Berlin 2011: Inside Makau's 2:03:38 World Record

Inside Makau's world record

Patrick Makau has broken Haile Gebrselassie's 3-year old world record with a sensational run in Berlin, clocking 2:03:38.  You can read my real-time comments and thoughts at the live post I did during the race, but below is a more detailed look at Makau compared to that 2008 record of Geb's also in Berlin.

The graph below compares the 5km splits for Makau's 2:03:38, to those in the previous record of 2:03:59.  It's an ideal comparison because it's the same course and very similar race situations with an organized attempt, pacemakers etc.

The red line shows Gebrselassie's 5km splits, the green line Makau's.  Below each marker is the cumulative time gap between Geb and Makau in seconds.  And at the top are the times for the individual segments, with the 2011 vs 2008 gap for that segment beneath it.

Makau's is a world record that was clearly established from 10km to 35km, whereas Gebrselassie owed his to a very fast final 20km.  He got faster and faster from 25km onwards, whereas Makau put himself in an excellent position by halfway, blew the race wide open with an incredible sixth interval of 14:20 (25 to 30km), and then hung on to the finish.  

The precision of pacing

The first point about looking at pace is to emphasize just how incredibly precise one has to be to run at this level and to beat a world record.  Consider for example that we're celebrating Makau today because he broke the record by a fairly large margin - it was 21 seconds.  Geb had broken it by 29 and 27 seconds in his two records, respectively.

In fact, in the last seven records, the margin of record-breaking has been 23 seconds, 4 seconds, 43 seconds, 29 seconds, 27 seconds and now 21 seconds.  The point is - world records don't get "smashed" anymore.  And 21 seconds over 42 km, you can work out pretty easily is 0.5 seconds per kilometer and thus only 2.5 seconds over each 5km interval.  I think it's safe to say that any runner who goes out and runs 2 sec/km faster than a world record is heading for a major meltdown!

That's amazing precision, and should be borne in mind when considering the splits, because the elite field are on a razor's edge, and so particularly the first half has to be very accurately paced.  Today, the first half was run just under one second per kilometer faster than the previous record, and it was consistently paced up to halfway.  There was talk of it being too fast, but in the end, it was a sustainable pace...just...and so very good to begin with.

Makau's race - building a 'buffer' and hanging on 

The dashed line in the graph above shows the required pace per 5 km segment to match the world record - it's 14:41.5 per 5km.  Today, Makau was faster than this all the way to 20km, building up a 'reserve' so that by 20km, his time of 58:30 put him on course for a 2:03:25.  Also, up to this point, the pace had been very consistent - those "bumps" you see in the first half represent a range of only 1 second per kilometer, and the field was always faster than the required pace.

The comparison to Gebrselassie in 2008 is also interesting.  The opening 10km was basically the same - 29:17 for Makau vs 29:12 for Geb.  Then, from 10km to 20km, Makau began building his buffer - had there been a "virtual Gebrselassie" on the road, Makau would have been pulling further and further away from him.  

A 14:34 and a 14:39 put Makau up by 12 seconds and then 20 seconds at 20km.  He'd have been over 100m ahead of Gebrselassie at halfway.  At this stage, the commentators were talking about it as being "suicidal", which it certainly was not.  It was fast, but the way the race was to unfold (hindsight is wonderful!), this was actually very solid pace setting, because the race was pretty much on course for a mid 2:03:20 from the outset.

Makau's surge

Then things started to open up.  The race fragmented after halfway - a group of 11 was cut very quickly, and soon it was Makau, Geb and about three pacemakers.  I'm not 100% sure what happened at around 27km, but it was here that Makau made what would be the race's decisive surge.  He went to the front, Gebrselassie stepped off the road just after 27km and Makau just kept going.

He hit 25km in 1:13:18, fully 47 seconds ahead of what Gebrselassie had done in 2008.  That segment, covered in 14:20, was easily the fastest of any of the three world records set in Berlin since 2007.  It was a huge surge, and to put into context what Makau was doing, Geb, despite stopping, was 1:10 down at that stage.  That means that Geb was only 23 seconds slower than he'd been on route to his world record, but now found himself trailing by over a minute (and he'd stopped, of course).

The question now was not so much whether Makau could sustain the pace, but rather whether that surge would cost him?  He had built up such a buffer that he could afford to run the final 10km in 29:49, which was much slower than anything they'd been producing up to that point.

Hanging on, holding pace and running alone  

The next 5km were actually the most significant period for the record - having surged at 14:20 pace, it was here that we'd see a substantial drop-off if there was going to be one.  Makau was now all alone, just the clock and the race car for company.  But he held it together enough to run a 14:38 for the segment.  That was actually quicker than Geb over the same interval in 2008, and meant that Makau extended his "lead" over the virtual Gebrselassie to 49 seconds (as seen on the graph).

Now, with only 7km to, only a massive blowout would cost Makau the record, and the question was whether he'd hold that pace, and break the record by over 45 seconds, or whether he'd come back slightly.  But the world record was now pencilled in.

The answer was that he would pay for the early pace, but only a little.  His final 5km segment, from 35 to 40km was easily the slowest of his race, but it was still 14:59.  Compare Makau's line to that of Gebrselassie, who built from halfway to get faster and faster when he broke 2:04 in 2008.  Makau's graph is "going the wrong" way, but not quickly enough to save the record.

The result of Makau's slowing, however, was that suddenly, that gap of 49 seconds to Gebreselassie in 2008 was cut to 19 seconds.  However, there were only 2.2km remaining, and the pace he needed to run to match the record was now 3:04/km, so his "fatigue" was not costing him enough to save the record.  Also, Gebrselassie had slowed significantly in his final 2.2km back in 2008 as well, so Makau just had to hang on and claim the record.

At 37.2km, with 5km to, I timed Makau as needing to run 15:20 for the final 5km, and it was clear that though he was fighting by this point, he wasn't going to implode to that extent!  The 6:23 for the final 2.2km was in fact similar to Gebrselassie in 2008 (6:25) and it gave Makau the record by 21 seconds.

Optimal pacing?

The first half was run in 61:45, the second in 61:53.  That's near enough even pace (you'd be picky to argue over 8 seconds), but the two halves were constructed very differently - look at the green line in the chart to see that.  Of course, the theoretically perfect race is exactly even, no variation, but you never see that.  Makau was definitely more variable than Gebrselassie in 2008, but that's largely because of the 14:20 he put in just after halfway, and the fact that he slowed to a 14:59 at the end.

The most striking difference between Makau 2011 and Gebrselassie 2008 is the shape of the curve over the second half.  You can see this in the graph very clearly - Gebrselassie got faster and faster after halfway, Makau was up-and-down.  Again, this was the result of his 14:20 and the 14:59 that it 'cost' later.  

The first 20km were a talking point during the race - it was considered too fast to sustain, but in the end, it was pretty precise, based on the halves.  All in all, I'd say it was very well paced, from a global perspective.  But a more conservative sixth interval (25km to 30km) might have left Makau with a bit more for the final 10km, and that may have helped him a few seconds faster.

Future of marathon running - 2:02:59 looms

But really, that's all we're talking, a few seconds.  And it's hard to fault Makau for this performance, which is simply exceptional.  It means that we've seen three performances under 2:04 this year alone, though of course two of them came on the Boston course, which was massively wind-aided this year (and won't stand, though not for that reason!)

Makau is now the marathon king, with three wins in five starts.  It's actually extra-ordinary to think of what Kenya in particular have in the marathon.  Geoffrey Mutai and Moses Mosop were the Boston runners who nearly broke 2:03, and then Abel Kirui dominated the World Championship marathon in Daegu with the kind of performance that would suggest that he's capable of something very close to 2:04 as well.  This record, exceptional as it is, may not have too long a shelf-life!

As always, attention will now turn to the limits for marathon running and people will start talking about sub-2 hour clockings.  That's a little premature - if each record lasts say 3 years (because weather and pacing can undermine even the most gifted athlete), and if the times improve by around 20 seconds per record, we'll be commenting on another eleven world records and will be waiting until 2044 for that to happen.  If it happens at all.  That kind of speculation is always fun though.

It certainly seems as though it's a matter of time before 2:03 is challenged, perhaps a decade or so, if the next generation are as exceptional as the current one.  If it happens, we'll be on it!

Ross