Is 'Consistency' Running More Often, or Not Stopping? We Tested It on 8,060 Runs
A viral Terra chart says more runs per week goes with faster marathons. We ran it inside 167 runners. Frequency vanished once we held volume constant. The real signal was the longest gap in your training block, and it slowed people down even at matched volume.
The Chart That Says Just Show Up More
Last week we took the viral "the easier you train, the faster you finish" chart and tested it inside each of our runners instead of across strangers. The huge between-people slope flattened to nothing, and volume turned out to be what actually moved fitness.
That chart came in a pack, and the pack has another panel that gets quoted just as much. Alistair Brownlee made it for Terra, and it plots "training consistency", defined as your mean number of runs per 7 days, against marathon finish time. It slopes the friendly way. More runs a week, faster marathon. The advice that rides on it is simple and everywhere: be consistent, just show up more often, the frequency is the magic.
We had the same suspicion as last time, so we ran the same test.
Why a Scatter of Strangers Can't Answer This
The chart compares 425 different people to each other. And faster runners train more, which usually means they run more often too. So "runs per week" and "fast" hold hands without one having to cause the other. A between-people scatter cannot tell you whether running more often made anyone faster, or whether fast people simply run a lot and a lot of running means a lot of days.
The only way to answer it is to look inside a single runner over time. When the same person ran more often than their own normal, did they get faster? That question has no stranger-to-stranger confound, because everyone is compared only to themselves.
How We Tested It
We pulled 22,683 runs with heart-rate zone data, then narrowed to the 8,060 that were genuinely aerobic: at least 60% of the time in Zone 1 to 2, and 20 minutes or longer. Those came from 167 runners with enough history to build a personal baseline.
On each aerobic run we measured speed at the same heart rate, the cleanest "is this person getting fitter" signal you can pull from normal training without a lab. Then we described that runner's previous 6 weeks three ways: how many runs, how many total minutes, and the longest run-free gap. Each one was compared to that runner's own normal, so zero always means a typical block for them.
Running More Often Is Just Running More
The first thing the within-runner data does is take the legs out from under frequency as its own lever.
Inside a runner, the number of runs in a block and the total minutes in that block move together at r = 0.95. They are almost the same number. That makes sense the second you say it out loud. You do not usually add a Tuesday run and then hold your weekly mileage fixed. You add the run, and the mileage climbs with it.
So we pulled the two apart with a partial correlation. Holding total volume constant, the number of sessions did nothing for efficiency (r = -0.034, if anything a touch negative, because at a fixed volume more sessions just means each one is shorter). Holding the number of sessions constant, volume still tracked getting faster (r = +0.054). Same verdict as the easy-share post. The thing the chart calls "consistency" is the volume story again. Running more often is one of the ways people run more, nothing more than that.
That also squares with the lab. A 2021 review on endurance frequency versus volume found that when intensity is matched, aerobic adaptations follow the total work done, not the number of sessions you chop it into. Spreading the same volume over more days has practical upsides, fresher legs per session and an easier fit around life, but it is not a physiological bonus on its own.
The Real Signal Was the Gaps
Here is where this stops being a rerun of last week.
The single strongest within-runner signal in the whole analysis was not frequency and not even volume. It was the longest run-free stretch in the block. Across 8,000-odd runs it correlated with efficiency at r = -0.081, the biggest number on the board, and it pointed the obvious way. The bigger the hole in your last 6 weeks, the slower you ran at the same heart rate afterwards.
And it was not just a roundabout way of measuring volume. When we held total minutes constant, the gap effect barely moved (r = -0.066). A block with the same mileage but built around a long break still ran worse than the same mileage laid down without one. The dose-response is clean: a block with near-daily running sat about +0.11 standard deviations on efficiency, a block with roughly a 9-day gap sat about -0.15. In real units, that is on the order of 8 sec/km at the same heart rate, between a runner who never let a break open up and the same runner who did.
That is what "consistency" should have meant the whole time. Not a high session count. The absence of a long stop.
How Much a Break Actually Costs
It is worth being honest about size, because the chart's sin is making everything look enormous. The gap penalty is real and it is measurable, but it is sec/km, not minutes/km.
The same frequency-versus-volume review found only a slight sign of aerobic decay after two full weeks of reduced volume. The muscle-level oxidative machinery that drives aerobic fitness is fairly stubborn over a week or two. Nobody loses their engine in a week. They lose a sliver. The reason gaps still matter is that the slivers add up across a season of stop-start blocks, and the runner who keeps the wall intact banks all of those small amounts the stop-start runner keeps paying back.
One caveat we will flag, because it is the honest one. A long gap is often caused by illness, injury, or travel, and those things make the next few runs slower on their own. We checked "days since your last run" as a separate measure of post-gap rust and it was weak (r = -0.025), which argues against the effect being pure next-run rustiness. But we cannot fully separate "you stopped" from "the thing that made you stop." Read the gap number as a strong nudge, not a physical law.
What We Would Do With This
Stop chasing the run count. If you are deciding between four runs a week and six of the same total volume, our data says the split barely matters to your fitness, so split it however keeps you healthy and able to keep going.
None of this makes frequency a dead lever, and we do not want to overcorrect into "never add a run." Adding runs is one of the main ways you add volume, and volume is what moved fitness here, so more days can absolutely help. What the gap result changes is the order you pull the lever in. Get consistent at your current frequency first. Then add days or minutes slowly enough that your recovery keeps up, because the surest way to manufacture the exact long gap that slows you down is to ramp too hard, run yourself into a cold or a niggle, and lose two weeks. Frequency helps when it adds volume you can absorb. It costs you when it buys an injury.
What we would actually defend on the calendar is the no-long-gap rule. When life gets in the way, a 20-minute jog to keep the streak alive is worth far more than its tiny training load, because what you are buying is not the minutes, it is the gap you did not take. Three runs a week with no breaks beats a brilliant week followed by ten days off.
And the same close as last time. Terra's chart is an average of 425 strangers, and it will always make frequency look like the cause. The version that tells you something is your own training block sitting next to your own aerobic efficiency curve, with the gaps marked.
athletedata connects your Garmin, Strava, COROS, WHOOP and the rest, watches your training load and aerobic efficiency over time, and tells you what tomorrow's session should be based on your own runs instead of an average of strangers. Start a free trial.
