Haile Gebrselassie

Haile Gebrselassie: 2:03:59 WORLD RECORD in Berlin
Analysis, splits and race report

Haile Gebrselassie has broken the Marathon World Record AND the 2:04 barrier - 2:03:59. In an incredible performance, he first dealt with the surprise challenge of James Kwambai, who had the race of his life to keep Geb company all the way to the 35 km mark, and then went on to fly to the finish line and take 27 seconds off his one year old record. It was a spectacular final 12 km that did the trick, because he was more or less exactly on the pace up to the 30km mark.

Below is my "real time" analysis of the race, including split times and paces of the magnificent performances.

Race report

The Berlin Marathon, and Haile Gebrselassie's shot at his own world record, is underway. Welcome to our "real time" analysis of the race. As far as Marathon analysis goes, this race will be relatively easy - no pesky race tactics or surges, just one man, 4 pacemakers, and the clock! The analysis is below...

Pacing and splits

The graph below shows the split times for each 5km interval, and the pace and projected times at each point. Below that is my "real time" commentary of the race, as it unfolded.

Note: These times are off the TV coverage - the official splits may differ by one second - they change the projected time by a whole 1 second, at most. So I hope you'll bear with the margin for error (1 sec) in return for getting these up within 15 mins after the race...

5km - too fast early?

The 5km mark was reached in 14:35, a pace of 2:55/km. I feel that is too quick, which is actually what was predicted in the preview - perhaps the pacemakers and Geb got a bit carried away with the occasion. Unless it slows down pretty soon, the chances of the world record may be gone by the 15km mark...

Speaking of pacing, since our preview post on Thursday, I learned that Geb had spoken about a first half in the same time as last year - 62:29.

That might have been fine for the world record (2:04:26), but if Gebrselassie is serious about cracking 2:03-something (which he did say he was chasing), then his projected pace is too slow. To run a first half of 62:29 would require a second half of 61:31, which would represent a massive negative split of one minute. That's not only undesirable, it's probably impossible. Turns out it's a moot point anyway - they are going to hit halfway in 61-something at the current pace.

10 km - still too quick

The second 5km were covered in 14:37, the same pace as the first 5km. So the signal to slow down has clearly not come yet...whether that's a conscious decision, that the pacemakers are being spurred on by Gebrselassie, I don't know. Regardless, if he continues at this tempo, he'll hit the half marathon in 61:11. That's not going to happen (they'll slow down), but in my opinion, they have about 5km to sort this out - if they don't, the chances of the record will disappear under the quick pace...

Perhaps Geb believes that this pace is managable - in Dubai, he hit 10km in 28:35, or thereabouts, and maybe that gave him confidence that a brutal early pace is in fact possible. They're coming up to 15 km now...

15 km - pace slows down, back to "reality"

The 5km interval from 10km to 15km was covered in 14:51, which is a substantial slowing in pace. At the 10km mark, the pacemakers were visibly discussing something with one another - perhaps they made the decision to back off slightly? The pace certainly has dropped, although the projected time is still sub-2:04. That interval is probably just what was required, to settle the race down a little. The chances of the record are still very much alive.

20 km - another 'steady' interval

2:57.4 per kilometer for the last 5km interval, which was covered in 14:46. If you look at the graph above, you'll see that the last TWO intervals have been run SLOWER than what was required at the outset to break the world record. As a result, Geb's projected time is drifting slowly back towards to the world record - it now projects 2:04:07. If that trend continues, Gebrselassie will need to find a brilliant final 10km to break the record.

The commentators are currently waxing lyrical about how he's going too fast. If you take the halfway split in isolation, then they're right, of course. But the signs are ominous, because having started so fast, the pace has now settled down into what might be too slow. If they keep up the pace from the last 10km, then he'll finish in 2:04:36. So rather than being too fast, the last 10km have been too slow - it's only the incredibly fast early pace that has created the illusion of going too fast. The race really begins at 30km, however, and so that's when the action will begin.

25 km - settled pace continues

A third successive 5km interval has been run SLOWER than the intial pace required. That might turn out to be a blessing, because Gebrselassie may just be gearing up for a big final 10km, when he is almost certain to be alone. But his projected time, having been 2:03:12 at the 10km mark, is now 2:04:20, and so the last 5km are certain to be captivating. Gebrselassie will be gathering himself for that final assault. He still looks very comfortable, no signs of trouble yet.

30 km - the second half begins...

We're in a bit of a 'holding pattern' - the same consistent pace has been maintained at the front - 14:45 for the last 5km (projected 2:04:21). But the pacemakers have been thinned out, and only Abel Kirui has survived beyond 30km. Gebrselassie has other company in the form of Kenyans James Kwambai and Charles Kamathi, and it is interesting that they've stuck around for quite so long. However, the real race begins now.

35 km - down to two

The pacemaker and Charles Kamathi both dropped off at about 32km, and it was down to two. Gebrselassie and James Kwambai, who was racing out of his socks to stay on this pace. The last five kilometers were run in 14:40 - they've dipped back under the world record pace.

The presence of Kwambai may prove to be a major boost for Gebrselassie - he has some credentials, having finished second in Boston before. Gebrselassie is pressing hard at the front, and Kwambai is hanging on half a stride behind him, looking reasonably comfortable, it has to be said. He may well be in this the whole way. Will he be the spur Gebrselassie needs...? Then again, Gebrselassie has never excelled in marathon RACES, only time-trials. Perhaps he prefers being alone, and Kwambai will be a hinderance. Tactical games may slow the athletes down, as a result of watching one another, as happened in London this year. It's certainly a pleasant, unexpected angle to the race.

40 km - Gebrselassie breaks away, the clock is all that remains

It's now a solo race against the clock, which is what was expected all along. The last 5km were run in 14:29, the fastest of the race, and brings Gebrselassie within sight of the sub 2:04 clocking! Fabulous running, he's visibly increased the pace, and it's a majestic run for the line now.

Finish line - 2:03:59

Enough said! The record is gone, by 27 seconds, in the greatest marathon (time-wise), ever run! An incredible final 7km in particular, and the sub 2:04 performance was delivered. Haile Gebrselassie didn't need more confirmation of his status as the greatest distance runner ever - that position was secure. He gave us a reminder anyway, with world record number 26.

Considering the young generation coming up behind him, perhaps he did need this race, but it leaves the marathon in good shape for the future!

Join us tomorrow for more insight, once the dust has settled. We'll also look at Irina Mikitenko's fantastic sub 2:20 performance (fourth fastest ever).

But for Geb's race, we'll analyse that pacing strategy in more detail, give some thoughts on the race and also compare this performance to other great runs, including last year's Berlin World record!

Join us then!
Ross