Watt for watt in flower quantum boards or any board or strip or cob of equal efficiency would crush a 400w hps in.
Its hard to compare bulbs just from cheap to top end, some of the cheap ones perform almost as good as the top end HPS bulbs but there are plenty that perform way less too.
Same can be said for ballasts.
You should be able to find some test results for various bulb comparisons online.
In terms of yield per square foot there's not much comparison. The leds get better yield per watt but you gotta use them at less wattage per foot so the overall outcome is pretty much the same.
It depends on climate. Leds needs +25-26C, and many times, depending on deaign, fo not give out enough heat to heat a room.
Generally id say leds beat hps, yiu can design much better spread of the light than with a reflector.
This is why grams per watt is a poor metric for comparing yields. In a 2D grow (where we aren't side lighting) I think the best metric is grams per square foot. This allows yield of different lighting technologies to be better compared. If you are using LED and pulling less per square foot then yields aren't as good even if the wattage per square foot is less. Grams per watt is only relative to cost of production. Compare the yield per sqft and if the drop in yield is worth more than the cost of energy saved then you really didn't save anything, you lost.
Now I am not saying that LED can't yield, I am just showing why grams per watt is a poor metric.
Depends on the cheap bulb. I will say that if you can replace the cheap bulb every cycle you will get more yield in the long run than if you ran an expensive bulb for 3 cycles. Of course replacing the high end bulb every cycle would be even better.
If a bulb looses 10% of it's output in say one cycle and lets assume that 10% less light = 10% less yield, is the 10% yield worth more than the bulb costs? Thats pretty much a no brainer. Lets say we have a 1kW SE HPS and with a new bulb we pull 1000 grams off it. 10% would be 100 grams of weed, and I don't know anywhere that wholesale prices have dropped so much that 100 grams isn't worth more than a new philips or hortilux bulb.
I think generalizing everything above what appears to be 560nm as "red" to get a bigger percentage might be a little misleading. Compare that to this spectrum where the majority of the 560+ is in a narrow band at 660nm, where the bestva light has a peak up around 760nm.
Infrared starts around 700nm so the bestva lights additional red peak thats around 760nm ish would likely contribute to warmer leaf surface temperatures. In this case one might need to run the canopy temperature cooler as would be typical of a HPS grow. This could also contribute to yield in so much as some LED growers experience difficulty achieving leaf surface temperatures that are warm enough to achieve optimal growth rates.
It would be interesting to quantify the IR heating effect of the bestva light.