The research is mixed and I don't believe there is a consensus on the methodology of drinking fluids during a marathon. One of the most common suggestions is to drink to thirst. Ultimately, I think it's a mix of do what you did in training, learn what your individual needs are, and then be willing to adjust on race day.
For anecdotal information, here is my consumption rate on an hourly basis for a Marathon (including carbs, electrolyte, and caffeine information):
15 min prior to start: 17oz of Maurten 320
Per hr consumption:
Water = 40oz/hr (estimated 3 oz per cup and 13 cups if racing)
Calories = 373/hr
Sodium = 646mg/hr
Carbs = 93g/hr
Potassium = 145mg/hr
Magnesium = 14mg/hr
Calcium = 40mg/hr
Caffiene = 66mg/hr
So in a 3 hour marathon, I estimated I drink about 137oz of water. That's about 1.07 gallons or 4.05 liters. I tend to be on the high side of things.
I definitely think there's a fine balance. Like
@camaker said, "Corral jumping" is always going to exist. How much cheating on POT occurs is an unknown, but I would venture to guess it's quite low. So it simply comes down to whether runDisney wants to collect more submissions of POT, but never really check them (what I had always assumed was the case pre-2022) or go with the idea of collecting a smaller number of POT submissions because of tougher standards, and then actually check them. For me, why not a little of both?
I have been an advocate that the POT standards be lessened instead of strengthened. I'd personally like to see them move the standards from a 2:45 to a 3:00 or 3:15 HM, instead of a 2:00 HM (or in the case of the M instead of 4hrs I'd do 6 hrs). In that case, you'll get more submissions. I think you say you'll check them, and if they're falsified, you go to the back. But in reality, you still only check the sub-2 submissions to ensure the first one or two corrals have a majority of appropriately placed runners. I believe the vast majority of runners won't submit falsified data when there's the threat of being put in the last corral. When it comes to an estimated drop-down method where there is no verification, then people are more willing to choose something that may be potentially faster than they'd ever be able to do. Because with the simple drop-down there is no information to prove otherwise. I personally don't see it as the same level of false information.
Now, the flip side is the new runners that don't yet have a POT for even a 3:00-3:15 HM. In a normal year, race registration is 8 months in advance of the race. It would seem to me that it's not unreasonable to do a 10k within that time period if you have a strong desire not to be in the last corral as a new runner. Certainly there will be those who are unable to do it within that time period, but I would think the vast majority of runners that want to, could. Same goes for a M. It's not an unreasonable expectation to do a 10 mile or HM before a M. Maybe even lessen the standard to a 10k if you felt strong enough (potentially with a massive hit on the race equivalency calculator).
There were two factors that would make the 2022 M course seem more congested than a normal year. First, they released the entire field in about half the time. So that means in essence the density of the runners doubled. At times it would feel like a field of 32,000 runners instead of a field of 16,000 runners. So for every open space that would have been there in a rate release over 60 min, you now have a person in that spot. So new POT standards or not, that's going to make things way way more congested.
Second, 2022 was a clear outlier in terms of distribution of runners across the corrals than the three previous races when the POT standard was 5:30.
View attachment 637895
From 2018-2020, about 42-45% had a POT of 5:30 or faster. That means 55-58% of the field did not have a POT, and where seeded in the last three corrals. In what was the 4:00-4:30 corral in 2018-2020 (Corral C) there was about 10-12% of the field.
Compare that to 2022. A total of 70% of the field submitted a 5:30 or faster (compared to the previous 42-45%). Only 30% of the field chose 5:30 or slower. In what was the 4:00-4:30 corral in 2018-2020 (Corral C), but was s3 for 2022 there was about 27% of the field. The 2022 numbers are rough and based on some assumptions, so don't take them as completely fact. But they're generally close to reality.
So the runners were overall released in a more dense manner (30 min release time) and there was clearly an influx of runners in the top corrals based on estimated finish times.
I think two things will actually drive it as an opposite effect. Similar to what we saw when comparing 2018 to 2019. You can see in the data that F, G, and H were reasonably even in 2018. But in 2019 and 2020 there was a huge swing towards submitting a 5:30 estimate. Presumably for more time on the course. Because moving from having 10% less of the field ahead of you means an extra six minutes of start time (roughly) versus the balloon ladies. You could gain as much as 10-15 min or so depending on placement within the corral.
So after people's experience in 2022, I think you'll see more runners say, "Well in 2022 I put the appropriate time of a 5:30, but then I was seeded in s6 behind 70% of the field when in 2018-2020 I was closer to the top 45% of the field if I started at the front of the corral. So I'm just going to fudge my estimate just like what seems like everyone else is doing." Just a hunch though based on past rD data.
Second, if a person who was suppose to have a 30 min buffer being the 50th percentile starter, goes to a 15 min buffer as the 50th percentile starter, I don't think they'll be inclined to put a more accurate estimate and put an even slower submission. That would mean they'd now have a smaller buffer than the 15 min they saw in 2022. They're more likely to put an even faster submission to increase their buffer time.