VO2 Max Race Time Predictor

The premise of a VO2 max race time predictor is straightforward: your aerobic capacity sets a ceiling on race performance, so knowing that number should let you predict finish times across distances. The reality is more nuanced, and most of the nuance comes down to one question: where did your VO2 max number come from?

The Source of Your VO2 Max Number Changes Everything

Garmin, Apple, Polar, and Coros all estimate VO2 max by comparing heart rate to running pace during outdoor runs and structured workouts. The algorithms differ, but they all introduce compounding uncertainty. GPS drift affects pace data. Optical heart rate sensors lag behind actual effort. Values shift based on heat, fatigue, and terrain. Research published in the International Journal of Environmental Research and Public Health found consumer watch VO2 max estimates can be off by 5-10% compared to lab values. For a runner with a true VO2 max of 50, that margin puts a marathon prediction anywhere between roughly 3:20 and 3:50.

Garmin’s race time predictor layers additional signals on top of the VO2 max estimate: training load history, heart rate variability trends, recent workout recovery. When training is consistent and heart rate data is clean, the predictions can be accurate. The problem is the system always looks backward. After illness, low-mileage weeks, or summer runs where heart rate spikes relative to pace, the predicted times lag actual fitness by weeks. That’s why discussions about how accurate the Garmin race time predictor is tend to split sharply in forums, with some runners calling it spot-on and others calling it useless, based almost entirely on whether their recent training has been stable.

Lab VO2 max testing is more precise but has a different limitation: it measures your oxygen-processing ceiling, not your running-specific performance. Two runners can share a lab VO2 max of 60 and finish a 5K 30 seconds apart, because running economy (the efficiency with which you convert oxygen into forward motion) varies between individuals. A raw VO2 max number fed into a predictor will overshoot finish times for runners with poor economy and undershoot for efficient movers.

If a lab VO2 max is all you have, you can still get useful predictions by finding the race time at each distance that corresponds to that value. Just expect the results to run slightly optimistic for most recreational runners.

How VDOT Solves This

Dr. Jack Daniels and Jimmy Gilbert developed a formula in the 1970s that sidesteps the economy problem by working backward from performance. Given a time over a known distance, the formula derives the runner’s “effective” oxygen uptake. Daniels called this VDOT. A runner who finishes a 5K in 18:30 has a VDOT of roughly 58, regardless of what a lab test might show, because the race result already encodes both aerobic capacity and running economy into one number.

That VDOT then maps to equivalent performances at other distances using the same physiological model. Because you’re anchoring to an actual performance rather than a sensor estimate, the predictions are grounded in something real.

The Daniels-Gilbert formula is:

VO2 = (-4.60 + 0.182258 × S + 0.000104 × S²) / (0.8 + 0.1894393 × e^(-0.012778 × T) + 0.2989558 × e^(-0.1932605 × T))

Where S is speed in meters per minute and T is time in minutes. For most runners, manual computation isn’t necessary. Pacesmith’s Race Time Predictor runs the full calculation from any recent result and outputs predictions across standard distances. The VDOT Calculator also shows training zone paces alongside the equivalent race times, which is what Daniels designed the system for in the first place. The training pace calculator guide explains how to actually use those zones once you have them.

Prediction Accuracy Across Distances

The VDOT model assumes races are run at a physiologically equivalent level of effort. That assumption holds reasonably well for distances between roughly 1500m and the marathon. Move toward either extreme and the predictions become less reliable.

At 800m, anaerobic capacity contributes roughly 50% of energy demand. VDOT is primarily an aerobic measure, so an 800m race time predictor result derived from a 10K or half marathon performance will tend to underestimate a runner with strong anaerobic capacity. It gives a useful floor estimate, not a precise target for a track race.

At the other end, marathon predictions built from 5K or 10K results tend to be optimistic for most recreational runners. The formula assumes equivalent aerobic output over the full distance, but marathon performance is also governed by glycogen management, fat oxidation efficiency, heat tolerance, and long-run training volume. A runner with a high VDOT from track work but limited long-run history will often finish 5-10 minutes slower than predicted.

The most reliable predictions come from adjacent distances. A race time predictor for a half marathon anchored to a recent 10K result will typically land within 2-3% for a well-trained runner. 5K to 10K, 10K to half marathon, and half marathon to marathon are all solid prediction ranges. Going further than one distance step, especially across distances with meaningfully different physiological demands, adds real uncertainty. The race time predictor guide covers how to read those numbers when you’re crossing larger distance gaps.

One edge case worth flagging: VO2 max from a cycling test does not translate cleanly to running predictions. Aerobic capacity transfers across sports, but running economy is specific to running biomechanics. A well-trained cyclist who starts running will have much worse running economy than a runner with the same VO2 max value, so predictions built from a bike-derived VO2 max will overstate running performance, sometimes by several minutes over a marathon.

VDOT vs. the Riegel Formula

Many older prediction tools, including earlier versions of the Runners World calculator, use the Riegel formula: T2 = T1 × (D2/D1)^1.06. It’s simpler to compute and works reasonably within a narrow distance range. The problem is the fixed 1.06 fatigue exponent, which doesn’t hold across all distances or runner types. Recreational runners slow more in the marathon than Riegel predicts. Master’s runners with strong aerobic development but reduced sprint capacity get inconsistent results at shorter distances.

VDOT models oxygen demand as a curve across distances, making it more accurate when predicting further from the input distance. For adjacent distance predictions, the two formulas produce similar results. For longer extrapolations, particularly 5K to marathon or vice versa, VDOT’s edge becomes practically meaningful.

From Prediction to Race Plan

A predicted time is a hypothesis, not a guarantee. The formula doesn’t account for weather (expect roughly 4-8 seconds per mile slower above 60°F depending on humidity and direct sun), course profile, or pacing execution. Use the prediction as an anchor and adjust from there.

The most practical step after getting a prediction you trust is building a split plan. Knowing your goal time is useful; knowing your per-mile targets for a specific course is what you actually use on race day. The Race Split Planner generates a mile-by-mile or km-by-km breakdown for any goal time and distance, with options for even splits or a negative split structure that you can print and carry to the start line.

The best sequence for using these tools: run a race or hard time trial between 5K and half marathon, plug the result into a VDOT-based predictor, cross-check the output against your honest sense of current fitness, then use the target time to build a concrete split plan. If the VDOT prediction and your gut instinct point to different times, trust the feeling more than the formula for your A-race, but use the formula’s output to understand where the disconnect might be.

If you want to work through that whole sequence in one place, Pacesmith handles VDOT scoring, race prediction, and split planning without requiring an account or internet connection.