The idea that you can measure how challenging an encounter is to a group of PCs is an old one; in current D&D, it takes form of a "Challenge Rating" (CR) that, as far as I can tell, doesn't work too well.
I don't use CR myself; I've been taking published modules at their word ("for 5-7 3rd level characters") and even when I create my own encounters I prefer being sensible than balanced (picking a fight against a small army of orcs? Yeah, you're probably going down.). The PCs can not rely on me to make sure that I'll balance the encounter if they want to face a big dragon as 2nd-level heroes.
However, I do think having some way of measuring "challenge" would be useful, at least to be sure a module is somewhat accurate when it writes "for 5-7 3rd level characters" on the cover. It is something I've been considering for a while.
The easiest way I can think of measuring this "challenge", in Old School D&D, would be simply using HD. I'm not sure just counting HD is accurate enough for existing D&D monsters (if you want a detailed analysis, try this post by Jens or follow Delta's blog), but here is how I would do it: just count HD on both sides, and it the number is similar, both sides have a more or less equal chance of winning.
Of course there are enough variable to make exact predictions impossible. But I wish we could abstract it all to HD when necessary - for example, in mass combat.
For example, take the 9 HD elephant - I'm using that because it has no special powers (no asterisks beside the HD). According to the table I used in Teratogenicon, it would have AC 14, 41 HP and cause 11 points of damage. The actual beast has indeed AC 14 but causes an average of about 15 damage due to trampling very often. In my game, a 9 HD creature has a +9 attack bonus, but in OSE it would be a +7 - the elephant has effectively about +10, however, again due to trampling.
Anyways, it is a decent approximation for my purposes.