When to replace running shoes: how to check, not what to count

Ask ten runners when to replace a pair and you get one answer: five hundred kilometres. Or eight hundred. Or five hundred miles, if the person learned to run somewhere that measures in miles. The figure is repeated by shops, by brands, by training plans and by whoever is standing next to you at the start, and it is not exactly wrong. It is a range being quoted as though it were a number.

Notice what that range is doing. Between 500 km and 800 km there are 300 km of daylight — for most people running three or four times a week, that is a whole season of training. A rule whose lower and upper bounds sit months apart is not telling you when your shoes are finished. It is telling you that whoever wrote it cannot know, because the answer depends on which shoe, which runner, which surfaces and which paces, and a printed rule knows none of the four.

This catalogue makes the point better than an argument can. Every shoe here carries an editorial durability score out of 100 — a comparative judgement compiled for this app rather than a laboratory result, which is why the app says so under every result. Across the 78 shoes those scores run from 38 to 92, with a median of 78. The lowest belongs to the Nike Vaporfly 3, a carbon racing shoe whose own entry warns that it is fragile and should be saved for races and key sessions. The highest belongs to the Brooks Cascadia 18, a trail shoe built to be dragged across rock for years. Both are running shoes. No single kilometre figure covers both, and any figure claiming to describes neither.

One thing to rule out before you read any further. If you are here because the pair you own has never felt right, the problem may not be wear at all — a shoe that never suited you does not become suitable at 400 km. Start with how to choose running shoes instead, and come back to this when you are replacing something that was working.

A running shoe is three things bonded together, and they fail on three different clocks. Almost all the confusion about replacement comes from judging the whole shoe by the only clock you can see.

The outsole is rubber, and rubber abrades. It goes smooth in patches, those patches spread, and eventually the rubber is thin enough that the midsole beneath it starts taking the road directly. This is slow, obvious and easy to photograph, which is why it dominates every conversation about worn shoes.

The midsole is the part that does the actual work, and it fails differently. Foam is gas held in a matrix of thin cell walls. Every footfall compresses it and it springs back — that spring is the entire product. Over tens of thousands of cycles, some of those cell walls stop springing back: they buckle, crease, and stay creased. The foam gets slightly thinner, noticeably denser and much less lively. Almost none of that is visible from arm's length.

So here is the whole problem in one sentence: the part you can see is not the part that fails first. Outsole rubber is chosen to survive abrasion. Midsole foam is chosen to be light and bouncy, which is close to the opposite specification. On a road shoe the foam is usually finished well before the rubber looks worn, which is exactly why a pair can look perfectly serviceable and feel like running on a plank.

The upper is the third clock and usually the slowest. Mesh stretches, the crease across the toe box deepens, the heel counter softens. Uppers normally outlive the foam. When one does not — a tear at the flex point, a heel that no longer holds — the shoe is done regardless of what the midsole has left.

One more thing worth knowing about modern foams. The supercritical nitrogen-blown foams that have made the last few years of running shoes so much livelier are, broadly, less durable than the old moulded EVA they replaced. That is not a scandal; it is the same trade the industry has always made between rebound and lifespan, now made at a higher price. It does mean that "shoes last longer than they used to" is not a safe assumption.

Turn the shoes over. What you are looking at is a record of how you land and push off — not a fault that needs correcting. That distinction matters more than it sounds: the older retail habit of reading a wear pattern and prescribing a shoe category to counteract it did not survive being tested, and shoes and running injuries goes through what happened when it was.

• Wear on the outer edge of the heel is the most common pattern there is. Most runners land slightly outside of centre at the heel. It is normal, it is not a problem to be solved, and on its own it is not a reason to change shoe category.
• Wear concentrated across the ball of the foot, with a clean heel, usually means you run fast, or on your forefoot, or both. Lighter shoes often carry rubber only where the brand expects contact, so this shows up sooner on them than on a daily trainer.
• One shoe wearing faster than the other is common and is not in itself a sign of anything. Road camber, which way round you run your loops, and ordinary left-right asymmetry all do it. If it arrives together with pain, that is a question for a physiotherapist rather than for a shoe box.
• Midsole foam showing through the rubber is the one unambiguous reading on this list. Once foam is in contact with tarmac, that part of the shoe is finished for road use, and it goes quickly from there.

Camber is worth a sentence of its own, because it is free to fix. Running the same direction along the same cambered road every day tilts every single footfall slightly, and the outsole registers it before anything else does. Alternating direction on out-and-back routes costs nothing and evens the wear out.

How the shoes are treated between runs matters more than most people expect, and it is the one lever that buys life back for free. Looking after your running shoes covers the things that shorten a midsole for no good reason — radiators, car boots in summer, washing machines.

Neither of these is a measurement. Both are comparisons, and that is the point: you are asking whether this shoe still behaves like a shoe, using a newer pair or the other half of your rotation as the reference. Five minutes, no equipment.

1. 1

   Get them clean and at room temperature

   Knock the grit out of the outsole and let the shoes sit indoors for an hour if they have been in the cold or in a hot car. Foam feels softer warm and dead cold, and softness is the exact thing you are about to judge.

2. 2

   The compression test

   Hold the shoe in both hands and press your thumb hard into the midsole from the side, first at the heel, then under the ball of the foot. New foam gives and then pushes back at you. Tired foam gives and stays given — it feels packed, like polystyrene that has already been sat on. Do both shoes, then immediately do the same to another pair. The difference between the two is the measurement.

3. 3

   The twist test

   Hold the heel in one hand and the toe in the other and wring the shoe gently, the way you would wring out a towel. Every shoe twists a little. A shoe that twists easily and keeps going, or that folds almost flat across the middle with no resistance, has lost structure it had when you bought it. Again, compare it against another pair rather than against an idea of how much twist is correct — a racing flat and a max-cushion trainer are legitimately different here.

4. 4

   Read the sidewall

   Look along the midsole from the side, at eye level, in decent light. Fine surface creasing is normal from the first week and means nothing. Deep wrinkles that run through the foam and do not spring back when you flex the shoe, foam bulging out over the edge of the outsole, or a heel that has visibly slumped to one side all say the same thing: the foam has taken a permanent set.

5. 5

   Stand them on a flat floor

   Put both shoes on a hard, level surface and crouch to look at them from behind. A heel that leans noticeably inwards or outwards is compressed midsole, not a manufacturing quirk, and it will get worse rather than better.

Two runners can buy the same shoe on the same morning and finish it three months apart, both of them running sensibly. Four things do most of that work.

• Body mass. Every footfall loads the midsole with a multiple of body weight, and foam fatigue depends on the size of those loads as well as their number. A heavier runner compresses the same foam further on every stride and reaches the same total compression sooner. That is not a judgement about anybody; it is what happens to foam.
• Pace. Fast running lands harder and pushes off harder than easy running does. A pair that has only ever done easy kilometres has had an easier life than a pair that has done intervals, at identical mileage on the log.
• Surface. Tarmac abrades rubber and concrete is harsher still. Trail is gentler on foam and harder on everything else: lugs round off, rock strikes bruise the midsole, uppers get torn open. In this catalogue the 17 trail shoes carry a median durability score of 80 against 75 for the 59 road shoes — which reads the right way round, because trail shoes are built tougher out of necessity. Road versus trail running shoes covers the rest of that difference.
• The shoe itself. This is the largest of the four and the one people most often leave out of the calculation.

The plate makes that last point clearest. Split the catalogue by what is inside the midsole and the durability scores separate hard: the 14 carbon-plated shoes have a median of 48, the six nylon-plated ones sit at 73.5, and the 58 shoes with no plate at all sit at 80. Carbon racing shoes are not badly made. They are made for a small number of very fast kilometres out of the liveliest and least durable foam available, and priced accordingly — a median €250 against €150 for the unplated shoes here. If you race in a carbon shoe and train in it too, expect it to be finished in a fraction of the life of a daily trainer. Carbon plates and super shoes works through what the money actually buys.

The same logic runs all the way down the list. The Nike Invincible 3 is the heaviest shoe in this catalogue at 310 g and among the softest, and its own entry notes that the outsole wears faster than the price suggests. That is not a defect either. It is what maximum softness costs, stated honestly on the box.

The single most effective thing you can do to get more kilometres out of a shoe budget is to stop wearing one pair for everything.

Part of the reason is physical. Midsole foam does not recover the instant you take the shoe off: mechanical testing of running-shoe foams shows that rebound lost during loading returns over hours rather than seconds, which is the real basis for the old advice to give a pair a day off. A pair that gets 48 hours between runs starts every run closer to its own baseline than a pair that goes out every morning.

There is also an observational finding here, and it deserves reporting carefully rather than enthusiastically. A prospective study of recreational runners that tracked which shoes each person used found that those who rotated several pairs reported a lower injury rate over the study period than those who ran in one. It is observational, so it cannot separate the rotation from the kind of runner who rotates — people who own several pairs also tend to run more varied sessions and to have been running longer. Treat it as a reason to rotate if you were going to anyway, not as a reason to expect a second pair to protect you.

Be honest about the arithmetic, though. Three pairs do not last three times as long as one. They each last somewhat longer, and — more importantly — the hardest sessions land on the shoe built for them instead of grinding down the shoe you also use for everything else. The result is that you replace pairs less often and each pair spends its life doing work it suits. Building a shoe rotation shows what two-pair and three-pair weeks actually look like, and how little the second pair has to cost: this catalogue starts at €100.

Here is what actually tells you, roughly in the order the signals tend to arrive.

• The ride has gone flat. This one comes first and is the hardest to describe. The shoe stops giving anything back, the same pace costs more effort, and landings feel like landings rather than like something absorbing them. Most runners recognise it within a single run of switching to a fresher pair — which is why the comparison matters so much.
• A new ache that turns up only in that pair. Not discomfort on every run: something that appears when you wear these shoes, is absent in another pair, and does that repeatedly over several runs.
• Deep creasing through the midsole sidewall that does not spring back, or foam bulging out over the edge of the outsole.
• Midsole foam visible through the outsole rubber at the heel or under the ball of the foot.
• The heel counter has gone soft enough that the shoe folds as you pull it on, or the upper has torn at the toe flex crease.
• You have quietly stopped picking them up. Rotations sort themselves out on their own; the pair you reach past every morning has usually already told you what it thinks.

That second signal needs one qualification, and it is the most important paragraph in this guide. A new ache that only appears in one pair is a good reason to stop running in that pair. It is not a diagnosis, and a different shoe is not a treatment. If something hurts on more than a couple of runs, hurts when you are not running, or is getting worse, the next step is a physiotherapist or a sports doctor — not a shoe shop. Shoes and running injuries sets out what the evidence does and does not support, and the short version is that it supports considerably less than the marketing does.

Replace a pair before it becomes unusable rather than after. The comfortable way is to buy the new pair while the old one still has a few weeks left, run the new one on short easy runs first, and let the old pair see out its life on dog walks and gym sessions. You also gain a fresh reference for the compression test on everything else you own.

If a pair worked, the safest replacement is the same model, and the second safest is the same model one generation on. Brands change midsole foam and last shape between versions far more often than the unchanged name suggests, so check the fit rather than assuming it carried over. Getting the fit right covers what to check in the first ten minutes, and how much to spend on running shoes covers what the money above this catalogue's median €160 is actually buying.

If the old pair was not right, this is the moment to change something deliberately rather than by accident — and to change one thing at a time, so you find out which change did it. If the thing you want to change is the geometry, read stack height and drop explained first, because the two numbers are different and most boxes print only one of them.