Are Ironman Races Close to Kona Easier to Qualify At?
Facebook
X
LinkedIn
Threads
Reddit
EmailI’ve been asked this question a lot. I don’t really like the assumptions behind it: that all the fastest athletes are in Kona and not racing at other events. I get the logic, but what’s the evidence? My position has always been that it doesn’t have much impact. Or, at least that I’ve not seen anything to convince me it was the case.
The argument cropped up again during the debate over the new qualification methodology. Ironman itself has used it as a reason why the initial system wasn’t delivering as many slots for women as expected. Honestly, I felt this missed the point. If a third of your races have fewer women qualifying (because the fastest are in Kona) the remainder will need significantly more to bring the numbers up. (I did the maths: at the point of change women had around 20% of slots, to get to 30% they would need to be taking 35-40% of slots at every other race. Quite a change.)
There isn’t a monolithic, unchanging group of top age groupers – athletes come and go. Many – but not all – of the fastest age group athletes race in Kona.
And then it occurred to me that looking at how athletes perform in Kona and when they raced during the qualifying season might give an insight into the strength of qualifiers. I’m not sure why I’ve never thought of that before. Over the last 3 days I’ve been plotting graphs and examining data. I’ve taken tangents I felt necessary to support my case. I went very deep. I was going to put it all in this article, but there is so much it’s only reasonable to break it up into digestible chunks. At least I’ll start with my conclusions to spare those of you who don’t like wading through endless charts.
Conclusion
No*, if you’re a man. Also, no+, if you’re a woman.
* Although they’re a bit more likely to encounter someone who makes top 10 in their age group between March and May. Otherwise. there are enough men to ensure fast athletes are available even when Kona is happening.
+ There may be a small difference for women, likely due to lower number of athletes in all. I’m not convinced it’s enough to affect their race selection.
If you’re happy to take my word on this, you can stop here. Otherwise, buckle-up we’ve got a lot of data to review.
Kona Times and Placings
The basis for believing early season qualifiers are easier is the idea that the fastest athletes aren’t there. If the fastest athletes aren’t there, does it follow that the Kona finishing times from qualifiers at these races aren’t as fast as those from later races? Seeing a difference in the pattern of finishing times might be an indication in a difference in strength of qualifiers.
The chart above plots the distribution of age group male finishing times in Kona for 3 different groups. Those who raced early-season (late August to December), mid-season (March to May) and late-season (June to early August). Along with the distribution there are two sets of lines indicating the top 5% and the median for each group. In this instance I’ve considered Kona between 2015 and 2019 for consistency. Recent years with two championship locations and the impact of Covid make the data less clear.
All 3 groups show remarkably similar distributions with their top 5% and median lines tightly bunched together. There’s no significant difference in performance for these qualifying groups. Although, we might observe that the red March to May group have a slight edge. This certainly wouldn’t lead me to conclude the men who qualified early season were weaker than those that came later.
Notes on the data: the Ironman World Champs results are not included for any athletes in the September-December data even if they raced there. Only Ironman races are considered in this dataset. Athletes who raced multiple Ironmans in the season will be included in multiple groups.
The picture is different for women. The separation between groups is clearer, although no more than 15 minutes at the median and 10 minutes at the top 5% points. It’s enough to suggest that the qualifiers from the early-season races aren’t performing as well at Kona, although it doesn’t necessarily indicate why. That uncertainty, plus my own bias in believing there’s no real difference lead me to dig further.
Rather than looking at the distribution of results, scattering the results gives a visual representation of the very front of these races. Each point represents a female finishing time at Kona between 2015-2019 plotted against the date of a race they competed in during the qualifying season. Segregating this by Age Group would probably make some sense, but there are limits to how long I can make this article.
This graph includes results from any athletes who completed the World Champs during the qualifying season (in red). When we look at the lower half of the graph we can reasonably assume that most of those red diamonds are not blue crosses in that region. With women making up a smaller proportion of total Ironman numbers it’s not unreasonable that some of the fastest women aren’t racing elsewhere in that period, but there’s not a huge difference in the scatter of those early races from those higher up in the chart.
I’ll leave myself a reminder here that I need to make an interactive version of this particular graph to see how it works at an individual age group level
I was so dissatisfied with that last chart that mid-article I had to make the above graph to clarify the point. Rather than considering finishing times, looking at finishing places gives a clearer picture of competitive potential and eliminates differences between age groups. March through May remains the time period with the highest ranking female age groupers in Kona. A 7 place difference at the median, but only a difference of 1 place for the top 5%. June to August is comparable with the early-season qualifiers.
None of this suggests a huge difference in performance and that variance in age groups is a factor in the time distribution charts. Racing in October won’t avoid you facing someone with the potential to podium in Kona.
We shouldn’t forget the men.
Many more men race Ironman, so the equivalent chart for age group men shows a much denser scatter plot. Again, we can expect that the majority of athletes in the red diamonds are not blue crosses in that period. But the greater numbers helps ensure that there are many other athletes equally capable of hitting the same times. As I realised with the women’s chart, this isn’t the most useful way to view the data, at best it gives a sense of how distributed top performers are across the race season.
So, we come back to a distribution chart. With a large field the distributions become a bit more consistent. There’s only a 4 place difference in the median placing for each group with September-December coming last. At the 5th percentile there is very little difference, 2 places between early and mid-season and no difference between early and late season. We can clearly see the bump in the red group for the top ten places otherwise they are all much alike.
Let’s note here that for both men and women March through to May seems to produce a good number of the fastest results in Kona. I suspect this is those who raced Kona recovering and then qualifying for the next year as early as they can. As I’ll show in a moment racing Kona back-to-back like that is actually quite race, less than 20% of men and women race Kona in consecutive years. Although, for those who make the top 10 in their age group that number jumps closer to 40%.
Competitor Numbers
The numbers of athletes involved in Ironman was another area of consideration. For Kona to be influencing the competitiveness of other races it needs to have taken in sufficient fast athletes that there aren’t enough to be entering other events.
In the period I’ve focussed on we quite consistently have around 3.5% of the Male competitors in Kona and a touch over 5.5% of the female competitors there. Or consider it a pool of 45,000 men and 10,500 women who aren’t in Kona at that time. Has Kona cleanly snapped up the top 3.5% and top 5.5% of men and women respectively? It’s unlikely.
As I’ve already mentioned over 80% of men and women in Kona weren’t there the previous year. Or, based on the numbers in the table above around 300 men and 100 women in Kona were there the year before. You’re unlikely to have encountered those individuals at an early season qualifier, but are they the strongest athletes?
When we focus on those who manage to make the top 10 in their age group we see a shift in the numbers. Top 10 age group athletes raced the World Champs in the previous year about 35% of the time. The very fastest are more likely to race again and from our earlier charts are probably gaining their qualification in March to May if they can. But still, the majority of the age group top 10 were not in Kona the year before so could have easily qualified early season.
Just considering previous World Championship experience shows around 60% of athletes were first timers in Kona in any given year. Numbers rapidly decline beyond that point. The competitive-span of a Kona competitor is relatively short, a small percentage keep coming back. New Kona athletes are an important part of the makeup of the race.
Again focussing on those who make the top 10 in their Age Group in Kona we see slightly more longevity. Almost a third are new to the World Champs, that’s a good rate of change over. Almost a quarter have only raced once before. The fastest age groupers are a more stable group than Kona athletes as a whole.
But their numbers are small as a proportion of unique athletes racing Ironman in a qualification year. Not enough in my opinion to truly diminish the potential for faster athletes to be racing at early season events. At the very least there’s sufficient turnover to expect new names to be coming into the field.
What About Now?
What if we apply the same principles to more recent times? There are going to be issues. Covid had a huge impact on early numbers and then we alternated between two different Championship locations. I’ve only considered those racing in Kona for this analysis. Our trust in these numbers should consider the many other factors that influence athlete choice in these times.
In this recent period there’s a shift in male qualification with a clear gap between those racing early and those racing late in season. The median is fastest for the late season qualifiers, but as before the mid-season dominates the top 5%. Along with changes of locations and delays in championship races there was a huge increase in numbers with the gender split championships. A likely reason for the large difference in the middle of the field is the increased qualifying numbers in these years. At the front of pack we see no difference between those who raced early in the season or those who raced late.
As championship numbers are significantly reducing with the return to a joint race I would expect the time gap to tighten at the median.
If we check how placings have distributed based on prior race months the mid-season has a couple of places advantage at the front and the bigger field takes the medians outside the bounds of this chart. Were I choosing when to race, on this evidence I’d probably avoid March to May expecting the fastest athletes then.
But, does the larger championship field, varying timing and race location, and delayed racing from Covid allow us to trust these results in full? I’ll need a few more years of data to really be sure. With lower numbers and more stability we may see a return to 2015-2019 behaviour. The new allocation system could also reshape how qualification numbers look and raises the question of whether I should segregate on age group as well.
We have a similar pattern for women in this time range. June to August racers are fastest at the median, but not at the front 5%. September to December produces the slowest group of athletes, but the time range is slightly smaller, especially when compared with March to May. Greater numbers of women racing in a championship gives a smoother dataset. And, it should be noted, also reduces the skew of the finisher distribution in Kona.
Considering the distribution of finisher placings for women we again lose the median from the chart due to the high numbers, but like the men the top 5% are drawn most heavily from the March to May period with little difference between early and late season racers.
The very fastest tend to race March to May, by a small margin at least. In a system with very few slots for women that may well mean there are advantages to racing at other periods of the year. Not enough to account for a large discrepancy in the proportion qualifying over a season, but if only an age group winner gets a slot you’d be at risk in March to May.
Summary
It’s been a journey. I’ve gone from believing there was no advantage to qualifying while Kona was happening to thinking there might be a small advantage for women, to thinking it’s so small that it’s probably not worth considering. While there was evidence that women’s times were slower in Kona for those who race in the early qualifying season, and more of the top 10 at Kona come from the March to May period, it’s not a huge difference. Men showed even less difference in times for the three groups, placings followed a similar trend.
More recent data could be taken to indicate a more significant difference, and in particular to avoid March to May as the most competitive period. But it comes with large caveats. A lot more variance in those years and factors that won’t apply to future racing in Kona. I’d not place too much weighting on these results.
Ultimately, man or woman, I’m not sure I’d factor any of this into my thinking. Nor, were I planning an allocation system for slots, would I expect it to have a huge impact. Maybe a small uplift in slots during March to May, but back to the same from June onwards.
There’s a lot I still haven’t shown you. More analysis on athlete numbers and history, including how performance relates to past Ironman racing. I’ve many graphs to share, but if you’ve made it here you’ve seen enough.
Finally, while I don’t think there’s any significant qualification advantage to racing early-season, one advantage to qualifying early is you get a full season to prepare for Kona.
