Metathon race predictor syncs your strava data to predict your marathon time

I have. First time I did it in October/November after being sick for about 3 months after a knee injury. I had done some jogging and easy bike riding and the result was 3:13 which was way off but because almost no training in previous 8 weeks, I can understand why they predicted like that but I wasn't in so bad shape. Now I checked it again and Metathon predicts 2:58. I haven't started marathon training blog yet and run no races whatsoever so far. Hence, I am okay with that result. A friend of mine who is plans to run a sub 2:40 marathon in spring is according to Metathon in 2:50 shape.

So based on his and my experiences I think that now in winter time when runners run some of their runs extremely slowly and don't race at all, the predictor is pessimistic but in my opinion it's better to be pessimistic than optimistic because it gives motivation for training. I am planning to check once a month where I am according Metathon and hope that also that predictor will start to show more promising results when the weather warms up.

2:59 for me. I ran 2:33 a few weeks ago.

This is uncanny, but the "yesterday" prediction is my EXACT marathon PR, set last March. Prediction is 7 seconds faster as of today, 4 minutes off my goal.

FWIW, this is week 4 of my Boston marathon training cycle. Prior to that, I had been doing a 5K cycle after coming back from gluteal tendinopathy, to keep my mileage a little lower and get some speed back. I'm assuming the prediction will drop after a few more weeks of marathon training.

So, seems relatively accurate for me. I'm hoping to be faster, but don't really know what to expect after coming back from injury.

Interesting. Mine says 3:25 while my PB is 3:30 (age 60) and I've been hoping for 3:40 this year. I have been running higher mileage lately so that must factor highly. Hope it's a good indicator for me!

It's fastest prediction for me from 8 weeks ago when I was in pretty good shape is still 20 minutes slower than my fastest marathon.

Currently predicting 3:03:40 for me. While I know I'm not currently in marathon PR shape (2:44 low), I could go out and jog that time any day of the week. I also nearly PR'd in an indoor 5k 2 weeks ago so it's not like I'm grossly out of shape.

I do tend to take my easy runs very easy and my workouts are on the more conservative side so this probably skews the data for me.

The numbers actually seemed really reasonable to me. They had me at 2:51 four weeks ago and slipping to 2:56 as of today. I got injured the last couple weeks and have had to severely reduce mileage and believe I was in about 2:50ish shape at the beginning of December. I'm impressed!

Yeah, probably a better model would be to throw out recovery/easy pace in terms of average training pace, at least for high mileage runners. As a guy hoping for sub-3, my "recovery" runs are often 8:45ish, sometimes even over 9:00 pace. Doing high mileage and running doubles, there are a lot of those that are going to drag my average pace up. My current 8-week average pace is 8:25.

Today it has me at 3:03 low after a relatively fast med-long run last night and an easy 6 this morning.

Study of 20 ?confirms the old adage, Miles make champions.

8 weeks ago it said my prediction was 3:24,9 weeks ago I ran 3:07 on Frankfurt,it punishes those who run slowly.

Here's the formula: Pm (sec/km) = 17.1 + 140.0 exp[-0.0053 K(km/week)] + 0.55 P (sec/km)

Pm = marathon pace; P = mean training pace, K = kilometers run per week.

I plugged my training in and it predicted 3:08, but I ran 2:59. Another year I ran 3:04 on the same training, and I paced it poorly, so not a good race. Miles are up this year, so maybe I'll do well under 3 at Houston.

I'm a 50s male, been running since high school.

20 is too small a sample size, and I bet stratifying the samples (male/female, age, etc) would improve the accuracy of the model.

Most responses state that metathon seems accurate or a little slower than actual results. My prediction is always around 3:22, but I think my goal of 3:40 this spring is a little aggressive. My pb was 3:30 three years ago.

The interesting thing is that I'm following the Daniels paces for a 3:40 target and those are the miles metathon is using in the calculation.

Any thoughts on the result I'm getting being that far off?

How hard could I crash? wrote:

Most responses state that metathon seems accurate or a little slower than actual results. My prediction is always around 3:22, but I think my goal of 3:40 this spring is a little aggressive. My pb was 3:30 three years ago.

The interesting thing is that I'm following the Daniels paces for a 3:40 target and those are the miles metathon is using in the calculation.

Any thoughts on the result I'm getting being that far off?

What makes you think that 3:40 is aggressive? Do you have any recent (shorter) race results that suggest you might be overreaching.

As for my metathon, it's hard to say as I haven't run a marathon near my current data. Though I'm running similar mileage and paces to what I was last year and was just over 3 hours. Metathon is giving me a 3:16 today.

So if it's giving you 20 minutes faster than your goal time, and most of us are being given slower times, I'm thinking your goal time is on the doable side.

Mine predicted a 3:27 marathon now, and 3:51 8 weeks ago. But I actually ran 3:03:23 at CIM 8 weeks ago. It listed 515 miles in last 8 weeks at 9:45 per mile. These were winter miles, some on snow and ice, some at 7,200 ft in Laramie. I am 60 years old. I run slow a lot of the time, often pulling my dog on a leash. But I usually run a few miles each week at sub 7:00. Last week I ran 13.1 @7:03 and 7 @6:55. The week before I ran 40 miles @9:59 but also a 5k cross country race in Yakima in 19:42. I also ran 2:58:28 in a marathon in October before I turned 60. I guess total miles and average pace doesn't tell the whole story. Running every mile at 9:45 is not the same as running some long runs at 8:30, others at 10:00, some jogs with the dog at 11:00 to 12:00, and a few miles at 6:40.

3:38 apparently

Not entirely true. I am aerobically weak, I ran 3:50 for my first marathon not too long ago. VDOT says it should be closer to 3:30 based on 5k.

Metathon now has me at a sub-2:59 as my mileage ramps up for Boston (in the midst of 100+ mile weeks). I would be ecstatic with that time. Guessing it might extrapolate to 2:55-ish by taper time, which is probably out of reach for me. My goal remains sub-3, and only if the weather cooperates.

20 is a small sample size, but the formula seems pretty reasonable to me. There will always be outliers.

People that are really fast, like 2:10, run a lot of mileage, but not very much consecutively at that 2:10 pace, 5 flat. They just touch that pace periodically within longer runs. I bet it would not predict this person to run very fast.

Yep same kind of prediction, predicting 3h16 for me and I ran... 2h44 in Frankfurt last October.