Accelerometer, pace and GPS

I have an old SDM4 (branded Adidas actually !). I tried it, and had a couple of significant errors at time. but it does seem more accurate. Maybe it needs to be calibrated more than the Stryd ? The Stryd power stuff seems useless to me, I'm not running on roads anymore, and not at a current pace / power on long nature runs. It's not my main indicator, so it's a bit expensive to get an accurate distance...

The 6X can go from accurate to quite poor when it comes to distance, mostly by underestimating it. The old watches were often doing the opposite, overestimating the distance. But sometimes, it's really really underestimating too much, the cure is worse than the disease. But it's a very complicated pb to solve i'm sure. Between different GPS conditions, the results must be very different and it must be hard to adjust. 

That's a shitload of data :)
A waterproof Stryd without power indicators would be all I need. i'm curious to know if a SDM4 would be reliable for nature runs, when i actually, well... don't always run :) Steep mountain runs are a mix of all paces, from very slow to decently fast. It must be more challenging for a footpod. If I can put my hands on a Stryd, I will try to do a few runs like that with both the Stryd and my old SDM4 (well, if this Adidas footpod is REALLY an SDM4, not really sure !). 

The SDM4 were Garmin branded Dynastream foot pods (see http://www.garmin.sk/img.asp?attid=8735) and I believe your Adidas one is the same thing (there were quite a few different brandings).  As far as I know, both Dynastream foot pods and RunScribe are really only designed for running, as walking has a very different cycle and would need a different algorithm.  I recall seeing in some sort of marketing blurb, that walking is something that Stryd did tune their pods software to cope with a couple of years back; so we're back to you spanking the $$$ on Stryd again

yes, that's what I figured out... I'm trying to find a second hand one, but it's not like there are tons of them ! and the fact that it's apparently NOT waterproof is a concern. A few times my feet spent some time not just under a heavy rain but actually under some mud / water / whatever I have to go through. So I hope it's decently water resistant !

I was also wondering what the water resistance for Stryd is too.  IP67 is the answer https://support.stryd.com/hc/en-us/articles/115003992334-Can-I-use-Stryd-in-streams-and-rivers- so it's OK to run in the rain (and probably wet grass) https://support.stryd.com/hc/en-us/articles/360035284853-Why-is-my-power-off-when-running-through-rain-or-a-puddle- but not through streams or standing water.  As I recall, RunScribe are the same.  The SDM4 doesn't have the barometer sensor hole like Stryd and RunScribe do, so I assume it's more water resistant (I've certainly taken mine through deep standing water on cross country runs).  So we've now saved you money , as the Stryd pod can't cope with your use case and your Adidas one seems to fit the bill.

(I've just done some looking around and it seems you can't even get those Dynastream foot pods anywhere now; whether Garmin, Suunto or Adidas branded - The Zwift Runpod (formerly Milestone pod) https://zwift.com/uk/shop/product/zwift-runpod seems to be the only cheap option now, but it's BLE only though).

i was able to get good to very good or consistent results without a footpod on the 910 / 920 / 935; fenix 5 was also good-ish. i know i cannot convince you :-) - but for me that error in gps measurements or pace display is the difference between a sub 3h marathon or ar 3h marathon; my personal oppinion is that technology should have advanced as well, but with regards to the new sony chipset, things got worse (as said, at least for me and for the location i'm running)

I think it's as simple as the form factor and drive for better battery life in newer GPS watches, has forced all manufacturers to move away from a GPS antenna in a plastic 'bulge' below the screen and move it to a sub-optimal location behind the bezel, as well as chips that use less power at possibly an accuracy trade off.  My sense is it's the antenna placement that is actually more important than the chip though.  Everyone berated Garmin's move from SiRF Star to Mediatek GPS chips a while back, assuming that was why accuracy took a dip (not also noting the concurrent antenna design change).  Then Suunto released the Spartan Trainer Wrist HR with the same Mediatek chip, but uniquely the antena was in a plastic bulge below the screen; guess what? the Spartan Trainer Wrist HR was instantly hailed by all reviewers as having the best GPS accuracy of the watches being released that year.... Basically, if you want a GPS watch that gives you optimal pace/distance accuracy from the GPS data alone, then you'll need something that resembles a helicopter landing pad on your wrist; like the good old FR 305

Well, it's also that the manufacturer of the SirfStar chipset was bought by I don't remember who. Broadcomm maybe ? So you don't have that much choice, and battery life seems to be a strong marketing argument. It's indeed great not to have to check the battery all the time, especially as I wear my Fenix 24/7. I knew i would be sacrificing a bit of GPS accuracy compared to my Ambit 3, and I'm delighted with the 3 weeks I can get out of my 6X ! It's really impressive. I just wish it was possible to turn off the black magic of the accelerometer tweak, it's giving worse results than pure GPS data from what I see. 

And indeed, the antenna is a real problem. Even dual frequency GPS may not drastically improve the accuracy if the antenna is a strong limit factor. We'll know that in a couple of years I guess

At least, I am not the only one ... 

Strange thing: pace is more accurate at higher speed than lower speed, at least in my case. Maybe because my cadence is closer to standard then?

I will try with my old 620 to compare.