Built-in optical HR sensor experience while running (image inside)

Please look on the image. The point of confusion is sharp HR drop when I accelerated from 3.25 to 3.09 and HR dropped from 174 to 165 :) 

It is very unlikely that it can be true. 
I understand that built-in HRS is not perfect but just curious what can cause such results.
It looks more like Connect sync problem (note the HR spikes look very similar to the pace spikes but shifted left - but, on other hand, the previous interval looks fine) or optical HR started to count two HR beats as one because of very high HR that incorrectly appeared as HR drop.
This is not single event, I noted this many times.
Anyone experienced something similar? 

  • If you run faster, maybe your watch moves more, so light can enter the sensor, cause distortion of the OHR measurement. Does it change when putting the watch a little tighter?

  • If you spend just a little time looking at the multitude of threads discussion optical HR you will see exactly the same thing reported a multitude of times. And you'll find out why, why not, and any other of millions of suggestions and hypotheses about what causes it. 

    Ultimately it comes down to the technology.

    Ultimately, ultimately, the solution is get a strap.

  • Finally, I've got a strap and it's awesome experience compare to optical HR. If one will once found this topic I strongly encourage to do the same. Even though YouTube tests show little difference in precision, the lag make a difference. The strap is really another level.  

  • It's unfortunate that we still need a strap for 'active' activities but that's limitation of the technology. WHR has got better but it's not quite there and I am not convinced it will be there consistently. However, I do find myself using the strap less and less for steady state activities - long steady run and bike rides - but would definitely use the strap if doing interval involving rapid changes of energy expenditure.