Volume Control setting corresponds to how many db down (-db) from 0db at full volume?

I’ve wanted to be able to estimate how many watts my amps are driving as it corresponds to my preamp’s volume control setting. More specifically with a typical volume control having a rotational range from say 7:00 o’clock around to 5:00 o’clock, (my Croft preamp actually goes from Noon to 9:00), roughly where would it be -3db or -6db, etc. In one system with my 50 watt push-pull pentode amp, I rarely go past 10:30 o’clock. With my 25 watt triodes in my other system I rarely go past 3:30 o’clock on the Croft. Interestingly that’s roughly the same degrees of rotation!? Can I estimate approximately how many watts I’m operating at in each system simply based upon the volume control setting?
elunkenheimer

I'm sure there is a clever way to do this estimation but it's not coming to me right now.
i'm thinking out loud here & if i end up writing gibberish please excuse me....
most volume controls are 40-45dB in attenuation unless you paid a premium for your unit or the volume control in which case you would get something like 60dB of attenuation. I also believe that you would have something like 24 steps in the volume pot going from full attenuation of -44dB --> no attenuation of 0dB.
so, now you can calculate how much you've attenuated the preamp signal going to the power amp.
even if you know this, you still don't know what amplitude signal is being fed into the power amp input. You could measure this at the inputs of your power amp (probably need to pop the top lid & attach a voltmeter (which will give you the average voltage). 
Even if you know this you don't know the transfer function of the power amp i.e. what input amplitude corresponds to what output power. It's not linear; it's most certainly logarithmic. 
One of my amps has a VU meter & when the volume knob is set to -48dB, I see that i'm using 0.1W (of 120W/ch) on average just eyeballing it.
 

yeah, Almarg, you are correct - the xfer fn of a power amp is linear. don’t know what I was thinking when i wrote that??? ;-) thanks for catching this.

re. VU meters - the manual of my vintage Yamaha says that the response time is 1uS. I don't think it mentioned a bandwidth per se. I'll have to go back & re-read it....

guys, 
i was thinking about this topic this morning & tell me why this won't work: 
elunkenheimer should use a voltmeter & measure the average voltage on the speaker binding posts (of amp or speaker - better at the speaker as it will take into account any I*R drop in the speaker cable). 
Then to calculate power simply do V^2/R where R is manuf specified average resistance of speaker (4 Ohms/6 Ohms/8Ohms, etc). 
elunkenheimer can do this calculation at several (or each if he wants to do the work) preamp volume knob settings.
Speaker impedance varies vs. freq but this will be a rough estimate of power being delivered to the speaker.

thanks Almarg.

But the larger question I would raise about all of this is what's the point?

let's hope the OP chimes in here & let's us know what is on his mind re. this.....