Honestly, I have come to hate PAR as it is about as useful as Lux due to the reasons you stated. A Red/Blue light putting out 800 PAR is equivalent to around 2000w of MH if you simply measure the same spectrums, but of course the MH will have more spectrums overall than a red/blue light. Then you add warm white LEDs to the mix and all bets are off, you simply can not make any type of legitimate comparison unless you have a way to break down every spectrum/intensity and show individual numbers for specific spectral output, but we don't have that ability at home. This is why I have come to hate PAR, people seem to put so much into PAR #s when in reality it is a crude form of measurement that can't really make fair comparisons between different lighting technologies. It is best used to compare bulbs within the same technology.
In the pictures above, I compared a 10k XM MH bulb to a DIY Cree LED light with optics, but I used the same electric setting and the LED light has much more blue output, so in reality, the LED PAR reading is actually quite a bit higher than what is pictured. Changing it to 'Sun' mode gives you a more accurate reading with LEDs, and adds about 100 PAR to the number, then doing the math corrections for the blue spectrum adds another 200 PAR to the light, so the LED output is closer to 2200 PAR vs 800 PAR, but to my eyes, the 400w MH actually looks brighter because most of the spectral output is in the yellow range, which does very little for photosynthesis.