Most running apps give you a number. “Your predicted 10K time is 52:30.” Clean, precise, satisfying — and dishonest.

Because no model can predict your race time to the second. The weather on race day, how you slept, the course profile, whether you paced the first mile correctly, the person next to you who started too fast and pulled you with them — all of these affect your finish time in ways no algorithm can anticipate.

Pacewright gives you a range instead. Here’s why.

Precision vs Accuracy

A prediction of “52:30” is precise. A prediction of “50:45 to 54:15” is less precise but more accurate — because your actual finish time almost certainly falls somewhere in a window, not on a single point.

False precision is worse than honest imprecision. If the app says 52:30 and you finish in 54:00, you think you underperformed. But you didn’t — 54:00 was always within the realistic window. The app just didn’t tell you that.

Where Ranges Appear in Pacewright

Race predictions use a P20-P80 confidence interval — roughly an 80% chance your actual time falls within the range. The interval width depends on data quality: recent race data produces a tight range; old training-only data produces a wide one. The minimum range is ±3%, because even the best data can’t account for race-day variables.

Recovery windows are ranges, not single numbers. An easy run needs 0-24 hours of recovery. VO2max intervals need 48-72 hours. Your body doesn’t recover on a precise schedule — biology is messier than that.

Volume increase caps vary by mileage level rather than applying a universal percentage. The “safe weekly increase” is 5-20% depending on where you are, not a single number that pretends all runners are the same.

Detraining estimates are ranges too. Seven days off costs 2-4% of fitness, not a precise 3%. Fourteen days costs 4-8%. The variation depends on your training history, fitness level, and what you did (or didn’t do) during the break.

Why This Matters

A single number creates a target. A target creates pressure. Pressure creates bad decisions — running faster than you should in mile 1 because the prediction says you “should” hit a certain pace.

A range creates a window. A window gives you room to execute. If your range is 50:45 to 54:15, you know that anything in there is a solid performance. You can run your race by effort, check in at the halfway point, and adjust — rather than chasing a specific number that might not have been realistic for today’s conditions.

The Transparency Layer

Every range in Pacewright comes with an explanation of what determined its width. Race predictions tell you which models were used, how old the data is, and what factors widened the interval. Recovery estimates tell you what workout type drove the recommendation.

You don’t just see a range — you see why it’s that wide. If you want a tighter prediction, the app tells you what data would narrow it (a recent race at a similar distance, for example). That’s more useful than a false single number.