Thursday, 25 April 2013

A Tiny Bit More on Pacing, and Trail Running Sussex Develops


Tonight's post will be just a short post.  Firstly, a little bit more on race pacing, and then an announcement of an exciting development for Trail Running Sussex.

Last Sunday morning I had a great time watching the BBC TV coverage of the London Marathon, whilst at the same time, watching on the London Marathon website the live progress of five runners I were tracking.  The London Marathon website was excellent.  It would automatically update every 5 kilometres and at halfway, with the runners last 5km (1.1km) running pace, and a predicted finish time.  It would also predict (based on the runner's pace to date) where they were on the course every ten seconds or so.  So a great feature that added to the enjoyment of watching the elite races.

The women's elite race had a pretty small field, and straight away it became clear that the elite women runners weren't interested in the pace that the two pacemakers were setting.  So, in terms of elite women standards, a reasonably large bunch 'dawdled to the 20 km mark, passed through halfway in a time of 71:49, and then they started racing.  The 5km split time decreased from 17:11 for 15 - 20km, to 16:01 for the 20 - 25km split.  The race was won, with a negative split of 3:23, running 68:26, for a finish time of 2:20:15.  The second placed women, also achieved a negative split of 2:06, but all of the other 15 elite women finishers positive split the race, with the largest positive split of 13:17 by the current Olympic champion, although she did get knocked down by a wheelchair racer!

The elite men's race was quite different with the men's field keeping in contact with the pacemakers who were running at sub world record pace.  They actually ran the first 5km in a time of 14:23, which is quite amazing being a marathon pace resulting in the time of 2:01:23!  No wonder that was the fastest 5km split of the race, and pretty well every subsequent 5km split got progressively slower.  Now this was a demonstration of taking my motto of "Run as fast as you can, while you can!" to the extreme!!!  So with the men's elite field going unbelievably quick, there was actually still a lead bunch of 7 runners at halfway, reached in 61:34, still under world record pace, with the reigning Olympic champion just 14 seconds back. 

All of these eight runners filled the first eight finish places, all slowing down significantly so none of them, in fact, none of the entire men's elite field achieved a negative split.  The winner ran a positive split of 2:52, and the largest positive split was 5:31 (8th place) and the smallest positive split was 0:19 (9th place).

Around two years ago I analysed the London Marathon 2011 results to help understand the 'stupidity' of trying to run a negative split in a marathon.  (Check out the Pacing article.) So I didn't feel that there was a need to repeat this analysis, so I thought I would look at some results data in a different way, and this time just looking at the top end of the field.

I therefore downloaded the top 500 finish times and their half marathon split times from the massed start field.  With the belief that running a negative split is detrimental to ones performance, I thought it would then be simple to produce a scatter diagram plotting the amount of the positive/negative split against finishing place, to illustrate this performance inhibiting relationship, with those who run a negative split finishing in lower places. 

First I ranked the 500 runners in terms of the positive split, with a rank of 1 is the runner that slows down the most, and the rank of 500 is the runner that speeds up the most during the second half of the race.  Well take a look at this scatter diagram below, with a coefficient of determination of 0.00001.  This data is pretty well as close as as you could ever get to illustrating that there is absolutely no relationship at all between the amount of the positive/negative split and finishing pace.

I then instead of looking at the relationship between the positive/negative ranking and finish place, I looked at the relationship between the actual positive/negative time and the finish time.  Again there appears to be pretty well no relationship! Why?

I then had to really think, why no relationship?  Surely those that run a negative split, have only achieved this because they have run the first half of the race so slowly.  It then jumped out at me.  Yes, those that negative split perform poorly, but they just simply move lower down the finishing place list.  Throughout the field people will finish with a certain time due to many factors, so there is a massive spread of finish times.  Within this spread of finish times, some runners will perform well in relation to their fitness, ability, belief, whatever else, and hence move further up the field, finish faster.  Whereas some runners will perform poorly, i.e. finish slower.  But the good or poor performances will not stand out within the results, they simply move upwards or downwards amongst the finish results.  Hence the zero relationship.

However, if we look at the very top end of the field, those who run a poor performance, which I believe is due to running a negative split, will move down the results list, and those who run a good performance, those who I believe who run a positive split, i.e. start out fast and then slow during the second half of the race, will move up the result list.  And being at the top end of the field, simply moving up or down into the next portion of the results list can't occur, so this true relationship wont then be lost!

So what do the results fro the London Marathon 2013 show us.  As already described, all 19 elite men finishers ran a positive split, and 15 out of the 17 elite women finishers also ran a positive split.  But what the elite do may not translate to the non elite.  So looking at the top level club runners from the London Marathon, as expected, those that run a good performance move up the result list, and those that run a poor performance move down the results list.  For the massed start field, out of the first 118 finishers, from a finish time of 2:17:10 through to 2:37:30, so the best non-elite performers, ONLY ONE runner achieved a negative split.  All of the other 117 runners ran positive splits.  I don't think I need to discuss this topic again.  The idea that running a negative split is a sign of a good running performance has now been totally 'blown out of the water'.  End of story!

Well that first subject of tonight's post wasn't quite as brief as I thought!

My second topic tonight is related to the running venture I am involved with, Trail Running Sussex.  You may recall that I mentioned last month that Trail Running Sussex has recently been established and provides guided runs and trail running camps within Sussex.  The first running camp has an ultra trail running focus, and takes place from Friday 21st June through to Sunday 23rd June, based at the National Trust Slindon Estate, near Arundel. Well, Trail Running Sussex has just set up a relationship with the newly established online running store The Ultra Runner Shop.  So, in association with The Ultra Runner Shop, all trail running camp participants, and those that attend a guided trail running weekend break, will receive a 15% discount on all running kit purchased from The Ultra Runner Store.

I am really excited with this development, as The Ultra Runner Store, which has been set up by ultra runner Tony Holland and his partner Helen, not only stocks top quality kit, but they are also raising money for and awareness of the Downs Syndrome Association.  So if you have a spare moment, check out their website at the following address and if you need any running kit, see what they have and give them a go.

Time to sign off.  I shouldn't really be having any late nights at the moment, with my first key race of the year taking place this Saturday, being the 53 mile Highland Fling.  To those of you racing the Fling, I wish you all the best.  I'll see you on the start line at 6:00am, and whatever you do, don't try to attempt to run a negative split!

I will sign off with some appropriate words that were within a comment left on my blog by Tom back in September 2011 following my UTMB DNF:
"Training is just that, (the ability to tune in), and it is during this period you need to build your confidence and physical ability. This is the time to build mental strength and grow and in that respect I agree 100% with your model.  Once this is achieved you have the physical and mental tools to perform on race day but you also need to have these set to auto so that you can trust without thought. I see it like the auto neuro response that makes your heart beat - you don't give it thought you just trust it will happen.  The secret, in my view, is on race day to be able to forget everything (trust the heart will beat) but tuning into the environment that nature throws your way and flow with it and thus enjoy it (vital).

I believe we can all 'mentally horde' without realising it and maybe all this preparation is taking away from what matters. Maybe we can forget to trust and not separate training from the joy of racing.

Training = build the physical off the mental.
Race Day = Trust and enjoy."
See you on the trails,



  1. Your keenness on positive splits being a good thing seems to make you blind to some of the facts.

    The men's race went out way too fast and they suffered badly in the second half. The winners time was slower than it should have been had the pace in the first part of the race had been so suicidal.

    If you look at the world records for Marathon, these have been down with very even splits or negative splits. So the very best of best performances seems to come with even or negative splits.

    What your analysis does show is just how hard it is to achieve an even pace, how hard it is not go out too fast and suffer in the later stages. If you are racing it's close to impossible the further you go as the limits of your late racec performance are so difficult to detect in the early parts of race.

    The fact it's so hard to do an even pace or negative splits doesn't make a positive split a good approach to aim for, nor does a positive split show that individuals performance on the day was maximized. The only fact that you have show is that positive splits are normal and one shouldn't beat yourself up for doing one.

    There isn't any physiological advantage to a positive split as far as I'm aware. I don't recall you ever putting one forward.

    One reason to go out slow is that your body isn't yet warmed up and the enzymes in your muscles don't work well aerobically until they are warmed up so you burn carbs faster at the start for a given pace. Going out fast squanders this precious resource.

    Second reason, an even speed and associated even split minimizes the amount of carbs burned for given speed. As depleting your reserves of glycogen is a factor in crucial factor in marathon and longer race performance optimizing your use of carbs is important. An positive split simply doesn't make the best of available fuel you have.

    Going out fast also accumulates muscle damage more quickly than going out at steady even pace. Muscle damage to speed isn't a linear relationship, going 10% faster incurs muscle damage at a faster rate than 10%, this means you get to the half way point with more muscle damage if you run a positive split, and slowing down in the second half won't magically fix this - you have to carry this source of fatigue all the way too the end.

    These aren't reasons that can be easily dismissed, I've loved to see you try to come up with some compelling reason why it's a good thing to go out hard and how this overcomes the well established reasons for an even split I mentioned above.

  2. Hi Robert

    Thanks for your comment above, and your previous comments on my post two weeks agao where I also discussed pacing. Many apologies for not directly replying to your previous comments, however, I was hoping my post last night would provide some more evidence that running a positive split is the best strategy.

    From your above comments, which I really do welcome as it does get me to consider and question my beliefs, I can see that you are still not convinced. Rather than respond directly here, I think I will do one more pacing blog post next week, where I will be able to provide for you the "compelling reason why it's a good thing to go out hard and how this overcomes the well established reasons for an even split". I will also be able to add some comments regarding pacing during the Highland Fling, as I will also have the 2013 data, so able to reply to your earlier comments.

    Thanks for your input to my UltraStu blog. Are you doing any trail races this year, if so maybe we can discuss the pacing strategy in person if we are both at the same race. Alternatively, why not come down to my Ultra Trail Running Camp at the end off June, and we can have the entire weekend discussing and debating trail running.


    1. This comment has been removed by the author.

  3. (reposting this reply to fix some typos :-)

    I remain totally unconvinced because as the actual London marathon data doesn't support your hypothesis that positive splits are a good thing.

    Might I suggest you have a look at the splits of men's and women's world records, if positive splits was a good thing then you'd see significant positive splits. What you do do you see are negative or even splits for women's and men's respectively. If you go back to previous men's record then it was a negative split as well.

    If you also look at the splits of the men's London marathon races. This years men's race was provided the slowest finishing time in 7 years, yet had the elite group go through below the world record pace at the half way point. It had the fastest half way split yet the slowest time? The previous two years where the way splits were slower had three runners below this years winning time. This year's winner also has been quoted that the elite's went out too fast. This evidence certainly doesn't support the idea that the positive split is a good thing.

    Far from a positive split being good for this years elite men's marathoners it destroyed their performances, runners who have run several minutes faster on numerous occasions fell apart in the later stages. They ruined their race times when just doing a positive split of a few percent so it looks to me like even a modest amount of positive split is performance sapping.

    Just because most runners typically don't negative split doesn't mean that a positive split is a good thing, it just proves how hard running the perfect race is, so many things can go wrong in the later stages, almost all exacerbated by going out too fast.

    Your own PB at the Fling was done when holding yourself back more than is usual, but still not enough as you lost places in the second half. Even with your slowing you still had splits that were better than average w.r.t slowing. So the majority of runners slowed more than you, so might suggest that for them they went out relatively even more aggressively than you?

    My recommendation would be to yourself would be to reign things in for the first half, sure it'll take a bit of patience and confidence in your ability to have a strong second half, but it'll reduce your use of carbs, reduce the strain on hydration and eating and create far less muscle damage than going out aggressively. Remember the previous times you've run the Fling you got beat last time by several runners that exhibited just this tactic I'm advocating.

  4. This comment has been removed by the author.

    1. Oopps, removed and replaced with typo's fixed again:

      Stuart, I went ahead and did a comparison of splits and finishing times of London Marathon Winners, their PB's and the marathon world records:

      I believe this shows that anything more than a very small Positive Split is detrimental to marathon performance. It's also clear from this your analysis that most runners including elites Positive Split does not sustain the conclusion that a Positive Split is good thing. Your analysis simply shows how hard it is to do an Negative or Even split.

    2. Hi Robert

      I went to your blog but couldn't find your analysis. there doesn't appear to be a link.