Thursday, 24 April 2014

Good Day/Slow Day Highland Fling Splits

Running ultra marathons using splits is popular, with a range of approaches used, from using ones own splits from previous races on the route, using splits from training through to adopting average splits of previous races.  In this post I'll publish two sets of splits for the Highland Fling race, one set is are based on previous winner time and represent the splits that are likely to what we might achieve is we run the perfect race and finish strongly, while a second set is based on the average splits recorded during the 2013 race but as I believe most runners set out too fast and suffer in later stages using these splits for me represents when things don't go to plan and how you are likely to slow.

Winners splits are consistent year on year

The winner splits I am using are Andrew James's 2011 winning performance of 7:12 where he set the course record.  This course record was broken Lee Kemp last year (2013) where he lowered the record to an astonishing 7:02:50.  However, the route was altered last year due to path upgrading near Drymen, and Lee made a navigation error on route to Rowardennan which lost him time using his splits aren't likely to be as representative of a perfectly run race on the normal route, so I stepped back to Andrew' performance as my gold standard of splits.  

My confidence in these winning splits being pretty near optimal is based on analysis of previous years winners and course records, where there the difference between male winner splits is less 2% from each other, while others in the top ten and behind on average moving increasingly away from how the winners have consistently paced their races.  The following graph shows previous winners in men's and women's category in relation to Andrew's splits.  

Curiously only the two fastest female winners Lucy Colquhoun and Emma Rocha come close to this type of pacing and their times are way out ahead of other female winners.  I suspect the smaller number of international class woman attending the Fling makes the races less competitive at the front end compared to the men's race that has consistently seen near course record race wins over a number of years.  With less competitive races it's possible for a winner to get the pacing wrong and struggle on the day but still able to win.  Tracy Dean's courageous win last year saw here nursing a injured calf through the majority of the race and time and splits consequently suffered but she was still able to pull out a win. In the men's race it's been so competitive that a small slip up in racing will see them overtaken.

The key take home is that winners pace more consistently and evenly than the almost the whole of the rest of the field, the start at an relatively easy pace for them and finish strong while everyone else around them falters.  The "almost" the whole of the rest of field needs to be qualified, when you analysis all splits there are runners further back in the field that pace evenly as well, and my experience with using a HR monitor to guide pace has resulted in pacing that is far closer to winners splits than the average.  

Runners should take encouragement from this, while most go out too fast and suffer as consequence, it is possible to run your own race well, starting off easy enough that you can keep the stress on your body lower enough for spare glycogen stores, avoid early muscle damage and enable the body to keep digesting food and water.  If you start off easy then there is much greater chance of finishing strong, with the bonus of the experience being a positive one where one enjoys the race rather battling through from early in the race.

How to use splits when planning your own race

Knowing just how easy you should start off is difficult, first you need to know what time you feel can likely achieve on the day, then from this you can select the splits to match this time.  As there is no way we can know exactly how well the race will go on the day I believe it's best to not to try and pick on single target time.  

Instead I would advocate thinking about what your perfect race would turn out like and what time you might achieve if everything went to plan and you raced the perfect splits and the best of your ability, this is your time would define your "Good day" target time and the splits to select from this would be from the Winners splits, where you start steady and finish strong.

Not all races go to plan, it could be that we over estimate our fitness and choose splits way beyond our capability and go out too fast and then struggle, it could be that we get cramp or gastric stress that slows us down, or injure ourselves and then struggle for the rest of the day.  Choosing splits that account for this possible slow down provides us with a fallback set of splits to allow us to estimate our finishing time - these would be result in our "Slow day" target time and associated splits where we start a too strongly for our capabilities on the and finish slow.

If you choose your splits well then it's likely you'll be in between your "Good day" and "Slow day" splits, the nearer you are to your "Good day" splits the better your judged the day and paced the race. What will be apparent from the splits I present below is that the "Good day" and "Slow day" splits start close together near the start and steadily the "Slow day" splits drop off the pace and result in a much slower finishing time.  

For instance if I set off with with 1:50 split to Drymen (9:04 min/mile pace) then the "Good day" splits would see me finish in a time of 9:29:51, while the "Slow day" splits for a 1:50 initial split would see me finish in 11:13:21. If had chosen to use the "Slow day" splits for a finishing time of 9:30 we would have to do the Drymen split in 1:33 which is blistering 7:41 min/mile pace.  Such a strategy is likely to ruin our chances of succeeding on hitting our target time as even the "Slow day" splits would likely become impossible to sustain as we'd be burnt out by the time we reach Balmaha.  

Follow are the splits for the various legs and based on different times for the initial leg to Drymen, note the splits for sub 7 hours course record starts at 1:21, just a shade faster than last year.  The "Good day" splits are based on Andrew James 2011 race, while the "Slow day" is based on the average splits from the 2013 race.  I could label this as "Average day" but I don't believe average splits are best to aim at, instead we should aim for a perfect race, and if things don't go to plan we know we'll still finish in perfectly decent time. An important takeway is that for the perfect race you don't need to start fast to get a great time.

Leg ratio Completed Leg distance Leg ratio Completed Leg distance Leg ratio Completed Leg distance Leg ratio Completed Leg distance
Winner 0.1930345572 12.13 12.13 0.2703255168 27.09 14.96 0.3003702561 41.04 13.95 0.2362696699 52.87 11.83
Average 0.1633621928 12.13 12.13 0.2641205352 27.09 14.96 0.3171045563 41.04 13.95 0.2554127157 52.87 11.83

Drymen Leg Pace Rowardenan Leg Time Leg Benglais Leg Pace Tyndrym Leg Pace

Good day 01:20:00 01:20:00 06:36 03:12:02 01:52:02 07:29 05:16:31 02:04:29 08:55 06:54:26 01:37:55 08:17
Slow day 01:20:00 01:20:00 06:36 03:29:21 02:09:21 08:39 06:04:38 02:35:17 11:08 08:09:43 02:05:05 10:34

Good day 01:21:00 01:21:00 06:41 03:14:26 01:53:26 07:35 05:20:28 02:06:02 09:02 06:59:37 01:39:09 08:23
Slow day 01:21:00 01:21:00 06:41 03:31:58 02:10:58 08:45 06:09:11 02:37:14 11:16 08:15:50 02:06:38 10:42

Good day 01:25:00 01:25:00 07:00 03:24:02 01:59:02 07:57 05:36:18 02:12:16 09:29 07:20:20 01:44:02 08:48
Slow day 01:25:00 01:25:00 07:00 03:42:26 02:17:26 09:11 06:27:25 02:45:00 11:50 08:40:19 02:12:54 11:14

Good day 01:30:00 01:30:00 07:25 03:36:02 02:06:02 08:25 05:56:05 02:20:03 10:02 07:46:14 01:50:09 09:19
Slow day 01:30:00 01:30:00 07:25 03:55:31 02:25:31 09:44 06:50:13 02:54:42 12:31 09:10:55 02:20:43 11:54

Good day 01:35:00 01:35:00 07:50 03:48:02 02:13:02 08:54 06:15:52 02:27:49 10:36 08:12:08 01:56:17 09:50
Slow day 01:35:00 01:35:00 07:50 04:08:36 02:33:36 10:16 07:13:00 03:04:24 13:13 09:41:32 02:28:32 12:33

Good day 01:40:00 01:40:00 08:15 04:00:02 02:20:02 09:22 06:35:39 02:35:36 11:09 08:38:03 02:02:24 10:21
Slow day 01:40:00 01:40:00 08:15 04:21:41 02:41:41 10:48 07:35:47 03:14:07 13:55 10:12:08 02:36:21 13:13

Good day 01:45:00 01:45:00 08:39 04:12:03 02:27:03 09:50 06:55:26 02:43:23 11:43 09:03:57 02:08:31 10:52
Slow day 01:45:00 01:45:00 08:39 04:34:46 02:49:46 11:21 07:58:35 03:23:49 14:37 10:42:45 02:44:10 13:53

Good day 01:50:00 01:50:00 09:04 04:24:03 02:34:03 10:18 07:15:13 02:51:10 12:16 09:29:51 02:14:38 11:23
Slow day 01:50:00 01:50:00 09:04 04:47:51 02:57:51 11:53 08:21:22 03:33:31 15:18 11:13:21 02:51:59 14:32

Good day 01:55:00 01:55:00 09:29 04:36:03 02:41:03 10:46 07:34:59 02:58:57 12:50 09:55:45 02:20:45 11:54
Slow day 01:55:00 01:55:00 09:29 05:00:56 03:05:56 12:26 08:44:09 03:43:14 16:00 11:43:57 02:59:48 15:12

Good day 02:00:00 02:00:00 09:54 04:48:03 02:48:03 11:14 07:54:46 03:06:44 13:23 10:21:39 02:26:53 12:25
Slow day 02:00:00 02:00:00 09:54 05:14:01 03:14:01 12:58 09:06:57 03:52:56 16:42 12:14:34 03:07:37 15:52

Good day 02:05:00 02:05:00 10:18 05:00:03 02:55:03 11:42 08:14:33 03:14:30 13:57 10:47:33 02:33:00 12:56
Slow day 02:05:00 02:05:00 10:18 05:27:06 03:22:06 13:31 09:29:44 04:02:38 17:24 12:45:10 03:15:26 16:31

Good day 02:10:00 02:10:00 10:43 05:12:03 03:02:03 12:10 08:34:20 03:22:17 14:30 11:13:27 02:39:07 13:27
Slow day 02:10:00 02:10:00 10:43 05:40:11 03:30:11 14:03 09:52:32 04:12:21 18:05 13:15:47 03:23:15 17:11

Good day 02:15:00 02:15:00 11:08 05:24:03 03:09:03 12:38 08:54:07 03:30:04 15:04 11:39:21 02:45:14 13:58
Slow day 02:15:00 02:15:00 11:08 05:53:16 03:38:16 14:35 10:15:19 04:22:03 18:47 13:46:23 03:31:04 17:51

Good day 02:20:00 02:20:00 11:32 05:36:03 03:16:03 13:06 09:13:54 03:37:51 15:37 12:05:16 02:51:21 14:29
Slow day 02:20:00 02:20:00 11:32 06:06:21 03:46:21 15:08 10:38:06 04:31:45 19:29 14:16:59 03:38:53 18:30

Good day 02:25:00 02:25:00 11:57 05:48:03 03:23:03 13:34 09:33:41 03:45:38 16:10 12:31:10 02:57:29 15:00
Slow day 02:25:00 02:25:00 11:57 06:19:26 03:54:26 15:40 11:00:54 04:41:28 20:11 14:47:36 03:46:42 19:10

For myself my own targets for this years race (this Saturday) are:

Platinum : 9:30hrs, unlikely but if everything went perfectly and raced above expectations

Gold: 10:00hrs , a more realistic goo day target, I would love to get a sub 10 hours, even if just be few seconds

Silver: 10:20hrs, I think as long as I don't mess things up it should be possible.

Bronze: 10:46hrs, to match match my Highland Fling time of 2012 where I had a good day, but I'd honestly be a bit disappointed if two years on and all my good ultras in Autumn of 2013 that I wouldn't be faster.  Ultra's can be cruel though, you just don't know what's going to happen on the day.

This leaves me looking at a 1:50 to Drymen for me to get within range of my very lofty target of 9:30 hr time, to go out slower than this will likely start reducing my chances of achieving it.  If this was too fast then my chances of doing my more realistic target of 10 hours will be reduced too.  A conservative start would see me look at 1:55 initial which would make a sub 10 hour time more achievable.  

The low risk, lower reward approach would be to head out at ten minute mile pace and arrive at Drymen at around the 2 hour mark.  In my 2012 race I arrived at Drymen in 1:58 and went on to finish in 10:46, but I struggled with low energy levels in the last three miles and had to walk far more of the final miles than I was happy with.  If I am indeed fitter and in particular better fat burner as I suspect I am then I should be more capable of keeping the pace up right through to the end, so even with a slower start I could still realistically be thinking about my silver target of 10:20. 

Based on the "Good day" splits I know that I'd be crazy to start out faster than 9 min/mile pace as it'll likely be unsustainable. If I do end up going through Drymen in a sub 1:50 time I'll know that I need to back off and start conserving my energy better.  If I go through Drymen between 1:50 and 1:55 I know that I'm being competitive but not necessarily fool hardy if it is to be my day and I can manage to stick close to the "Good day" splits through to the finish.

Pacing by Heart Rate

In my recent post How to pace the perfect marathon or ultra marathon I advocated pacing using a Heat Rate Monitor, which will probably lead one to question this post, which one is best?  Race by splits or race by Heart Rate?  If the later what what heart rate to use?

My plan this Saturday is pacing primarily using my Heart Rate monitor with "Good day" splits as a secondary guide to whether I might be over cooking it and also to help project what time I might be on for. If I am able to stay with a heart rate in my target zone for a section and within range of "Good day" split will be reassuring, if I'm too slow or too fast then I will have a bit of feedback on just how well my body is coping on the day - if my HR is high for a given pace then I'll know to be more cautious, but if it's lower then I could be my day I have the headroom to chase a fast time and more optimistic splits.
Finally knowing that I might still be on for a sub 10 hours time could be quite motivating in the last section from Beinglas to Tyndrum.   

The advantage that pacing by Heat Rate has is that helps you even out all the ascents/descents and changes in terrain/weather so you avoid pushing on too hard on ascent, or taking it too easy on a descent.  Running across hilly terrain makes following average pace splits next to impossible, by contrast pacing by HR needs no mental arithmetic, just a quick glance down to the HR monitor watch to check that I'm within a target zone.

Pacing by HR can also be tweaked to fit with average target pace, if you are slightly too fast adjust the target HR down, if you are a bit slow adjust it up.  

In my 2012 Highland Fling race my HR monitor recorded an average HR of 152, in my 2013 Devil O'Highland Race my average HR was 154 (it's a shorter race so higher intensity).  I expect if my race goes to well then I'd see a HR around that of my 2012 Highland Fling race and may actually be a little lower as I'm two years older (max HR goes down with age), and hopefully fitter thanks to my West Highland Way race training,

Based on an assumption of an average HR of 152 by plan is to account for HR drift through the race, breaking the race into three sections:
  1. Milnagaive to Drymen, start first few miles at a target HR of 140, but keep below 150 to Dyrmen
  2. Drymen to Beinglas, the central 6 hours of racing I'll target 150 to 155, with a hard limit of 160 being a Heart Rate to NEVER exceed.  I have found when I exceed 160 in ultra's cramp and/or gastric stress have often been close behind.
  3. Getting to Beinglas in good shape is my key objective, running conservatively be holding myself back  to not got faster than the "Good day" splits and keeping the HR sensible should ensure that my body is able to conserve glycogen stores, avoid overheating and associated fluid loss, and keep my digestion ticking along nicely.  
  4. Once I get to Benglas I'll asses how my body is responding and look to race the last 12 miles to the best of my ability, splits will be out of the window, and I'll relax my target HR allowing it to go higher if my body is able.  If I've got cramp, suffering with gastric stress, muscle fatigue etc. then I'll keep the HR target range conservative.  If my body can handle increasing the effort level then I'll happily let it go up to around 160, and in the last few miles even that limit I'll happily ignore if my body is responding well.
If you are thinking about using HR monitor for pacing as well then using previous average HR for a previous race of similar race is good place to start.  If you don't have this then aiming for around 25bpm lower than the HR you'd see at lactate threshold or average 10k racing HR would also likely be a reasonable target.  For reference by 10k race HR averages around 178, while my max HR is around 190.

Best of luck with race if you are racing and see you there. In case your wondering what I'll look like come race day, here's a photo from last years very wet Devil O'Highlands Race.  Sadly the forecast isn't that great so I suspect my Montane running jacket, worn below, will see plenty of action again.


  1. I agree that even splits are likely to be best, and this can be demonstrated by retrospective examination of the splits of winners (at least for males). The simple lesson is that one should start fairy conservatively. But how conservatively? The challenge is knowing what is the fastest initial pace will allow even splits on a particular day. I agree that your strategy of using HR to guide you is fairly sound, provided you start in a well recovered state.
    I hope the Fling went well.

    1. My Fling race went very thanks, I'm in the process of writing a blog post for it, but am a slow writer...

      Now I've done the Fling again I can reflect on splits I achieved and compare them to the Good Day/Slow Day splits. At the halfway point it looked like I was on for a 10:15 time if I matched the "Good Day" splits based on Andrew James' 2011 course record. What surprised me is that by the 40 mile point I was on target for a sub 10:00 having run proportional strong than Andrew from the half way point. My finish time was 9:43 which is well inside 10 hours. My pace over the last three relatively flat miles was very similar to that of my first 12, so as much as is possible over such a hilly trail I achieved pretty consistent effort level through the race. I believe using a HR monitor as a guide is big part of this.

      Running a proportionally stronger finisher than a previous record holder surprised me, and makes me suspect that Andrew actually went out a bit too fast in the first half and couldn't quite maintain the pace through to the end. He was racing others rather than the clock so this may have been a factor, also I don't believe he's a trail specialist. I now believe my own split % from the 2014 race would probably form a better basis for "Good day" splits so will revise them in a later article.

  2. Robert its an interesting post, especially looking at your splits for this years Fling. I ran the Fling too and finished 34 seconds behind you. My target was sub 10 hours so I was delighted to be 9:43. My pacing was very different from yours as I went out harder than I would normally inspired by a blog post by Stuart Mills. I was at Drymen 8 minutes ahead of you, 17 at Rowardennan, 8 at Beinglas and then you finished 34s ahead of me at the end. I am still debating what is the best approach for pacing. The even pacing is more satisfying (my normal approach) but the going out harder at the start did make me feel I had put everything on the line and I felt I crossed the line and given my all. When I ran the D33 earlier this year I ran it all at 8:30 pace until the last six miles which I increased to 7:30 pace, it felt great to finish strong but could I have gone faster earlier and got a better time?

    I think pacing is something that needs to be experimented with just like food and kit. Not sure if I have found my personal answer yet. If you've not read it its worth looking at Stuart Mills views on pacing.

    1. Well done on the sub 10. Your own pacing was actually far more even than most runners, you were in top 12% of field when looking at second half/first half splits - so even though you went out harder than you usually do, you still went out far less hard than the vast majority of the field, including those around us.

      I believe Stuart's approach maximizes the perceived effort level rather than produce the best finishing time. Going out fast means that you are depleted and fatigued much sooner in the race and have to work hard throughout the whole race to maintain even the slowing pace that the approach "deems" acceptable.

      My own splits were in the top 1% of strong finishers, only 3 people ran proportional stronger second halves. I was 120 places behind you at Drymen but was able to finish a few places ahead, and you ran a relatively strong second half. From this I take the message that one doesn't need to worry about not banking time early in race, as long as you look after yourself properly you'll finish strong and make up all the time and more on your competitors in the second half when they slow and you don't.

      As important as the actual finishing time/place is I feel the manner in which the race is carried out is just as important. Pacing evenly meant that I didn't have any real low points in the race, there was no battle against fatigue or pain, finishing strong wasn't a battle of wills, it was remarkably easy to keep moving at a good pace.

      Even pacing requires discipline and practice but come race day it has to be the least strenuous way to achieve a given time. Theoretical analysis suggests this will be the case, and my experience of racing in ultras is backing this up. If you want to run fast and really enjoy whole experience then pace as evenly as you can - use your body in the way it works best and it won't complain.

    2. Thanks Robert, I meant to add it was my best relative performance to date. Last year all my races I was in the top 30%, this year the D33 and Katrine were 20% so the Fling at 12% was a big leap in improvement. All the others I aimed to have a better performance in the second half of the race. Perhaps I had been holding back too much. Its finding the balance of effort. Obviously heart rate is one way to balance effort, but I prefer to listen to my body.

    3. Hi Stuart,

      Judging how hard you can sustain an effort for the whole race is difficult, being too conservative with your expectations can limit the performance you achieve, but I believe that typically going out too aggressive if far more likely to result in an overall slower performance. In my previous blog article on "how to run the perfect marathon or ultramarathon" I touched upon the concept of "Probabilistic Pacing" where you view pacing as affecting the probability of different outcomes. Once you view pacing this way you can start to assess the risks/rewards involved with different pacing strategies.

      As for using heart rate vs listening to your body for pacing, the heart rate monitor listens to your body in a complete objective way, there is no subjective interference from mental state. If you really want to listen to your body then the heart rate monitor is pretty unmatched as tool right now in terms of convenience and relevance. If you use a heart monitor properly it won't exclude listening to other aspects of your bodies functions, instead it just adds more information that enables you to make more objective conscious and subconscious assessments.

      What I have found is pacing by heart rate typically holds me back more in the first half, but in the second half when I start to think wouldn't it be nice to walk this incline because I'm feeling tired, checking my heart rate monitor gives me to boot up the arse to say that my bodies doing just fine and there is no excuse to walk. In the later stages my subjective feel for how hard I'm working can well under what my body is actually capable of handling, it's quite easy to fool yourself that it's OK to take it easy if you just listen to internal chatter.