battery life reduced to only 4 day when the 'Pulse Ox Mode' was set to ’All Day‘

Probably not. Perhaps you need to ask yourself why you really need all day pulse ox. 

Some smart bands can last for a week while setting the pulse ox to all day.I think instinct 2 can perform better with larger battery.

Do you find these numbers from PulseOx reliable? For me it's a gimmick. After couple of tests I gave up with this feature.

My understanding is that keeping Pulse Ox continuously active is useful only if you want to check your body acclimatization to high altitude (i.e.: you go on a long high altitude hike or spend several days hiking and camping at high altitude).

Since Garmin threw everything but the proverbial kitchen sink in the Instinct 2 and it has become THE outdoor watch in their line-up, it makes sense they gave the user this option. It does not mean everyone needs it or should use it.

See here for some interesting consideration on Pulse Ox use for acclimatization:

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7916608/#:~:text=Conclusions%3A%20The%20use%20of%20finger,AMS%20progression%20and%20treatment%20efficacy.

and here for more than you probably want to know about Pulse Ox as implemented by Garmin:

https://support.garmin.com/en-US/?faq=SK2Y9a9aBp5D6n4sXmPBG7

Pulse Ox during the night can be useful to investigate potential sleeping issues (i.e.: sleeping apnea), always keeping in mind that Garmin watches (and most smartwatches, for that matter) are not medical-grade devices

"always keeping in mind that Garmin watches (and most smartwatches, for that matter) are not medical-grade devices" Exactly. And here is the problem. Would you trust to score you get with this watch according to PulseOx?

My personal test shows difference around 4-5 percent between Instinct 2 and readout from finger. Garmin shows mostly values around 95% which is lower end for typical healthy person. If I would trust to this value I could be worried, but this is one of this crazy features displayed by the watch which tell you everything with big BUT at the end of the sentence. I'm not alone with these values, check it out:

https://youtu.be/XEPjAS7MWP0?t=519

It's really annoying that from so many features you can really trust to few only.

That's why i wouldn't bother with PulseOx working all day.

I am not advocating the use or endorsing the accuracy of Pulse Ox, just stating something that may not be obvious to some of the users. 

If the difference between the Garmin's reading and a medical-grade finger oximeter is always the same, the relative difference from one night to another (of from one day to another if checking altitude acclimatization) might still be enough data for some. Accuracy and precision are two different things when measuring something.

Another thing to keep in mind is that a finger oximeter works in a much more controlled environment (clipped firmly, subject not moving and smooth skin). Totally different than having a watch on your wrist that has to contend with different degree of tightness of the band, movement of the subject, skin hair, tan and complexion. Not really an apple to apple comparison.

Personally I consider this feature a "better than nothing" thing and I would not expect any other smartwatch to perform much better. It would be interesting if Garmin added the possibility to calibrate Pulse Ox if the user has available a more accurate oximeter. But again, there would still be all of the variables I mentioned before. Making a spot check of Pulse Ox, with the watch always in the same position on the wrist, not moving and perhaps on the inside of the wrist may be the best one could aspire to with a smartwatch.

raf200 said:

My personal test shows difference around 4-5 percent between Instinct 2 and readout from finger. Garmin shows mostly values around 95% which is lower end for typical healthy person.

Interesting. My GI2 is well in match with the fingertip oximeter, showing both 97-99% most of the time. Of course, the skin and the state of the veins play a role, and I am apparently lucky it works well with mine.

In fact, I do a lot of apnea training, and fingertip oximeters tend to fail when the diving reflex kicks in, and induces the vasoconstriction. There isn't then sufficient bloodstream under the fingertip oximeter, and it shows unrealistic numbers. An earlobe oximeter is much better. I did not test yet the PulseOx measuring of GI2 under deep hypoxia, after several minutes of apnea, but it should be less impacted by the vasoconstriction than a fingertip oximeter.

Besides monitoring the altitude acclimation, and besides apnea training, the PulseOx may be also useful when you are sick, or have respiratory problems. Especially today, in the era of Covid, it may alert you when the SpO₂ becomes critical. Otherwise, it is true, it makes no big sense to let it on round the clock, except of some testing, because together with GPS, the PulsOx sensor is the most power hungry component.

I checked now again:

-left hand 93%

-right hand: 96%

-left hand again: 99%

All readings taken within 30-40sec.

To be honest, personally I would prefer to have only few features which I can trust to rather than "million" which take readouts from randomizer.

Batman, I value your opinion and preference; sometimes I'm just annoyed that marketing makes things like this. It looks perfect printed as a feature on the box, but reality verifies it and effect is far from perfect.

raf200 said:

To be honest, personally I would prefer to have only few features which I can trust to rather than "million" which take readouts from randomizer.

This is why it is not bad idea to let it on for a few days, and see how it behaves. I did, and the graphs are rather stable, most of the time between 97-99%, dropping to some 95% occasionally during activities, and during sleep. Both of it is comprehensible, and in coherence with measurements I did with a fingertip oximeter. So far so good (for me at least).

Lucky you ;) I crosschecked now with GI1 Solar, percents don't match GI2.

And GI2 gave me numbers from 91 to 96%... Ok, I give up.