Wednesday, 18 June 2014

West Highland Way Splits

I am now in the final phase of preparation for the 95 mile West Highland Way Race (WHWR) that commences at 1am on Saturday 21st June, part of this preparation has been looking at what splits. In this post I'll provide the spreadsheet so that others can tailor the splits to their own target times, some will be suitable for printing out and taking with a runner and crew, and others will be suitable for guessing finishing time based on how a runner is doing at each check point.

Those not at the race can follow progress online via the sportsident website which will report runners passing through check points in real-time, in combination with my spreadsheet you'll be able to

My own pacing plan : Pacing by Heart Rate

My own pacing plan will be to use my Heart Rate Monitor as a guide, aiming for a zone between 80 and 82% of my Lactate Threshold, which equates to a range of 140 to 145 beats per minute.  On the way to Drymen I'll attempt to keep my heart rate down to a 135 to 140 bpm range, and for the rest the race up to 140 to 145.  I'll also set a heart rate never to exceed of 150.  This range is roughly 12 bpm lower than when I raced the Highland Fling back in April.

Pacing by heart rate has worked well for me over the last six months, and especially well at the Fling where a combination of good training in the run up and good pacing on the day saw me finish in 9:43, a PB by over an hour.  This positive experience has given me the confidence to try it out in the full WHWR, I've never run further than 53 miles before so I really don't know how well it'll work in practice.

Pacing by heart rate during the Fling resulted in my being able to maintain my effort level right through to the end, something I'd very much like to repeat for the WHWR.  If I can then I should be able to get close to doing an evenly paced split for the first and second halfs (Mlingavie to Achtertyre, Aucteryre to Fort William).  Rosie Bell, the winner of the women's race, came very close to achieving an even paced split last year with her average pace in the second half 97% of her first half average pace.

Will I manage a strong finish like Rosie or slow down significantly like the average runner does?  We'll find out on the 21st!

Splits on the West Highland Way Race 2013

When working out splits I have taken the approach of working out the average time taken for each leg by all the finishers in the 2013 race, this average split data gives me a base from which I derive all other split data.

I don't believe the average splits provide an ideal example of what to aim for on the day - as you include both runners that have a good race and ones that don't, so have come up with a technique for adjusting the split times to achieve different levels of average first half pace/second half pace.  The technique applies a pace gradient to each leg, typically slowing down the early legs and speeding up the later stages.  Using this pace gradient adjustment it is possible to tailor the splits to different first half/second half splits - such as an even split, or adjust to a specific competitor, from the race that you feel paced well and would like to emulate, such as 2013 race winners Rosie Bell and Paul Gilblin, this then provides a split profile which you can use a template for your own race plans.

The first half/second half split adjusted data won't be the actual splits that the runner that they are inspired by achieved last year, it'll be a "Rosie Bell Style" or Paul Giblin Style" split profile that captures how strongly the started and finished.  The actual race splits they achieved will show some legs faster and other legs slower.  This is an advantage of the approach though, as all runners have strong and weaker stages relative those around them, and take different amounts of time at check-points etc.   As the adjusted splits is based on the averages of all runners these individual leg variations are removed.

The follow table shows the "Rosie Bell Style" splits when fitted to the 2013 average time of 25:31, compare the splits to the averages in the table above.

The difference between the two sets of splits is most marked in the early and later legs.  The leg to Drymen in particular shows that an conservative start and strong finish runner like Rosie will be able to run the first leg half an hour slower than the "average" runner, but finish the less leg not much less than half an hour quicker with a pace close to what they began with.

Tailoring splits to your own personal goals/expectations

I don't plan to aim for the average 25:21:58, nor would I expect anyone else to, so the splits tables above are only really useful for illustration.  To be useful to our own race plans we need to decide upon the split profile we want to target and a finishing time, in this post I can't publish a full range of time zones and split profiles as there is an infinite number of combinations!  Instead you use my spreadsheet directly, choose either the OpenDocument link if you use a standards compliant spreadsheet, or the Microsoft link if Excel gets your love.

Open up the spreadsheet and you'll see two sheets, the Splits Calculator sheet that creates the tables as above, or a second Split Projection sheet that I'll cover later in this post.  The Splits Calculator has the fields that are appropriate to edit colour green, the main one of interest is at the top and has "Finished Time" label to the let of it.  Just type in your target time in format hours:minutes:seconds and the spreadsheet will automatically populate the split tables for the various split profiles.

For fun I'll put in 14:59:59 to give everyone an idea of what splits would be required to sub 15hrs for the first time.  Paul cam every close last year, and he's back this year and will have some stella competition that includes Robbie Britton and Matt Williamson so if all the elite's are fit then perhaps this could be the year... :-)

First up we'll use the average split profile to show just how crazy the splits would need to be at the start - faster than the fastest of elites run the start of the Fling at...

For a perfect even first half/second split profile we'd see:

The "Paul Giblin Style" split profile:

Personally I think Paul went out too fast, last years 2nd place finisher Marco Cosani paced more even race, using 16hr splits inspired by Richie Cunningham's previous WHWR win splits.  Marco was catching Paul from Glen Coe an indicator that Paul did indeed go out faster than is optimal.  If a runner was to run with Marco's split profile for a sub 15hr it'd look like:

Even more sensibly/evenly paced sub 15hr would follow "Rosie Bell Style" split profile:

Personally I think Rosie Style split profile would be the one that makes a sub 15 most likely, but the elite runners will be racing each other as much as the clock so the idea of precision pacing is likely to be dispensed quite early.

Finally I'll include split profile based on John Kynaston's 2013 race, relative to Paul, Marco and Rosie he started faster and finished slower, but still started far more sensibly and finished stronger than the average runner.  John isn't an elite like these three runners but loves spreadsheets even more than I so I thnk he totally deserves a stab at a sub 15hr even he won't be racing this year (we'll see him Marshalling with Katrina at Lundavra):

Nothing like setting of at tempo pace and hoping that you'll be able to keep till the end, GO Kynaston GO! :-)

For your own race if you want to use the splits calculator pick one of the race profiles and the target time you want and then print off and laminate the splits.  If you are struggling to get things to work and need some splits just let me your target time and preferred profile as a comment below and I can plugin in your target time and post the splits in a follow up post.

Finish time projections based on splits as the race progress

The second sheet on the spreadsheet provides just the time through each checkpoint for a range of different split profiles and times.  If you have a runner that you are following that is like to pace like Marco Cosani then plugin the base time into the green field and the increment you want.  For instance for the sub 15hr time and look down to Marco's split profile entry, it'll look like click to see a big version.

Personalizing the split profile to your own requirements

So far I've only discussed and illustrated changing the target time but entering different times into the green "Finish time" box, in the split profile examples for Paul, Rosie, Marco and John you'll also see a "PaceGradient" green box with differing negative values in it.  You can enter your own values in here to control how much slow down/speed up relative to the average splits that are used as the base for the splits computation - in all the examples I've given the gradient is negative, although reflecting on it now it should probably be labelled slow down gradient... this work is bleeding edge so likely to be a few bits that redefining later....  

Anyway the details of the computation aren't actually too important, then effect of them is the key, to the right of the PaceGradient box you'll see two automatically computed values, PaceScale - a multipler used to make sure all the legs% add up 100% for a given PaceGradient, and the part that humans will be interested in the SplitRatio - which is the ratio of SecondHalfAveragePace/FirstHalfAveragePace which is 1.0 for even splits. 

You can try different  different PaceGradient values till you get the SplitRatio you are after or simply see the changing paces through the race that start to make sense to what you are aiming to achieve.  I used a couple iterations of trial and error to get the appropriate PaceGradient for the even split, Paul, Marco, Rosie and John's pace profiles.  If I were programming this rather than using a spreadsheet I'd be able to get the program to compute the pace gradient given the SplitRatio but I'm not good enough spreadsheet wizard to do this.  Perhaps John can work out how to do it :-)

How to use splits before and during the race

Before a race having an idea of the time for each section can give you a guide to the time of day you are likely to arrive and how long each section is likely to take.  How long it takes will help guide how much food and drink you'd want to take, and whether this means you'll need a pack or can rely on handheld bottles and pockets in your shorts.  

During the race you'll want to keep things simple and just carry the splits you need, this will probably mean editing the spreadsheet to get only the data you want and printed off in nice clear form then laminate.  John Kynaston creates mini split cards for each section and picks these up at each check point/aid point.  The splits I have aren't as fine grained so you can probably get away with just one splits card for the whole race.

As I pace by heart rate I don't need to worry about fine grained splits for pacing, but others find it useful, and rather than preach how I might do it when I don't use it, I thought I'd ask, John with a few quick questions, here's his answers:

RO: What is the bit you always look at when you are running? 
JK: When I'm running I like to keep an eye on my average pace for my mini-split but I also like to keep track of how far to go to the next checkpoint and once there compare my time to my plan.  

RO: Do you check progress/pace between the mini splits and compare it to your pace specified on the split card? 
A: Yes ... but I try not to get too discouraged if I'm behind on a particular split. I do like to have realistic splits and hope to be ahead of them as I find this more positive. If I have a 'poor' split I try to put it behind me and try and concentrate on the present one. 

RO: Do you keep pressing the lap button on the watch or do you just leave it running?
JK: Yes I press the lap button each time I get to my mini-split.  I make sure the watch is NOT on automatic mile splits.

Thanks and see you on Saturday!

Thanks to John for helping provide his own splits for different races and with helpful suggestions, no doubt there is still plenty I can improve upon.  Thanks also to John for doing the WHWR podcasts and all those that helped provide lots of great insights, advice and inspiration.

It's now full two sleeps and an attempt at a Friday afternoon nap till Race day.  I can't wait, or at least I can wait a little as not all my race kit has arrived in the post yet :-)

I look forward to see all runners, support crew and marshals on Friday, Saturday and on Sunday at the Prize giving.  May the Goblet be with you!!! 


  1. Good luck - not that you need it, you are very well prepared!

    1. Thanks Thomas, not sure about very well prepared as I've still lots of items on my prep TODO list to complete...

      I've been wondering if that one day you might come race the West Highland Way Race or one of the other Scottish ultras... can we tempt you? :-)

  2. I can't pretend to really understand the workings behind pacing an ultra, but its a fascinating post all the same!! All the very best for this weekend, Robert - hopefully we'll bump into each other in Milngavie, or after you collect your goblet in FW. x

    1. Thanks Sarah. You'll be undertaking quite a challenge doing support as well. Should be an adventure for us all, one more sleep and we head north yipeee!!

  3. Robert , you haven't come clean on what your own expectations are! I know, you will run comfortably within your HR limits and the time will take care of itself............... but you must have some idea. Based on your Fling time and your approach I'm hoping you should make it under 22 hours, but anyway, good luck however it turns out.

    1. I was planning a follow up post on my own race plan but am still getting my plan and kit together... but won't have a chance to now as time is too short.

      My goals are:

      Platinum 20hrs - I don't think this is likely, but you can dream,
      John Kynaston and Debbie Matrin-Consani both run sub 20hr off similar Fling times to mine, so
      I now if you are well prepared enough and run a great race it's possible... this was also not their
      first WHWR so doing it on a first try it a tall order... but as I said, you can dream the perfect race

      Gold - 21:17 This is 2.19 x my Fling time, the 2.19 multiplier is something John Kynaston worked out from
      the average of previous Fling/WHWR finishers. Seems like a good place to hang my realistic
      but still very competitive target.

      Silver - Sub 24hrs, I'd still be very happy with a sub 24, and think I should be able to do it as long as I run as
      sensible race and don't injure myself.

      Brozne - sub 35hrs so I can claim my first Goblet. Any finish of 95 mile race is pretty awesome in my book,
      for me to go over 24hrs it likely means I had a very tough race for some reason and had to
      rest up and walk long periods.

      Hopefully I'll get a chance to spot you in the crowd tomorrow evening before we head off and wish you the best before we head off, if not may your day be fun filled and awe inspiring as the scenery is set to be.