KQ 2026 performance pool qualifying

[Lurker4]

When all is said and done, this rollout should stand as a case study of how not to use statistics to design a product & marketing campaign.

[devashish_paul]

I wonder if you go outside of the top 20 percent of middle age men in Kona if the rest of us are closer to the 20 percent, than the rest of 20-39 year olds outside of the top 20 percent of young men in Kona. If that’s the case, it helps the local fast and fit guy a bit more at the local IM.

What if they took 100 percent of the Kona field, what do the coefficients look like? Because then you have the Kona pack fodder in every age group, and these pack fodder guys in Kona are the local heros who are largely the ones getting the tailwind of the coefficient and getting the KQ slots.

If you generate the coefficients off the most competitive at Kona but then you apply it to the next tier, of humans, maybe its not as applicable. It’s just a hypothesis based on what is playing out (having not gone thru the entire year cycle yet…maybe we get more young people from the European IM qual cycle next spring/summer).

[Sander]

One thing I don’t quite get here (or at least I’m not sure there is a consensus), is what we’re actually using finish times and rankings as a proxy for?

Would the ideal theoretical final ranking for slots be by:

1. Actual vs theoretical max for your age and gender? (So luck + genetics + how hard you trained)

2. Actual vs theoretical max relative to age and gender with adjustments for the fact that life gets in the way as you get older and women face greater societal barriers? (So luck + genetics + how hard you trained + societal adjustment)

3. something else?

like what’s the actual goal here? we seem to have already decided, whether we know it or not, what a fair outcome looks like, and now we’re all trying to work backwards to justify it. It might make more sense to do it the other way around.

[deantrives]

i thought the goal was simple, create a metric to distribute slots beyond one per age group. but with all the talk of who is getting those distributed slots it seems my understanding is stupid.

[Sander]

Makes sense, so we could just add a cap and floor to each age groups slots? Either absolute or relative to entries in that age group. So keep the ‘performance’ aspect, but say no group gets less than 3 slots, and no group gets more than 10 slots. Just a little tweak but along the same lines as what they’ve already got so should be doable without too much of a fuss (very naive statement I realise)

[Tri78910]

Floor doesn’t really make sense as it cant be over 1.

A floor of 2 would mean every race would have to have at least 52 slots - which is more slots than 2/3 of the qualifying races. so even if you increase the slots in each race this would be effectively just give slots to the top 2 in in each AG

@devashish_paul yes right now there isnt a huge gap which is why i dont think drastic action is needed. lets see how it balances out after the kona racers return. also this will be 100% different next year when the top men and women are all in kona during this period

[Tri78910]

it seems like course is making a difference here

in Cali (flat) the top AG time was 20 min faster than the top under 40 AG time
Maryland (also Flat) it was 23 min fast

in Wisconsin it was only 3 min faster

Chatt top qualifier was M30-34

[monty]

@Tri78910

it seems like course is making a difference here

You are mistaking one metric for another. What matters is who shows up. All those times you posted were done by racers, not the course. If the 8 or so guys in the 50 to 59 AG that swamped the spots didnt race that day, completely different qualifying outcome. Has nothing to do with he course.

Yes some courses will be tougher than others, some swims cancelled or shortened, but at the end of the day it just matters who shows up to race on the day. Thus far it has been men top heavy, while the top women had other fish to fry in the actual Kona race everyone is so amped up about..

[deantrives]

that is the biggest thing (and to me always has been the biggest thing); my question is for the folks who recently finished second and are quite upset, where would have they placed had Kona not been the week or so before hand… maybe they would not have been so upset… we need to let this system breath.

[as9934]

I appreciate the efforts of the folks thus far trying to bring some data to this.

I went ahead and scraped all of the data for all of the participants in the 2026 Kona qualifying races that have happened so far (did not include 70.3 Hawaii).

HERE IS THE COMPLETE AGGREGATE DATA FOR EACH RACE AND AG:

Kona Slots Summary

NOTE – This is just the people who were OFFERED slots, not who took them. Please let me know if you see any errors and I will do my best to fix them.

Please also send me ideas of new KQ systems to investigate. So far in this and the other thread I have seen:

1. Re-weight performance pool slots on a for each race (rather than comparing to Kona times — working on this one now)
2. Separate women’s and men’s performance pool + podium slots like with 70.3 worlds (AKA equal representation)
3. Top 3 in each age group get offered a slot
4. Time cut-off of some kind (would need a clearer way to do this)
5. Distributing number of slots equally across all races
6. Distributing number of slots proportionally based on number of entrants for that race

Idk who the @SlowtwitchSupport mods are here (@Bryancd ??) but I’d be happy to write up my findings from these in a full post.

(My bias – *In my day job I work as a data journalist for one of the bigger news outlets here in the U.S. I’m a M25-29 Ironman and I’d love to go to Kona someday, but I’m currently about 4 hours too slow and my chances of qualifying in 30-39 about about 0.)

(Note: I noticed an error with the gender percentages that I went ahead and fixed)*

[Bryancd]

@as9934

Idk who the @SlowtwitchSupport mods are here (@Bryancd ??) but I’d be happy to write up my findings from these in a full post.

Go for it! This thread is sort of a in perpetuity rolling analyses as races drop. You can have a separate thread with just your data.

[jwmott]

Thanks for sharing. In case you didn’t see, someone else on this thread posted a similar spreadsheet. The main comparison they’re doing is versus the old allocation system. Maybe you can work together.

Mods, is there any way to pin these posts in the thread so people can easily reference the spreadsheet links?

Anyway, I think other system tweaks people have mentioned are:

• Using some other Kona results metric to derive the coefficients (average of top 3, top 3-20, top 5%, top 50% versus the current top 20%)
• Using data from all Ironman races to derive the coefficients (either using the top 20% or some other metric)
• Excluding legacy entrants from the results used to calculate the coefficients

Another thing I think would be interesting is a larger analysis of how this current system would have played out in past years (probably even pre 2020). Someone has done some partial analysis of the past few years for the fall races earlier in the thread.

A long wish list of some not so easy analysis, but you asked…

[SharonMcN]

Data journalists are the best!!

[timbasile]

Cross posting this here as well. Using IMCA, I show that the issue with the older men getting slots has more to do with the fact that the depth of field in M55 is considerably deeper than elsewhere and that this is driving the slot allocations. I do think that the coefficients need tweaking (XC, Legacy, Kona basis), but in this case M55 was considerably deeper than M30, and the slots bear that out.

If anything the person on the outs (at least as I saw) was 2nd place in F30.

An Alternative Way Forward for Kona Slots

Coming back to this, I figured I’d take a look to see if this hypothesis is true. And it is - The changes in demand that we’re seeing also impact age, not just gender. I believe that the IM system does have a bias with respect to older men’s categories due to the Kona course, Legacy and XC slots, but I hope to show below that most of the issue for M50+ actually has to do with who is showing up. I took a look at IMCA, and highlighted 4 age groups: M30, F30, M55 and F55, over the last 4 years.…

[as9934]

There were 163 M55-59 at IM CA or ~7% of total participants who finished, 8.7% of the men who participated.

Kona Slots Summary

(Note: I noticed an error with the gender percentages that I went ahead and fixed)

[Tri78910]

At quick glance something does not seem right here - I just know that i was the final member of the M30-34 AG to get a offered a slot at IM WI and I was 7th in that AG (you have 9)

Also the total number of slots for that race was 40, but you have 47.

For cali you have 64 slots but there was only 55

feel free to cross reference my sheet posted earlier in the thread

[as9934]

I have 55 slots for CA and and 40 slots for Wisconsin.

I fixed another error earlier today but that shouldn’t have affected the slot number.

9 represents the number of people in that age group that would have been offered slots based on their time. If you were 7th in your AG, then your AG ranking wouldn’t matter at all — it would go to the performance pool. 2 others slower than you in the AG may have been eligible to claim one of those slots but did not show up/already had one from a previous race.

I’ll double check on this again though.

[Tri78910]

Maybe I’m reading it wrong but when you sum column c it’s 47

I can confirm at WI no one lower than 7 in my age group was eligible in the performance piok

[Tri78910]

finally look at whats happening in more detail

It looks like you are grabbing the top X (55, 40, etc) in the performance pool and assuming they would all be offered the slot. This is not the case - If one of the AG winners is outside the top X then it reduces how deep the pool goes.

In the case of Wisconsin and California there were AG winner that didn’t crack the top 40 or the top 55 respectively in their performance pool ranks. So the pools only offered slots to the people in the top 32 and 46 in the performance pool respectively.

[as9934]

@Tri78910

It looks like you are grabbing the top X (55, 40, etc) in the performance pool and assuming they would all be offered the slot. This is not the case - If one of the AG winners is outside the top X then it reduces how deep the pool goes.

Wait I’m confused — my understanding of the system was that:

1. Every AG winner gets offered a slot not matter their time

2. After every AG winner gets offered a slot they go to the performance pool for the remaining slots

   This appears to be what Ironman says on their website:

   

So I did this:

df['adjusted_time'] = df['Finish Time'] * df['Factor']
df['auto_kq'] = False

#if you win your age group you get a slot
df.loc[df['Age Group Ranking'] == 1, 'auto_kq'] = True
...
#rank people on their adjusted times for each event
df['rank'] = df.groupby('EVENT')['adjusted_time'].rank(method='first').astype(int)

#if you don't already have a slot (auto_kq == False) and your adjusted time rank is below the total number of slots, you get offered a slot
df.loc[((df['auto_kq'] == False) & (df['rank'] <= df['SLOTS ALLOCATED'])), 'auto_kq'] = True

Is this incorrect?