Today, did a 5k on treadmill in 19:50, at 1* incline. But this felt too easy, my Garmin 920xt said I did only 2.7 miles, thus showing a slower pace 7:xx. I can run in 19s (doesn't feel easy) on the road for a race. So for you how do treadmill times compare? I am starting to fix my gait, which is why I assume the road running hurts. I just want to make sure I am putting same effort on the treadmill, and not being lied to that I am running 40s faster per mile.
Is it your treadmill? Calibrate it. You can do this easily by measuring the belt with a chalk line and then counting the revolutions over a specific distance.
You could probably do this on a gym treadmill too, if you did it during a non-busy time and were low-key about it.
I tend to find that my efforts and paces on treadmills are very similar as long as I'm running above 5 minute pace. But I've also run 8x400m in 60 on a treadmill before with 2min rest and there's no way I can do that on a track, so I'm imagining at speeds below v02max it's probably easier, but I find tempos and easy runs to be pretty similar. Also depends on the treadmill, but 90% of treadmills feel the same for me
GPS watches can be lousy on treadmills - as can their calibration (especially in gyms or hotels). This is the main thing I use my Stryd to counter against.
Be that as it may, while a lot of people might say treadmills are easier due to lack of wind resistance etc, there's a significant minority of us who find them harder. For me it's the lack of breeze for cooling and the general mental 'wrongness' of running indoors. Effort-to-pace ratio is always harder for me on treadmills than outside.
i find them to be in the ballpark but they aren't as accurate as you are wishing. if you did them all out, your treadmill pr would probably be 20 seconds faster than your road pr.
As others have said - relying on your Garmin for treadmill distance is unreliable. Treadmills are often terribly calibrated - but you can measure the belt length yourself.
1* incline is suppose to compensate for the loss of wind resistance - but I've no idea if this is really an accurate compensation. I also doubt that the calibration of the incline is terribly accurate in general.