When discussing the performances from Ironman Texas, most of the discussion is focused on two extremes: On one side, the position by Ironman seems to be “most everything went right, so the results are legit”. On the other hand, a lot of people argue that the results are invalid, creating an empty feeling by the athletes who did well in Texas.
I want to argue for a middle position – to me, the results shouldn’t be accepted as valid records while at the same time I want to point out some of the amazing performances we have seen in Texas.
No Records in Texas
There were a lot of issues with the Texas race this year, among them:
- The bike courses short (by Ironman’s own admission), although there is some discussion as to the extent. The general consensus seems to be that the course was close to 110 miles (one example is the GPS data in the Slowtwitch post on Texas). This puts it short of the 112 miles of an Ironman-distance bike leg, but is well within the length of some other courses.
- There were a lot of reports of blatant drafting, both in the Pro fields and among agegroupers.
- All non-competitors (including the draft marshals) were pulled from the bike course for “safety reasons”.
By analyzing the results and comparing the Texas times to other performances, there are a number of observations that show how unusually quick the Texas times have been:
- On average, the Pro athletes in Texas posted bike times that are about 12.2% quicker (or more than 30 minutes for a 4:30 bike ride) compared to their times on an average course, the largest number I have observed in analyzing Ironman-distance races since 2005. The closest bike adjustments are just over 9% (Barcelona 2015, Florida 2013). Just for comparison, the largest adjustment for Challenge Roth (often the “poster child” for a short course) is 5.9% (2009 and 2011).
- If we accept the bike times, this would mean that 9 of the 10 fastest female bike times were set in the 2018 Texas race. The only non-Texas time would be Daniela Ryf’s 4:31 from Roth 2016.
- If we accept the overall times, this would mean that there were 9 national records set in Texas this year. Last year’s race was already quick but only created to four new records.
To me, all this is a clear indication that the data from IM Texas 2018 is far outside of the normal parameters. As a consequence, I will not use them to rate the athletes that raced in Texas and will also not use them for future predictions.
Of course, the data cannot show the reason why the results are off. One can only speculate – were there favorable weather conditions (apparently yes, but it’s hard to imagine improving times by more than a few minutes), was the course short (apparently not by much, at least not shorter than other races that did not produce abnormal data), was there drafting (apparently yes, but of course not uniformly – at least we can be sure that Starky didn’t draft any competitor when he rode his 3:55 bike split), was there a draft by camera motor cycle (unlikely as there wasn’t much coverage of the race and all motos were pulled from the course), etc. Even athletes riding legally will still get a good benefit when they are 12 meters behind a large drafting peloton. Speculating about what has caused a fast time is exactly what I don’t want to do. It’s also impossible to decide on an athlete-by-athlete basis to accept or not accept a performance without a lot more data then what is made available. Of course, this also means that the performance of some athletes get rejected even if they were not impacted by the factors above (or not quite as much) .. and of course I also understand the athlete’s frustration about this.
Still, the Texas times should not be accepted as records. To me, records are like a rubber band – you can stretch it a bit, but when you put too much weight on it, it will break. Following the fastest times across different races and courses was always fun, even if any list of “fastest times” immediately created a lot of “yeah, buts …”. We are now in a place where “raw records” don’t really make much sense.
For a list of fastest times to start making sense again, a few things have to happen:
- Certify course lengths.
This would be an excellent project for the ITU and Ironman to tackle. Define a protocol for how to properly measure (incl. what is called “short course prevention” for road race measurements), set up an organization for measuring courses, pre-measure courses and release these measurements before a race. Then define acceptable parameters for “a full Ironman-distance race”, such as will short courses be accepted and some triathlon-specific issues such as currents in the swim.
- Define minimum officiating standards.
A lot of the standards are already in place, but often not fully enforced. I can’t see how you can have a record when there are no draft marshals.
- Give the women a better chance not to be impacted by the agegroup race.
A ten-minute gap between the Pro women and the fast agegroupers is not sufficient, especially if the Pro swim is without a wetsuit and the agegroup race is wetsuit-legal. Mel Hauschildt wrote on her blog: When the AG men overtake the Pro women, “the women’s race gets put on hold, the safety car comes out and brings the field back together until they’re free to race again a couple hours later.”
Defining the rules on when a race circumstances are acceptable as a record would be a great sign of a further maturing sport. In the absence of these rules, any comparisons between different courses purely on finish time is pretty meaningless and I’m struggling with whether it makes sense to put together any record-lists for now.
Great Performances in Texas
Even if I have argued above that the Texas times are not acceptable as records, there were still a number of great performances in that race. Based on my analysis of race data I have also built an adjustment model that compares results from different courses or races on the same course but different years. These adjustments (creating a course-independent “normalized time”) is not transparent enough to be used as a basis for a “world records list”, but even if the Texas data is likely outside the limits of the model, it still gives a reasonable indication of how the Texas performances compare to performances on other courses.
Here are a few great performances that should get more attention.
Matt Hanson Runs a 2:34 Marathon
When Matt crossed the finish line in first place with a 2:34:39 marathon, I found it hard to believe that the course was accurate – it was almost eight minutes quicker than last year’s winning marathon and seven minutes quicker than his own course record. In addition, by all accounts Matt was working hard on the bike, first being forced to close a gap of three minutes after the swim until the first turnaround and then pushing the pace in the chase group. With the data from the other competitors and GPS data, it seems clear that the course was the same as in previous years and relatively accurate. In my race database (going back to 2005) I could only find 11 sub-2:40 marathons, with the fastest being a 2:37 by Bart Aernouts from IM France 2013 on a course that was likely to be a bit short. (There is also Peter Reid’s 2:35:21 run split at IM Austria 1999, which was short by at least 1.5k.) When looking at the normalized time of Matt’s run, my analysis lists only two better run performances: Patrick Lange’s Kona runs from 2016 and 2017 that were faster by 38 and 44 seconds. I can’t wait to see what happens in Kona when Matt figures out how to run well in Kona!
Mel Hauschildt Wins Her Third Regional Championship on Different Courses
When Mel is healthy and able to train consistently, she is almost impossible to beat. Since stepping up to Ironman racing in 2014, she has won six Ironman races and is unbeaten in any Ironman she raced outside of Kona (DNF in 2016 and a 14th in 2017 when she didn’t have much time to prepare after two surgeries). As is typical for her, Mel lost some time in the Texas swim, she had to make up three minutes to most of her competitors. It took her about 90k to close that gap and usually she’s riding even better in the second half of the bike. In her blog Mel describes how frustrating it was to her to work hard only to be overtaken by agegroup men being able to slingshot behind her and forcing her to sit up to move to a legal distance .. only to have the process repeated again and again. But then Mel was able to show her strong run, took the lead after the first of three run loops and won with a gap of more than 12 minutes. Texas was the third Regional Championship she raced in, and it was also the third she was able to win. Matt Hanson also has three Regional Championship wins as well, but while Matt’s were all in Texas, Mel’s wins were on three different courses: Melbourne 2015, Frankfurt 2016 and now Texas 2018. This shows that Mel can race well on almost any course. If her prep for Kona goes well, we’ll see her mix up the field in Kona – to me she’s a strong podium contender.
Andy Starykowicz Posts a 3:55 Bike Split
The “headline performance” in Texas that caught the most attention was Andy Starykowicz’s blazing fast 3:54:59 bike split – the first sub-4 bike ride in any Ironman race. He was probably the athlete that had the least benefit from the issues discussed in the section above, after all there is no draft at the front of the race. Andy exited the swim with the front group and quickly rode away from the rest of the field. In T2 he had a lead of more than 14 minutes to the chase group. His bike time was more than eight minutes quicker than Johannes Ackerman’s (who tried to go with him but had to let go after the half-way mark) and more than eleven minutes quicker than anyone else in the field. He was clearly riding a step above everyone else in the field. Regardless of whether you accept his bike split as a world record or not, sponsors should be happy to pay out the bonuses to Andy – it’s not his fault that the race had a few issues, and he delivered what he set out to do. After his amazing bike split, Andy also posted one of his best run splits – he ran a 3:00:57 (his best run split is a 2:58 from Florida 2013) and finished in eighth place, securing his Kona slot. He’ll be one more strong bike rider in the Kona field – and I’m sure that he’ll join forces with Cam Wurf to lower the 2018 Kona race even further than what we’ve seen last October. With their bike strength and the progress on the run both Andy and Cam are working for, they will be making Kona even more exciting.
Lauren Brandon Has the Fastest Swim Overall
In all the coverage of the Texas race, the female swim was only a side note. But Lauren Brandon managed something that is rarely seen: Her 48:19 was the fastest swim split overall! (The fastest male was Sean Donnelly with a 49:06, he DNF’d.) She also put more than five minutes into Meredith Kessler and eight minutes into the rest of the female field. With her swim she set a new course record (improving on her own time from last year) and also swam faster than she and Lucy Charles did in Hawaii last October (also a non-wetsuit swim). Comparing swim times across races is tricky because the conditions can vary quite a lot, but Lauren’s swim is one of the best swim performances we have ever seen by a female Pro. If the conditions are right, I’m sure she and Lucy Charles will attack the swim course record in Kona.
Great Returns From Injury by Will Clarke and Matt Russell
Some of the most satisfying results came from athletes that raced Texas after being seriously injured. Will Clarke was qualified for Kona but had to decline his slot after he crashed on the bike and hurt his shoulder. It has taken him some time to regain decent mobility in the shoulder again and his prep races at 70.3 Dubai and Challenge Roma did not quite go according to plan. But he was in great shape for Texas and a third-best run split allowed him to claim the final step on the podium. He’s now in a good position to qualify for Kona 2018.
Another athlete I was very happy to see race in Texas was Matt Russell. Matt was hit by a car during the race in Kona and almost got killed when he lost a lot of blood. After a few days in the hospital he needed some time to recover but by now he seems to have made a full recovery. He posted the third-best bike split making up the time he lost in the swim to the front group, but he wasn’t yet able to deliver his typical 2:50-ish run split in Texas. A 3:02 run saw him fall back to 14th place. It’ll be tough for him to qualify for Kona, but typically he’s been doing five to seven IMs per year – hopefully he’ll be able to continue to race often and well!