Thanks.
So I contacted Coros customer service and they asked me to send in a bunch of data files. They were not able to reach any conclusion and said I needed to run with both my Coros and Garmin so they could see the comparison. So the one time I ran with both watches, the Coros HR was actually in the normal range and had the exact same avg. HR as my Garmin (130 bpm). Damn. For nearly the same exact runs this week (pace, effort) the avg HR on the Coros was 160-180, so it obviously malfunctioned the first 3 times I ran with it. These were all easy runs for me.
The main difference today was that it was warmer, about 55 degrees vs. 30-40 degrees for the other runs. I know cold can be a factor, but that would really suck if it can't perform reasonably well in 30-40 degree weather. Again, I'm not looking for perfection, just a HR reading that has some basis in reality, ~95%.
As for why I chose the Pace 2 over the Garmin 55, it mainly had to do with reviews I read, in particular DC Rainmaker and reddit. I really like the simplicity of the 35, but thought an upgrade would be nice, mainly for some of the more advanced features (e.g. race time predictor). The Pace 2 did really well in GPS accuracy tests and the battery life is supposed to be phenomenal. I'm sure the 55 is good, but nothing stood out to me in any of the reviews I read.
I was kind of surprised by the GPS measurement of my run today. The Pace 2 measured 6.38 vs. 6.30 on the Garmin 35. That's a pretty big difference, more than I would expect. I assume the Pace 2 was the more accurate one but I will see if I can compare my run to Google maps to get a better idea.