50-65% is a good rule of thumb, but class AB amps efficiency often go higher than 65% at rated or higher load condition, More like in the high 70's. It also tends to do better with 4 ohm loads. The main point though is, amp manufacturers do based the transformer VA rating not for literally continuous load, but the so called "continuous" rating that could be for minutes, 20 minutes, or may be a few hours. Also, power transformers, even the small ones used in audio power amplifiers have excellent overload capacity. So a 1,200 VA rated Tx, could be by design, used in applications that have load demand of say 1,500 VA or even more for sustained short duration, e.g. 30 minutes so for audio applications you don't really need 1200 VA for 600 W rated amps.
Actually this is a great example, one thing I like about Monolith amps, and I posted this more than once before on this topic, is that they are not shy on using larger transformers. If you look at Emotiva's class AB amps, the once seemingly very popular XPA-5, rated 200 WX5 specified a 1200 VA transformer, that's half that of Monolith's. Does it mean the XPA-5's efficiency is 83% for 8 ohm load and 125% for 4 ohm load, of course not. By the way, even using the Monolith example, if you use the 4 ohm rated load, the efficiency you calculated will also be significantly higher than the 50-65% number right?
The Emotiva example:
Emotiva XPA-1 Manual Online: Xpa-5 Specifications. XPA‐5 Number of Channels: 5 Topology: Fully Discrete, Dual Differential, High Current, Short Signal Path Class A/B Power output (all channels driven): ...
www.manualslib.com
"XPA‐5
Number of Channels: 5
Topology: Fully Discrete, Dual Differential, High Current, Short Signal Path Class A/B
Power output (all channels driven):
300 watts RMS @ 4 ohm (0.1% THD)
200 watts RMS @ 8 ohm (0.1% THD)
Transformer Size: 1,200VA
Secondary Capacitance: 60,000uF "
There are no shortages of similar examples that you can find on the internet. It just shows that we cannot based on published power output specs and/or even measured output figures to accurately estimate transformer size. I know you are very good at that, so if you take the time, you will agree with me on this.
I rarely say this, but I will this time, i.e., been there done that,.. and having worked in the related heavy industry for many years, I know a little bit about power transformers with hands-on experience, from low voltage low va to high voltage high MVA ones so I thought I would share what I've learnt over the years.