1. I didn't use Holmes because it only goes through 3rd level. Since it follows 0e, there's no variation in attack rolls for those levels. (I just noticed Zenopus said pretty much the same thing in his comment on the previous post.)
2. I left some major old school offerings off the list, such as Castles & Crusades, Basic Fantasy and Dungeon Crawls Classics. I did this partly because I am less familiar with these systems but more importantly because they are not really 'clonish' of 0e, 1e or Basic/Expert.
3. I chose to look at only the first 10 levels because 10 is a round number and because it encompasses most of what I think referees and players are interested in when they look at these games. Also I just didn't feel like adding any more columns to the already somewhat squished ones of the five tables.
Now for my observations and thoughts on the results:
1. After I put the tables together I wrote an excited email to an OSR friend of mine--one of the only people I know who can occasionally be even wonkier than me on these sorts of matters. He wrote back:
Hmm... it looks like a bunch of numbers that are all very similar.
So there it is. The great anticlimactic lesson here is that, with one obvious exception, there isn't really that much difference between the attack progressions of the various editions and clones. Comparing the to hit numbers by class and level from one edition or game to another rarely yields a difference of more than 2 points or 10%.
2. The obvious exception is Lamentations of the Flame Princess. Here, James Raggi made the decision to give Fighting-Men a to hit bonus at 1st level and to increase this bonus at each level, while at the same time never giving to hit bonuses at any level to Magic-Users, Clerics or Thieves. This certainly pumps up Fighting-Men (at the expense of the others). I suppose it's either amazingly brilliant or spectacularly dumb.
3. There seems to be a big change between 0e and 1e, especially for Magic-Users and Thieves. However, part of this is illusory, or rather merely a natural consequence of expanding the normal armor class range such that it went from starting at AC 9 to starting at AC 10. Essentially 1e added an extra armor class at 6. This is because there was a spontaneous movement among the masses to give Studded Leather or Ring Mail plus Shield OR Scale Mail a unique armor class rating*. But if you expand the armor class range from eight ratings to nine ratings but at the same time keep the to hit progression at a smooth one to one, then either you must make it one easier to hit at AC 10 or one harder to hit at AC 2. Gygax elected to go with the latter. So, even though it looks like 1e makes everyone an average of one point worse in their to hit rolls, this difference would have disappeared if we had looked at AC 10 instead of AC 2.
4. 1e is quite tough on Magic-Users and Thieves--the toughest of them all if you exclude LOTFP.
5. On the other hand, if your dream is to be a melee-mad Magic-User, find a Swords & Wizardry Complete game.
6. In the differences among editions and clones, Clerics and Fighting-Men vary the least and Magic-Users and Thieves vary the most.
7. 0e has some interesting results in places. For example, in 0e a 6th level Magic-User has precisely the same to hit numbers as a 6th level Fighting-Man.
8. Basic/Expert precisely tracks 0e.
9. Of the clones (with the exception of LOTFP), all except Swords and Wizardry either precisely track their originals or closely do so. This is to be expected, of course.
10. There's a bit of controversy over what is legally required (or prohibited) in terms of emulating the attack tables of earlier versions of the original game. This is partly because the Systems Reference Document does not contain a protected template for this. Is the mathematical pattern behind an attack table a non-copyright protected rule, or copyright protected substance? Interestingly, OSRIC precisely tracks 1e in three out of the four classes (Fighting-Men differ slightly). Labyrinth Lord and Delving Deeper make small changes from their originals. But they are not necessarily random twiddles. For example, the author of Delving Deeper has a complex but logical, fascinating and satisfying rationale for why the Delving Deeper progressions are in a certain sense faithful (perhaps even more faithful than the original numbers) to the mathematical theory behind the 0e progressions, based partly on what they owe to Chainmail. For my part, in Seven Voyages of Zylarthen, I followed the 0e algorithms precisely but presented them (in my view) quite differently from the original and from the other clones. Thus, for example, instead of a chart containing rows of classes or class levels, my central attack table references weapons (which, if you will, is part of why I think I 'get away with it').
So given 1, above, what difference does any of this really make? Why did I agonize for days over my own attack progression scheme?
(No jokes please.)
With your indulgence, I'll try to answer that tomorrow.