HR sensor unreliable?

I am not completely sure about what I see, but I think that too steep increases or decreases in HR get caught by some error suppression algorithm. If I ramp my training intensity slowly up and down, the watch (in my case a F5X+) will pick up. But if I make sudden changes in intensity, the watch will take much longer picking up on the change, if it ever does.

That's because of wrist HR.... for me it's only information, for precise data you need a chest belt. This is a feature of all wrist HR sensors

Do you use the handrail?
How long does it take before the watch sees 180?

Do you use the handrail?
How long does it take before the watch sees 180?

No nerver use handrail. I can use the stairclimber for 5 minutes and still be in no hr zone or warmup, but my heart is racing and im sweating like a pig... Weird!

It's been the same for me since pretty much day one. When hiking it's generally good enough since i never encounter a steep wall after a completely flat ground, plus I naturally slow the walking down when going up.
For runs I wear a Polar H10 each and every time, because a) that's the only way to get accurate readings with dramatic and quick BPM variations and b) to take into account HRV.

Your optical watch HR monitor should be fairly accurate compared to a chest HR monitor. The only times that your watch HR monitor's accuracy is degraded is when you're using your hands to grab things, flexing your wrist muscles (weights, handrails for stabilty).

If you have any questions, feel free to reach out to me. I have a YouTube channel where I post lots of informative & entertaining videos that talk about running, swimming, and cycling. https://www.youtube.com/watch?v=u3hwFB8-ZuQ&t=226s

I'm happy to report that I've not experienced any problems at all with the wrist based HR. I've had the watch since March so must by now have many thousands of hours of activity data, and hundreds of hours of running and cycling data.

I've run and cycled periodically with a second device paired to a chest strap and they are both really close to one another.

Whenever I've seen posts about optical HR accuracy before, oftentimes it seems that tightening the strap is the solution. I note you have had yours tight, so perhaps there is a fault with the unit?

Have you tried it on the other wrist?

This is my second 645 Music watch and every time I have a workout the HR is not very precise. Especially at the start of the workout. I go on a staircase climber and the watch say I im no zone and my HR is like 80 when my heart is actually racing and in the 180's! During other workout also its the same.

I tried to shave my arm hair and to tighten the watch firm and to have it loose. Same result. My skin is Nordic.

Why is it like this? Is something wrong with the watch or is Garmin bad at wrist HR ?

Thankful for anyone who can help me :)

Wrist based HR has problems with fast pulse changes. they are suitable for aerobic activity (constant pace) but not for intervals for example.
your type of activity change the pulse very fast - so it is like intervals in that sense.
even during constant running, the watch sometime takes time to follow the pulse because the first lap is a bigger pulse change comparing to the following laps.
these are know facts, not just for this watch. you can read the DC Rainmaker reviews of this and other watches for reference.
as different people have different skin color and different running habits, it may vary from person to person but in general this is the situation.

Doing some indoor rowing last week with wrist HR (since my chest strap is broken), I noticed the watch to be approx. 50 beats below realistic values. At some point in time, the HR reported by my watch made a jump, closer to what my HR really was and then another jump to catch up with (almost true) values.

JensDoe Rowing is definitely one of the activities where optical wrist HR is at its most useless due to the arm movement and muscle flexing. If you need reliable HR during rowing I would advice you to either fix your chest strap or get a new one.

There's simply no point in spending time and energy on how to get more reliable readings from the OHR, the technology simply isn't there yet.