How to choose running shoes: the decision in order

Almost every bad running shoe purchase starts in the same place: with the shoe. Someone reads a review, or likes a colour, or notices what the fast people at the club are wearing, and then works backwards to a reason. It is a pleasant way to shop and a poor way to end up with the right shoe, because the shoe carries no information about you.

Start at the other end. Three facts about your running decide more than every specification on the box put together, and you already know all three.

• Surface. Where do the kilometres actually happen — tarmac and park paths, or dirt, roots and rock? This is a hard gate, not a preference. It is the first question the finder asks, and it is the one that cuts the pool fastest: 59 of the 78 shoes here are road, 17 are trail and 2 are built to do both.
• Distance and pace. A shoe for forty easy minutes is not the shoe for a marathon at goal pace, and neither is the shoe for a 5K. The difference shows up as weight, as cushioning and — at the sharp end — as a plate.
• Frequency. Three runs a week and thirty kilometres is a different wear problem from six runs and eighty. High mileage pushes the answer toward durability and toward more than one pair, for reasons that are arithmetic rather than taste.

Answer for the normal week rather than the aspirational one. The most common expensive mistake in this whole subject is buying for the running you intend to do: a carbon race shoe bought in January for a race in October spends nine months being the wrong shoe for every run in between, and carbon race shoes are the shortest-lived shoes in the catalogue. If the ambition arrives, buying the shoe then is a nicer problem than owning it now.

Once the surface is settled, fit is the next gate and it outranks everything that follows. A shoe that is wrong for your foot is wrong at any stack height, any drop and any price. This is also the part no specification can tell you, which is why it has to be tested rather than read.

Feet swell over a run — they get longer and wider as blood pools and soft tissue spreads under load — so a shoe that fits perfectly in a shop at ten in the morning can be too short at kilometre fifteen. The conventional allowance is roughly a thumb width of space in front of the longest toe, which is not always the big toe. Length is the easy half. Width and shape are the half that actually decides whether the shoe works.

That split matters more than it looks. If you have a broad forefoot, roughly one shoe in seven here is cut for you, and no amount of going up half a size fixes a last that is the wrong shape — a longer shoe is wider only at the point where your foot is not. Equally, a narrow foot in a roomy shoe slides, and sliding is how blisters and black toenails happen. Getting the fit right covers the checks worth doing, including the ones you can only do by running in the shoe rather than standing in it.

One rule survives every other piece of advice here: if a shoe hurts in the shop, it will hurt at kilometre ten. Running shoes do not need breaking in. The upper softens a little and the foam breaks down a lot, but neither process makes a badly shaped shoe fit.

Now the specifications. There are four that carry real information, and a great deal of marketing that does not.

Stack height is how much material sits between your foot and the ground, measured at the heel unless a brand says otherwise. Across this catalogue it runs from 24 mm to 46 mm, with a median of 38 mm — nearly double from one end to the other, and the single biggest difference in how two shoes feel. Drop is heel stack minus forefoot stack: the tilt of the platform, not its thickness. A 40 mm shoe and a 30 mm shoe can share a drop of 8 mm and feel nothing alike.

Drop is where most people are given bad advice. There is no correct drop, only the one your legs are used to: the distribution here is 20 shoes at 8 mm, 15 at 10 mm, 12 at 5 mm, 11 at 6 mm and a long tail either side, with 11 shoes at 4 mm or less and 3 of those at zero. Nothing in that spread is a recommendation. What it does say is that if you have been running happily in 10 mm shoes, dropping to 4 mm is a change your calves will notice for a few weeks, so make it deliberately and not in race week. Stack height and drop explained has the full version.

Weight is the number people overrate and underrate at the same time. Overrate, because 30 g on an easy run is nothing you can feel after the first kilometre. Underrate, because it compounds: at roughly 180 steps a minute, an hour of running is over five thousand times you lift each shoe. That is why the span matters at the extremes — from the 170 g of the lightest shoe here to the 310 g of the heaviest — and why it matters far less in the middle, where most shoes sit within 20 g of each other.

Two features get more discussion than everything else combined, and both are more boring than the discussion suggests.

Plates

A carbon plate is a stiff spoon in the midsole. It does not add energy — nothing in a shoe does — but it changes how the foam under it compresses and how your foot levers off the ground, and in combination with a tall, light, resilient foam it demonstrably makes trained runners faster over a race distance. It also makes the shoe fragile: the 14 carbon shoes here have a median durability score of 48 out of 100 against 80 for the 58 unplated shoes. A race shoe is a consumable.

What carbon reliably buys, in this catalogue, is weight: median 214 g against 278.5 g for shoes with no plate, or about 64 g a foot. It costs a median of €250 against €150 — though the cheapest carbon shoe here is €160, so "carbon means €250 or more" is simply false. If you race, carbon plates and super shoes is worth the ten minutes. If you do not race, the honest answer is that a plate buys you very little.

Stability

The old model — you overpronate, therefore you need a motion-control shoe, therefore you will not get injured — has not held up. The trials that tested prescribing shoes by foot type or pronation did not find the injury reduction the model predicted. That does not make stability shoes useless: plenty of runners simply prefer a wider, firmer, more guided platform, particularly when tired, and preference is a perfectly good reason to buy something.

It does change what you should expect to pay for it. The 10 stability shoes here have a median price of €150 against €160 for the 68 neutral shoes: stability costs nothing. What it costs is weight — a median of 288 g against 258 g, so roughly 30 g a foot. Stability versus neutral works through when that trade is worth making, and shoes and running injuries covers what the evidence does and does not support. The short version of the second one: training load and injury history predict injuries far better than footwear does, and anything that actually hurts is a question for a clinician rather than for a shop.

Prices here run from €100 to €310, with a median of €160. That is a threefold span, and the interesting thing about it is how unevenly the improvements are distributed across it.

Below about €140 you are buying a real running shoe with fewer of the expensive foams: 17 shoes sit in that band, with a median weight of 264 g and a median durability score of 80. Just above that, from €141 to €180, where 43 of the 78 live, you get the better midsoles and most of the max-cushion options — and the median weight goes up rather than down, because that is where the deep, soft, heavy shoes are. Above €220 the catalogue changes character entirely: all 12 shoes there are carbon race shoes, the median weight falls to 212.5 g, and the median durability score falls to 46.5.

So the money does buy something, but not a smooth upgrade. It buys cushioning in the middle of the range and speed at the top, and it buys less lifespan at the top than at the bottom. How much to spend on running shoes takes the whole spread apart band by band, including the case against putting all of it into one pair.

1. 1

   Name the surface

   Road, trail, or genuinely both. This is a gate rather than a weighting: trail lugs on tarmac wear flat in a season, and a road outsole on wet rock is a fall waiting to happen. If your week is mostly pavement with the occasional gravel path, the answer is a road shoe — see road versus trail for where that stops being true.

2. 2

   Name the job

   Easy daily mileage, long runs, fast sessions, or racing. Most people need the first one and think they need the fourth. A daily trainer that is good at everything and excellent at nothing is the correct first purchase for almost everybody, and it is also the cheapest slot to fill.

3. 3

   Fit it properly

   Try it late in the day, in your running socks, with your own insoles if you use them, and run in it if the shop lets you. Check length with a thumb width in front of the longest toe, then check that the midfoot holds without pressure and the heel does not lift. A shoe that fails any of those is out, however good the reviews were.

4. 4

   Read only the numbers that matter

   Stack for how much foam you want underfoot, drop against what your legs are already used to, weight against the job. Ignore the foam names — every brand has invented one, and none of them are comparable between brands. If you want the long version of the two that get muddled most, it is in stack height and drop explained .

5. 5

   Set the budget last, not first

   Decide what the shoe has to do, then find the cheapest shoe in the catalogue that does it. Done in this order, the median answer is around €160 and a lot of people spend less. Done the other way round, the budget quietly turns into a shopping target.

The order is not ceremony. Each step deletes options from the next: surface removes three quarters of the catalogue, fit removes most of what is left, and by the time you get to price there are usually five or six real candidates rather than 78. Run it backwards — start from a budget and a brand — and every step after it is a guess.

One pair is the right answer when you are starting, and stops being the right answer sooner than most people think. The reason is mechanical rather than aesthetic: midsole foam recovers slowly. Running consecutive days in the same shoe means the second run happens on foam that has not fully rebounded, which is both a worse run and a faster death for the shoe.

Two pairs of the same shoe already fix most of that. Two different pairs — a soft one for easy days, a lighter one for fast days — fix it and give you a shoe for each job, which is the whole argument for a rotation. At three or four runs a week, two pairs is plenty. Above about fifty kilometres a week, three starts to pay for itself. Building a shoe rotation sets out which slots to fill and in what order.

The pairs you own also outlast the ones you look after badly, which is a smaller effect than rotation but free. Shoes die of compressed foam, not of dirty uppers, and the fastest way to compress foam prematurely is heat — a radiator, a tumble dryer, a car in July. Looking after your running shoes is short, and when to replace running shoes covers how to tell the difference between a shoe that is finished and a shoe that is merely scruffy.

That is the method: surface, job, fit, numbers, price, and only then a shortlist. If you would rather see the output before doing the work, these are shortlists already assembled from particular sets of answers, with the reasoning attached. None of them is exactly your running, and that is the point — read two of them and you can see which shoe moves when one answer changes.

Or skip the shortlists and start from your own answers. The finder asks the questions in the order above and ranks the whole catalogue against them, and the prepared answers index the combinations people ask for most.