The Best Running Shoes

The straight answer, up front

Buy on fit, then on where you actually run. The technology in expensive shoes is real and measurable, and it is far more situational than the marketing suggests.

This is one of the few product categories in this series where the science is genuinely strong. Running footwear gets studied properly, with treadmills, oxygen consumption and mechanical testing of the shoes themselves. So there are real numbers here.

The number everyone quotes, and the one they leave out

Shoes with advanced footwear technology, the carbon-plated foam ones, improve running economy by roughly 3 to 4 percent on flat ground compared with traditional shoes. That is the well-established figure.

Then researchers asked what happens on hills.

Twelve competitive male runners ran on a treadmill at four gradients, level, 3 percent, 6 percent and 9 percent, in a Nike ZoomX Vaporfly Next% against a Nike ZoomX Streakfly control, in randomised order. Oxygen consumption by indirect calorimetry, and the shoes themselves tested for stiffness and resilience.

The advantage decayed exponentially with gradient. It went from 4.22 percent on the level, to 2.42 percent at 3 percent incline, to 1.05 percent at 6 percent, to 0.52 percent at 9 percent. Roughly a 20 percent reduction in benefit for every 1 percent of gradient.

Half a percent at 9 percent incline is effectively nothing.

Mechanical testing explained why. The Vaporfly stored more energy under compression but had lower resilience, and it was around four times stiffer in bending. That stiffness pays off in a flat, rolling stride, and stops paying when you are climbing.

Twelve competitive male runners on a treadmill, so this is a small, specific sample. But the mechanism and the dose-response with gradient are both clean.

The plate does less than the shoe around it

A systematic review and meta-analysis, PRISMA-registered in PROSPERO, searched four databases to July 2025 for crossover trials comparing carbon-plated shoes against non-plated shoes in healthy adults aged 18 to 70. It pooled 15 studies on step frequency, leg stiffness and peak joint power at the toe, ankle, knee and hip, using random-effects models with GRADE certainty assessment.

As the review states, whether carbon plates systematically alter running biomechanics compared with non-plated shoes remains unclear, which is the question it set out to settle.

The takeaway for a buyer: the carbon plate is the marketing hook, but the biomechanical case for the plate specifically is less settled than the price implies. The foam, the geometry and the stiffness of the whole system are doing work that gets credited to the plate.

Sole thickness, and why 40mm is a number you have heard

A systematic review examined sole thickness specifically, following PRISMA, and included 14 studies, mostly in male recreational or experienced runners.

It exists because World Athletics imposed a 40 mm limit on sole thickness to prevent unfair competitive advantage, and yet, as the review notes, the effects of sole thickness on biomechanics and economy are not well understood.

It also flags the core methodological problem in this whole field: sole thickness usually co-varies with other shoe characteristics, so isolating it is hard. A thicker shoe is generally also a different foam, a different geometry and a different stiffness.

That is worth holding on to when comparing two shoes. You are never comparing one variable.

How I would actually choose

Fit first, and it is not close. None of the studies above will help you if the shoe rubs, and running volume is what improves your running.

Then match the shoe to your terrain. If you run flat roads and races, advanced footwear technology gives a real 3 to 4 percent economy benefit. If you run hills, that benefit shrinks toward nothing by moderate gradients, and the money is better spent elsewhere.

Then decide what you are buying it for. A 4 percent economy improvement matters if you are racing a time. It matters much less if you are running to be fitter.

Then consider rotation. Carbon-plated shoes are expensive and their foams degrade. Using them for every run is the costliest way to own them.

What I would not buy on: the presence of a carbon plate as a feature in isolation, since the meta-analysis found its independent biomechanical effect unclear. Nor on stack height as a proxy for quality, since thickness co-varies with everything else.

On injuries

I looked for the injury answer and I am not going to give you one, because the sources here study economy and biomechanics rather than injury rates.

What the literature does record is that concerns about injury risk with advanced footwear persist, and that they are concerns rather than established findings.

If you are getting injured, the variables with better support are how fast you increase your training volume and your overall load, not which shoe you bought.