Note: I’m in the process of updating this page. If you still see this message at the top, check back in a day or two for a full explanation. But the embedded calculator below is new.
Use the calculator below to compare the marathon performance of one or more runners. Simply enter their gender, age, and time and click “Analyze Performance.”
The results will include:
- their actual time
- their age graded time
- their performance level percentile
- their percentile in the overall distribution of finish times
Continue reading below for some additional context.
Note: Note that the results for runners over 80 are very limited.
How Is This Different From a Regular Age Grade Calculator?
Traditionally, the age grading system has been used to calculate an age graded time for the purpose of comparisons. This system uses an age factor to predict how much time you would slow down – compared to a younger runner.
It is an imperfect system, but in many ways it’s the best one we have. I wrote more about this in RUN Outside.
This system yields two numbers: an age graded time and a performance level percentile. The age graded time is what’s considered an equivalent time – if you were a younger runner in your late 20’s or early 30’s. The performance level percentile is your time, expressed as a percent of the age standard – the theoretical equivalent of the world record for your age group.
One of the problems with age grading is that for a lot of runners, big differences in times can equate to small differences in their performance level percentile. Even faster than average performances can end up equating to mediocre PLP’s.
Another way to contextualize your finish time is to compare it against every other runner in your age group – and see what percentage you’ve beaten out. This is the value represented by the “Percentile” in the table above.
The graph visualizes this by showing you the entire distribution of finish times for your age group and then shading in the group of slower runners that you beat out.
Methodology and Context
I maintain a large dataset of marathon results in order to project the likely cutoff time for the Boston Marathon. This dataset comes in handy for other kinds of projects – like this one.
The dataset includes results from every race in the United States or Canada with 200 or more finishers. Or at least, I try to. Occasionally, I miss a race – and if someone points it out I add it to my list of races to track. The full dataset also includes the full set of Majors, but the purposes of this analysis I’ve excluded all international races. Each year, this includes ~500,000 to 600,000 results.
The results have been narrowed down to finish times between 2:00:00 and 8:00:00 to exclude outliers and potential data integrity issues. Two different datasets are available for use – one based on results from 2024 and one based on results from 2025. Early next year, I’ll update this calculator with new data from 2026.
Actual age grading also varies by year. Every five years, World Masters Athletics creates a new version of their age factor tables with updated world records and age standards. If you use the 2025 data, then the age grades will be based on the 2025 age factors. If you use the 2024 data, then the age grades will be calculated with the previous (2020) age factors.
Runners are divided into five year age groups, with the exception of runners under 20 and runners 80 and above. These age groups are used to group results for the percentiles and for the distributions. For the actual age grades, an individual runner’s actual age is used – since the age factors are adjusted for exact ages. Note that the older age groups – especially 80+ for men and 75+ for women – have very few results and are not really reliable distributions.
The graph at the bottom shows the full distribution of finish times for the age group of each runner. So if Runner #1 is a 42 year old man, you’ll see the age distribution for men 40-44. For these distributions, the visible data is restricted to 6:00:00 – but the actual calculation of the number of results includes those who finished from 6:00 to 8:00. If you hover over a given point, you’ll see the percent of runners that are slower that the given point. The data is grouped into one minute increments.
Nice calculator, but it doesn’t seem to accept ages greater than 79. In particular I get no results for ages 80 and &1.
Good catch, Tony.
The dataset I was working with didn’t have enough data on runners 80+ to reliably calculate percentiles and z-scores for them. So it can’t return those results.
I can still calculate the traditional age grading values – I just need to fix the calculator to fail more gracefully. Once I update things with the 2023 data, I’ll also see if I can add percentiles and z-scores for the 80-84 age group
Great insights! Your age-grade calculator offers a fresh take on comparing marathon performances using age-grading, percentiles, and z-scores. The alternative methods add depth, especially for non-elite runners.
Additionally, the z-score analysis highlights potential biases, offering room for further refinement. This balance of methodologies makes your calculator a versatile option for runners seeking a more comprehensive performance analysis.
I am a male, 66 with a recent marathon of 3:45:17. My age-graded percentile is 68.59, my percentile is 92.95 and my z score is -1.26. I assume a minus z score 1.26 standard deviations faster than normal. Is that correct?