How so?

I agree, and I've heard the same. In my experience, knee problems always disappeared with moderate running. Of course, you have to allow yourself time to recover after running. But as far as I know, the science is that joint lubrication in your knees needs activity to function properly.

The focus is usually on the increased impact during running, but arguably when averaged over the time it is actually somewhat equivalent to walking, the 'air time' of each joint means you are effectively under no load with some form of decompression.

It is effectively a higher load with a lower duty cycle. versus walking with a lower load at about 50% duty cycle.

Joint 'damage' is a misnomer. Joint surfaces are under load/impact and friction while running, that is just biomechanics.

Other mechanical parts like bearings have a load capacity and lifetime. It is not a stretch to model the same for articular surfaces on your hips, knees, ankles.

I was curious so I looked

https://health.clevelandclinic.org/is-running-bad-for-your-k...

> Running does not cause arthritis, and when you rest, your body — and your knees — recover from the effects of running

> The type of surface you run on does matter because it changes the impact on your body

Okay, so if running does not cause arthritis, how can the type of surface matter wrt to causing arthritis?

Then they suggest a new myth, that running prevents arthritis.

https://www.health.harvard.edu/blog/does-running-cause-arthr...

> Runners can get arthritis, but is running the cause?

This section ends with "While the answers are still not entirely clear, we're moving closer."

Then they have a large section on why it's hard to determine.

https://www.nm.org/healthbeat/healthy-tips/fitness/is-runnin...

> Knee and joint pain may be common complaints among runners

> More common sources of pain or injury in runners’ knees are iliotibial band syndrome (ITBS) and patellofemoral pain syndrome (PFPS), says Dr. Mayer.

So it doesn't sound like a myth, unless you restrict your definition of knee health to "osteoarthritis".

Each of these cited a number of studies. I know there's a lot of "these are professional researchers, of course they controlled for bias" and the studies said "we controlled for bias" but it wasn't clear how they controlled for bias or how it was effective.

For example, the one about marathon runners didn't analyze long term affects of marathon running, just whether the people who ran marathons at the time had arthritis (no: if you had arthritis you would probably quit running marathons). Another general meta study seemed similar (recreational runners have lower arthritis, but would you be a recreational runner if you had arthritis?) in addition to noting higher arthritis in long term (15 year) runners vs general population.