Boston Qualifying Times and Hills Revisited: Data from Boston 2025

Feature image by dorestbays on Flickr - CC BY-SA 2.0

Last month, I published an analysis of the impact of downhill marathons on the Boston Marathon qualifying process.

As a reminder, here are a few of the conclusions I came to:

1. The vast majority of races have a net drop of 5m or less. There are a small group of races with significantly steeper downhills.
2. When you compare runners’ qualifying times to their Boston times, there’s a much bigger difference for runners who qualified at downhill marathons.
3. Between 8% and 12% of qualifiers at the 2025 Boston Marathon earned their qualifying time at a steep downhill marathon.
4. Eliminating downhill marathons would have a small impact on the cutoff time – but some significant portion of runners could still qualify if they were forced to run a flat race.

That analysis was published before the 2025 Boston Marathon. I used the entry list for the 2025 Boston Marathon to determine how many runners qualified on downhill races. And then I looked at runners from the 2024 Boston Marathon to see how their times compared with their earlier qualifying times.

Some Reminders About Data and Methodology

First, here are some a few clarifications about data and methodology.

I’ve been collecting race results to track and project the Boston Marathon cutoff time. The full dataset for the 2025 Boston Marathon qualifying period includes approximately 550,000 finishers across 282 total races.

I categorized those races by net drop in elevation:

• Flat, <1m net drop, 231 races
• Minor, 1-5m net drop, 20 races (excluding Boston). Example: CIM
• Moderate, 5-10m net drop, 11 races. Examples: Colorado, Steamtown
• Steep, 10-25m net drop, 13 races. Example: St. George
• Very Steep, 25+ net drop, 5 races. Example: REVEL Big Cottonwood

Using name, age, and gender as matching criteria and some fuzzy matching, I matched qualifying times to entries at the Boston Marathon. Of 31,138 entrants, I matched 16,757 names to eligible qualifying times. That leaves ~7,250 time qualifiers unmatched – either because of data differences (possible due to name or age irregularities) or because they are not in my sample (more likely for international runners).

A little over 10% of the entrants either did not start or did not finish Boston. Just over 2,000 of the matched qualifiers did not start or finish the race, reducing the size of the group with both eligible qualifying times and finish times.

Due to the weather at last year’s Boston Marathon, I’ve excluded those qualifiers, and I’m focusing on runners at the other races. Taking all of those exclusions into account, there are about 11,000 finishers from the 2025 Boston Marathon matched to qualifying times who form the basis of this analysis.

What’s the Difference Between Qualifying Times and Boston Times?

Boston is a tough course. Although it’s net downhill, there is considerable elevation gain throughout the race – especially in the last six miles. As a result, it’s safe to assume that most runners will run slower at Boston than they will at their qualifying race.

Conversely, if someone runs a downhill marathon to qualify, they’ll reach a qualifying time that is faster than they otherwise could achieve.

If you operate under the assumption that steep downhill races should result in faster times, the corollary to that assumption is that runners who qualify at steep downhill races should see larger differences between their qualifying times and their Boston finish times.

The visual below breaks runners out by gender, age (under 40, over 40), and the elevation profile of their qualifying race. The dot represents the median time difference between their qualifying times and their Boston finish times.

The dropdown toggles between men and women.

No matter what group you look at, there is a consistent trend. Runners at flat and minor downhill races have the smallest differences, while runners at steep and very steep downhill races have the largest differences.

Within that trend, there’s quite a bit of variation by age and gender. Within each group, younger runners have a smaller difference than older runners. In most cases, that delta is about five minutes.

There’s not a ton of difference between men and women.

For younger runners, the base difference between a flat race and Boston is around 3 to 4 minutes. That rises to 6-7 minutes for moderate downhill races, 8 minutes for steep downhill races, and 15-18 minutes for very steep races.

For older runners, the base difference is 8-10 minutes. It’s only slightly higher for moderate downhill races (~11-12) and steep races (12-14). But the difference at very steep races jumps to around 22 minutes.

Notably, the percentage of runners who are over 40 also increases with elevation. At very steep races, there are more than twice as many runners over 40 – a far greater share than any other category.

But Is Everybody That Much Slower?

There’s a danger in assuming that that median difference applies to all runners. That leads to the conclusion that all runners from downhill races are necessarily much slower at Boston, but that’s not the case.

The visual below takes this group of runners and plots the share of each group according to how much faster or slower their Boston time was.

A negative number means they were faster at Boston, while a positive number means they were slower at Boston. The x-axis groups things by 2 minute increments – so 0 means runners who were 0 to 2 minutes slower at Boston.

You can click on a category in the legend to hide it and focus on one or two groups.

For now, just focus on the flat, steep, and very steep categories.

The flat category is the largest, so the line is the smoothest. But if you ignore the ups and downs in the other two lines, they have a pretty clear pattern, as well.

The lines for the steeper downhill races are both shifted to the right. This is an indication that, on average, runners are faster on downhill races – and then lose a greater amount of time when they run Boston.

But find the zero on the x-axis and look to the left. Everyone to the left of the zero ended up running faster at Boston than at their qualifying race.

Maybe they had a great training block or the weather was bad at their qualifying race. Maybe those runners saw a smaller benefit from the downhill race – or were more suited to tackling the hills at Boston.

The point is, there’s some variation – and the median doesn’t tell the whole story.

How Likely Are Runners to Requalify?

One way to gauge whether runners would have qualified anyway is to look at how they performed at the 2025 Boston Marathon – and whether they met the qualifying time they needed to get in.

Some people may have run a downhill race and achieved a very large buffer. And they ran a little slower at Boston. But since they were so fast, they still ran a qualifying time at Boston … a clear sign that they could have qualified on any other typical course.

In this case, I’ve calculated these qualifying times using the 6:51 buffer that runners had to meet to gain entry into the 2025 Boston Marathon.

As you go from flatter courses to steeper courses, the percent of runners requalifying goes down. That makes sense if you assume that runners qualifying through downhill marathons ran faster than they otherwise could have run.

If you look at runners qualifying through flat races, a little over half of them met their qualifying times again at Boston. So it’s normal for some number of runners not to make it.

Some of them may have been on the cusp, and the tougher course at Boston pushed them over the edge. And some of them may have just had a crumby training cycle, and they weren’t prepared for Boston.

If you look at runners qualifying through very steep downhill races, the percent of runners not requalifying jumps to 78%. So again, this supports the conclusion that runners from downhill races are more likely to have gotten a little help qualifying.

But the flip side of that bar is that 22% of runners qualifying at very steep races – along with 40% of runners at steep races – met their qualifying times at Boston. It’s very unlikely that these runners needed help from their steep downhill qualifying race to get into Boston – and very likely that they would have gotten in no matter what course they ran.

If you put all that together, let’s assume that:

1. A small number of runners who didn’t requalify at Boston could have qualified on a flat course
2. All of the runners who requalified at Boston could have qualified on a flat course

That means that of very steep downhill qualifiers, slightly more than 22% could still have qualified. And of steep downhill qualifiers, slightly more than 41% could have qualified.

In the previous analysis, I estimated that somewhere around 40% of downhill qualifiers could have gotten in on a different course. I think that’s still in the right ballpark – and this way of looking at it supports the idea that somewhere around 30-40% of downhill qualifiers would still be able to make it into Boston without the assistance of a downhill course.

What If We Modify Their Qualifying Times?

Here’s one final way of looking at things.

Let’s assume there’s a standard modifier, such that a downhill course improves your time by X. We can then add X back into the runner’s qualifying time and see if they still would have qualified.

I took the data from the previous graph – showing the delta between runners’ qualifying times and their Boston times – and estimated a set of modifiers.

For runners under 40, I added 2.5 minutes for moderate downhill races, 5 minutes for steep downhill races, and 15 minutes for very steep downhill races.

For runners 40 and up, I added 2.5 minutes for moderate downhill races, 5 minutes for steep downhill races, and 10 minutes for very steep downhill races.

Then, I recalculated how many runners would still have qualified – verse how many runners now missed their qualifying times.

• For moderate downhill races, 76% still qualified and 24% did not.
• For steep downhill races, 57% still qualified and 43% did not.
• For very steep races, 31% still qualified and 69% did not.
• If you combine steep and very steep downhill races, 44% still qualified.

You could quibble over the actual modifiers. To show an extreme counter example, let’s use 10 minutes as the modifier for steep downhill races and 20 minutes for very steep races for runners of all ages:

• 30% of runners at steep races still qualified.
• 13% of runners at very steep race still qualified.
• Combined, 22% still qualified.

The exact figure is going to vary. But I think this supports the conclusion that some significant number of downhill qualifiers would still make it into Boston – even if they had to qualify on a flat course.

Based on these numbers, I think somewhere in the ballpark of 30-35% seems like a fair estimate – with a broader possible range of 25-40%.

So What Did We Learn?

To a large extent, this confirms what I found in the previous analysis.

When you compare runner’s results from the 2024 Boston Marathon to their qualifying times, runners that qualified on downhill races lost more time at Boston. The runner’s at the 2025 Boston Marathons showed a similar pattern.

The difference between a flat race and a minor downhill race – like CIM – is negligible. The difference is a couple of minutes for moderately downhill races (like Colorado or Steamtown), around 5 minutes for steep downhill races (like St. George), and 10-15 minutes for very steep races (most REVEL races).

Despite that boost, some significant number of runners have such large buffers that they could likely have qualified on any other course. Many of the runners who qualified on steep downhill courses still beat their own qualifying times at Boston. And if you apply a modifier to their qualifying times, many of those runners still would have qualified.

Circling back to the original question of how much this actually impacts the cutoff time, let’s combine this with the conclusions of the previous analysis.

There are about 24,000 time qualifiers, and roughly 10% (2,400) of them achieved qualifying times on steep downhill races (the high end of the estimate). At least 30% of those qualifiers (720) would still make it in if they were forced to run another qualifying race, while up to 70% (1,680) of them would be removed from the qualified applicant pool.