Everyone keeps throwing this one-fifth figure around but I personally have never found it to be valid.
And I believe one drawback seems to be beam angles. for example comparing
We have a 5 fitting chandelier in our lounge
, originally we used 25W golfball bulbs as originally supplied with unit but they always had a very short life, eventually we had 2 blow at same time and fitted candle bulbs from stock. No obvious difference in light level and cheaper, they ran much cooler or rather the shades did and bulbs lasted longer - much longer, I concluded it's a ventilation issue.
Moving on and started replacing with LED, initially trying to do a colour match with the filament bulbs using 3.5W/2700K but far too dingy when all old bulbs had failed, we have worked our way up through colours and power to 4.8W/6000K.
The ratings shown are for the LEDs we have fitted and generic from internet for incandescent.
Yes sure we have a nice bright patch on the ceiling but the shade attenuates it soooo much there is very little to the sides and of course nothing below the horizontal. We also have some wall lights, similar sort of design but inverted, 7W and I'll guess around 5000K. We now habitually run with all lights on (52W) if it's anything requiring light (such as reading) and we tend to be in certain places for that where previously it would have been either/or and quite even.
Our hall lights have settled down, I think we got through about 4 or 5 full sets including expensive and highly recommended brands (>£3 each) in the first couple or so years but not yet had a failure in the current cheap things from Ebay in several years.
However it's not these sorts of styles of LED products I gripe about, I accept changes are afoot and it's where we are.
Comparing Fluo with LED is so different to incandescent and IMO at best it's a 1/2 rule but I still come back with lack of colour spectrum and beam angle.