Stood at the 22 mile mark of a big city marathon a few years back, watching. The guy who'd been leading comfortably an hour earlier just fell apart right in front of me, legs gone, pace dropping by nearly two minutes a mile inside about ten minutes. Brutal to watch, honestly. Meanwhile someone who'd been sitting quietly mid-pack the whole way cruised past him barely looking troubled. Same race. Same distance. Completely different outcome, and it had nothing to do with who was fitter that day. It came down to a bet each of them had made with their own pacing an hour and a half earlier, whether either one would've called it that at the time.

The actual numbers behind why runners blow up

Here's the physiology bit, because it explains everything that follows. Roughly 87 percent of marathon finishers run a positive split, meaning they slow down in the second half, and only somewhere between 1 and 8 percent of recreational runners ever manage a genuine negative split, finishing faster than they started. Go out even 5 to 10 percent too fast and you can deplete glycogen stores up to 30 percent earlier than planned, which is exactly why so many runners hit a wall somewhere between miles 18 and 22, precisely where I watched that race fall apart. It's the same calculation a spectator makes flicking through this round-up of current free bet offers before backing a favourite in the elite field, the safer runner with a controlled, predictable finish, or the one who might blow up spectacularly but also might just run away with it if the pace holds.

Why the elite record books actually prove this

Before 1988, most record setting marathons were run on a positive split, fast start, sharp fade around the 25km mark, essentially betting everything on early speed and hoping the tank held out. That's changed completely since. Paula Radcliffe's 2003 world record was built on remarkably even, controlled pacing rather than an early burst, and it stood for well over a decade precisely because of how disciplined the effort distribution was. Eliud Kipchoge's sub-two-hour attempts followed the same logic, tightly controlled pace with a slight acceleration held in reserve for the closing stages rather than spent early. The fastest people to ever run this distance have essentially all reached the same conclusion independently, going out conservative and finishing strong beats going out hard and hoping.

Every pacing decision is a wager, whether you frame it that way or not

Think about what a runner's actually doing when they choose a pacing strategy. Positive splitting, going out fast, is a high variance bet, potentially a spectacular time if the legs hold, a genuine collapse if they don't. Negative or even splitting is the conservative play, protecting against the downside in exchange for giving up a bit of that early theoretical upside. Neither is objectively wrong, plenty of race conditions and course profiles (Boston and New York's tougher elevation profiles, for instance) actually favour a more conservative approach regardless of ambition level, but every single runner on a start line has made this calculation whether they'd describe it that way or not.

What separates the ones who get it right

Watching enough marathons, you start noticing the pattern in who actually executes their chosen strategy versus who just hopes adrenaline carries them through. The best pacers treat the whole race like a genuine risk management exercise, banking energy early precisely so there's something left to spend later, rather than spending everything up front and hoping the body doesn't send an invoice it can't cover. That discipline is the actual skill being tested over 26.2 miles, arguably more than raw fitness alone, because plenty of genuinely fit runners still blow the pacing and pay for it in the final six miles.

Still think about that runner at mile 22 sometimes, the exact moment his race fell apart in real time. It wasn't a fitness problem. Was a pacing decision made an hour and a half earlier finally calling in what it owed. Same as any bet that looked good at the time and just didn't hold up once the conditions actually played out.