Riegel vs VDOT vs "double it and add": which race predictor should you trust?

There are three ways runners predict race times, and people get weirdly tribal about them. There is Riegel's formula, the power law behind most online calculators including this one. There is Jack Daniels' VDOT system, which converts a race into a fitness score and back out into equivalent times. And there is the folk arithmetic passed around at club nights: double your 5K and add a minute for a 10K, double your half and add ten for the marathon.

I fed all three the same runner to see how far apart they actually land. The answer surprised me, and it settled how I use each one.

The test: one runner, three methods

Our runner has just raced a 45:00 10K, a genuine effort on a fair course. That is 7:15 per mile, a solid club-level run scoring a VDOT of 45.3.

Predicting the half marathon:

• Riegel says 1:39:17.
• VDOT equivalence says 1:39:34.
• The clubhouse has no half rule, so it abstains.

Seventeen seconds apart, over 21 kilometres. Predicting down to the 5K, Riegel says 21:35 and VDOT says 21:40. Five seconds. And the marathon, the distance everyone actually argues about:

• Riegel: 3:27:01.
• VDOT: 3:26:55.
• Double the predicted half and add ten: 3:28:34.

The two formal methods landed six seconds apart on a three and a half hour race. The pub rule was within 95 seconds of both. I have seen bigger disagreements between two GPS watches on the same wrist.

This is the part nobody tells you: for a typical club runner in the middle of the pack, the formulas basically agree. Riegel and Daniels were fitting curves to the same underlying reality, and the folk rules survived decades of clubhouse scrutiny because they approximate the same curves. Arguing Riegel versus VDOT for your half prediction is arguing about seconds while your training decides the minutes.

So when do they split?

At the edges. The agreement above holds nicely through the middle of the pack, roughly VDOT 38 to 55, and frays outside it.

The folk rules fray first because they add fixed amounts. "Double your 5K and add a minute" flatters slower runners: doubling a 21:40 5K and adding sixty gives 44:20 against a true 45:00-ish, a bit keen; do the same from a 30:00 5K and the rule misses by more than two minutes. The fixed minute was calibrated on the fast club runners who invented it. If you are further back, make it "add two to three minutes" and the rule works again. Same story with the marathon: for a 1:20 half, doubling and adding ten is close to fair, but an elite running 1:05 would double-and-add to 2:20 when the formulas say 2:15, and a 2:30 half runner gets 5:10 when the honest number is nearer 5:13 and probably worse on the day.

Riegel and VDOT stay closer to each other, but they diverge in what they silently assume. Riegel has one moving part, the exponent 1.06, which is the average fade of trained runners across distance. VDOT assumes you are equally trained at every distance on its equivalence table. Both assumptions fail the same way for the same person: the speed-trained runner with thin endurance, whose real marathon sits well behind anything either method prints. No formula sees your long run history. I tested this against my own races and the marathon miss was consistently in one direction, and it was not the formula's arithmetic that was wrong.

What each one is actually for

Once you stop asking which predictor is correct, each one turns out to have a real job.

Riegel is the tool for race predictions. It is transparent, it takes one input, and its single knob is genuinely useful: if you know you fade over distance, nudge the exponent toward 1.08 and your long predictions stop lying to you; if you are a diesel who negative-splits everything, ease it toward 1.04. Two of your own past races will tell you which you are, which is something the predictor's advanced setting lets you do directly.

VDOT is the tool for training. Its killer feature was never the equivalent race times, it is the pace prescriptions: feed the VDOT calculator your 45:00 10K and it hands you easy days at 8:31 to 9:49 per mile, threshold work around 7:25 to 7:37, intervals at 6:41 to 6:51. A race prediction tells you where you might finish. VDOT tells you what to do on Tuesday, and that is the more valuable sentence.

The folk rules are the tool for conversations without a phone. They are calibrated for mid-pack club runners, and inside that population they are honestly fine. Just know whose pace they were built for, and pad them if that is not you.

The decision rule I actually use

Predicting a nearby distance, one step up or down from a recent race: use Riegel and trust it. The error at that range is one or two percent, smaller than the effect of a windy day.

Setting training paces for the next block: VDOT, no contest.

Predicting a marathon from a half: run both methods, note that they agree, and then treat that number as what you have to earn with your long runs, not what you have banked. If your longest run is 14 miles, the real answer is slower than all three methods, and no formula will tell you so.

Predicting a marathon from a 5K: do not. Predict your 10K instead, race it, and climb the ladder one honest rung at a time.

The formulas agree because they are all describing the same animal from different angles. The variable they cannot see, the one that actually decides your race, is whether you did the endurance work. That one is on you.