responses to What time can I run for my marathon?

  1. The pragmatic physiologist’s point of view (and 2:47 marathon runner at age 47): http://christofschwiening.blogspot.co.uk/2016/01/predicting-marathon-performance-from.html (and his subsequent post regarding Tanda’s 2011 study). This has worked well for me (although I add cycling and swimming into the mix, which throws the prediction off somewhat). Over the last few years Christof has added many other runners’ data to the original Tanda n=22.

    • Thanks Andy, I’ll try to enter those equations into an Excel sheet next to the ‘rule of thumb’ predictions and see how they compare.

      There are a number of other methods, of cousre, such as using a Conconi test and calculating percentage of deflection heart rate from that (that’s one of the recommendations of Canova/Arcelli in IAF’s ‘Marathon training – a scientific approach’) but that’s a bit more technical for the layman.

Leave a reply