Negative splits aren't a brag — they're physiology
The fastest marathons in the world — Kipchoge's records, Kiptum's Chicago — were run with the second half slightly faster than the first. That's not coincidence. There's a stack of physiology behind it.
A 2025 mini-review in Frontiers in Physiology synthesized the case for negative splits across four mechanisms:
Glycogen. Conservative early pacing burns proportionally more fat and slows glycogen depletion. Running 10 seconds per mile too fast in the first half costs disproportionately more glycogen than running 10 seconds per mile too slow saves — the curve is non-linear because higher intensity tips you toward carbohydrate combustion.
Thermoregulation. Lower early intensity means less metabolic heat generation early, leaving more thermal headroom for the late-race miles when core temperature has been climbing for an hour.
Cardiovascular drift. Heart rate creeps up over the course of a long run even at constant pace — the well-documented cardiovascular drift phenomenon. Starting too hard accelerates this drift and can push you out of aerobic territory in the last 10K, where most "hitting the wall" actually happens.
Biomechanical efficiency. Form deteriorates as fatigue accumulates. Conservative early pacing preserves the neuromuscular coordination that lets you maintain stride length and cadence late.
Caveat: research that compared even and negative splits in elite marathoners found that absolute-fastest performances are usually run as even or only slightly negative — at the very top of the sport, runners are already so close to their physiological ceiling that there's nothing to hold back. For age-group runners, who have more headroom between starting pace and ceiling, the case for negative splits is much stronger.
Practical takeaway: bank seconds early and you spend minutes late. Start 1–2% slower than goal pace and let the second half come to you.
Source: Grivas, G. V. (2025). The physiology and psychology of negative splits. Frontiers in Physiology.
RunNerd looks at your last several long runs and tempo runs to model the pace you can hold for the full race distance, then biases the prescribed race-pace work toward a slightly conservative first-half target. On race day the coach's pacing plan will explicitly suggest a 5–15 second per mile buffer below goal pace for the first 6 miles, with the option to drop the hammer at halfway based on how your HR is tracking relative to your easy baseline. If your training data shows you consistently positive-split your long runs, the coach will flag that as a discipline issue worth fixing before race day, not just a pacing quirk.