What the research
actually supports

This is the strongest finding in the whole audit, and it is the guardrail most training apps do not have. A study of 588,071 recorded sessions from 5,205 runners found that a single run exceeding the longest run of the previous 30 days by more than 10% carried a clearly raised injury hazard, and that a run more than double that longest effort roughly doubled it again.

Notably, in the same data set the week to week volume ratio, the number most plans police, showed no such relationship. So Runvel guards the session, not just the week.

Garmin RUNSAFE cohort, British Journal of Sports Medicine, 2025. n = 5,205 runners, 588,071 sessions, 1,820 injuries.

Polarised and pyramidal intensity distributions, where the large majority of running is genuinely easy and a small share is genuinely hard, are what trained distance runners actually do and what performs best in controlled comparisons. Runvel holds a hard ceiling on the share of weekly distance run at intensity, and never schedules two hard days back to back.

Meta analysis of taper protocols points the same way: reduce volume substantially, keep intensity, and keep running as often as you were. Cutting the number of sessions is the part that costs you. Runvel tapers by distance and holds your frequency and your quality work right through to race week.

In a randomised comparison of novice programmes, the group that began at a lower weekly distance finished with roughly a third of the injury rate of the group that started higher. Where you begin matters more than most runners assume, so Runvel anchors week one to what you have actually been running rather than to what the goal would like you to be running.

The most repeated rule in running has been tested directly, and it did not hold. A randomised trial of 532 novice runners compared a graded programme against a standard one and found essentially identical injury rates. A larger observational study of 874 runners found no significant difference across increase bands, and a systematic review concluded there is very limited evidence that changes in training load predict injury at all.

Runvel still ramps gradually, because a slow build costs you very little. What we will not do is describe it to you as protection.

Buist et al., randomised controlled trial, n = 532, 20.8% vs 20.3% injured, P = .90. Nielsen et al., n = 874, GPS measured, no significant difference. Damsted et al., systematic review.

Training guided by heart rate variability does have evidence behind it, but for performance, not for safety. In the trials that beat a fixed plan, the guided group ran fewer hard sessions and still improved more. That is a good reason to gate your hard days on how recovered you are. It is not evidence that doing so prevents injury or overtraining, and we do not claim that it does.