Bob N., I also dug out the Kratville book for another look. Since the numbers you quote (from p.15) don't totally agree with the roster diagram numbers on p.64, I'm left with the impression that the usually quoted Big Boy numbers aren't totally safe for unqualified use in any comparisons. Kratville doesn't address this; his book seems to be in the nature of an appreciation coupled with a documentary record, with no obvious axe-grinding.
The Huddleston & Dixon book on the Allegheny takes a somewhat different tack, though. That the originally published weight of the initial batch (724 500 lb) was understated seems to have been adequately demonstrated; indeed the C&O roster diagram for #1600-1644, shown on p.192, quotes 771 300 lb. H&D make a case that there were differences amongst the individual batches in this group.
Incidentally, numbers like 771 300 lb for the Allegheny, and 772 250 lb for the second series Big Boys imply mind-boggling levels of precision! They must have had really good scales in those days! Although to be fair, many numbers and statistics get bandied around without apparent regard for implied precision - rigorous inclusion of the precision and bias aspects in test methodologies seems to be a relatively recent development. I wonder too, how good was the casting and machining; did the weights within a batch of nominally identical locos vary much? Let's guess a quarter of a percent. Hmmm, at 777 000 lb, that's close to a ton!
But as Bob Q. has outpointed, the "real" numbers are likely by now inaccessible.
My own take is:
There's some evidence that the heaviest Allegheny might have been a bit heavier than the heaviest Big Boy, but the difference was probably small.
The evidence suggests that an optimized Allegheny (i.e. without the excess weight) is probably a bit lighter than an optimized Big Boy.
But both statements must be qualified by the fact that quoted Big Boy weights don't appear to have been as closely scrutinized as the Allegheny case, perhaps because there's no prime facia case to suspect foul play.
- Really, too close to call with any precision on weight.
On power output, the limited measured data available, roundly 7500 dbhp for the Allegheny and 6300 dbhp for the Big Boy don't seem to be wildly inconsistent with the relative proportions of each. I'd expect the BB to have more machinery friction, and so a bigger gap between ihp and dbhp.
So, one might expect the Allegheny to have the edge in (drawbar) power, and the available data don't contradict this.
- Allegheny is the probable winner on power.
On the 7000 hp number for BB raised by Bob Q., I suspect that this comes from the Alco advertising of the 1940s, which did use this number. Ihp or dbhp - who knows, probably bshp, aka hype.
On the Allegheny cylinder sleeving, consider that the designers might have been working to the original (mythical) weights, which had 471 000 lb on the drivers, giving a 4.27 factor of adhesion (FA). Now as far as I can determine, 6-coupled engine units need a higher FA than 8-coupled engine units for similar slip resistance. And early x-6-6-y articulateds had a reputation for being slippery on the front engine. Another inch of cylinder diameter, all else unchanged, would have taken the FA down to around 3.9, a bit on the low side. Or, the weight on drivers would have to be increased to around 513 000 lb to maintain the same FA. So maybe the designers were playing safe.
Turning away from statistical comparisons, here's something I'd like to see from someone qualified knowledgable enough to address it. I'll state it in the form of an "essay question".
"In the context of the immediate post-WWII period, compare and contrast, in a semi-quantitative way, the relative utilities of the following articulated steam locomotives (list below). On what kind of operations - loadings, topograhies, speeds - would each be optimally and adequately deployed. What kind of operations is each definitely unsuited for? For these purposes consider actual US rail routes, and take into account interaction of locomotive and track. What special requirements in respect of facilities, operating procedures and/or maintenance procedures might be needed for specific locomotives? For the purposes of this exercise, assume that main lines generally will accommodate axle loads up to 70 000 lb, with higher values requiring special construction. Also consider what circumstances might indicate the use of a non-articulated locomotive (suggested list below) rather than say an x-6-6-y articulated. Incremental or evolutionary changes to any of the locos may be proposed provided the reasons for and expected benefits from these are supported. Highlight any individual locos that lack any distinct, or have a very narrow utility benefit as compared with all of the others."
The list of articulated locomotives is: BB, Allegheny, N&W A, DM&IR 2-8-8-4, N&W Y6, UP "big" Challenger, late "small" Challenger (D&H or Clinchfield variant). The suggested non-articulateds for consideration are: ATSF 2-10-4, C&O 2-10-4, and KCS 2-10-4. (Apologies to anyone if your favorite is not there!)
OK, so that's a "tall order", likely a pipedream, but if it could be done by some steam locomotive enthusiast/genius out there, I think that it would tell us a lot more about these locos than statistical analysis!

)