So I'm still looking seriously at Holmes D&D as a basis for a campaign, and digging up my megadungeon maps and working them into a new, bigger design. Particularly I think Holmes is more suited to the campaign dungeon than the wilderness sandbox, and I'm primarily interested in the dungeon environment anyway. (I'm not much of an artist, but I certainly can make dungeon-lookin' shapes on a piece of graph paper.)
Still, there are a few concerns with Holmes, mostly in the short combat section. Initiative based on Dexterity - I leave that to long time Holmes fans to make the case to me. Honestly rolling 3d6 and counting down Dex scores sounds cumbersome compared to my preferred initiative method, which is each side rolls 1d6, high roll wins it, ties occur simultaneously. (This blog is, after all, named after a long run where the PCs lost the initiative in every fight.) But more worrisome are the rules for weapon speed.
Basically, daggers are said to roll twice per round, most one-handed weapons go once, and two-handed weapons go once every other round. There is no "weapon type vs. AC" chart, and everything does d6 damage - so a two-handed sword means you're rolling to hit 1/4th as often as a dagger with no benefit. In fact, everybody but clerics (who have to use maces) should be combat gods according to this comparison.
I've been thinking about raising the damage dice for other weapons and lowering it for daggers, but that still doesn't work. To actually have a better average effective result as a dagger, with the dagger doing 1d4 twice (average 2.5 x 2 = 5), the long sword would need to do 1d10 (average 5.5) and the two-handed sword would need to be rolling a d20 (average 10.5 / 2) for damage.
Looking at the actual data, I think the solution lies in relative effectiveness, which can be done without the high degree of complication present in Greyhawk or AD&D. If you say that daggers do 1d4 damage and suffer a -2 penalty against characters with AC 7 or better, they are deadly effective against unarmored characters, but their impact goes down dramatically with lower ACs. If you say that 2-handed weapons have 1d12 damage and take the opposite effect, treating AC 7 or better as two categories higher, e.g. AC 7 becomes AC 9 and so on, they become more effective, on the whole, than standard weapons in the d6/d8 range. I'm envisioning 4 "classes":
Light weapons - daggers, mainly - 1d4 damage, Fighting-Men and Thieves get 2 attacks / round, -2 to hit versus AC 7 or better.
Normal weapons - maces, spears etc - 1d6 damage, 1 attack per round.
Heavy weapons - swords and similar - 1d8 damage, 1 attack per round.
2 Handed Weapons - two handed swords and polearms, 1d10 or 1d12 damage, 1 attack per 2 rounds, +2 to hit versus AC 7 or better.
Holmes does have one other "hole" in the combat system - specifically it says that Fighting-Men get more attacks but doesn't go into detail. I'm considering that at Hero level, they get 3 attacks per 2 rounds, with the 2 coming in round 2, with normal weapons, 3 attacks per round with light weapons, and 3 attacks per 4 rounds (essentially skipping a "rest" round) on 2-handed weapons. When they reach Super-Hero, this goes to 2, 3, and 1 respectively; I realize that the consistent thing would be 4 dagger attacks per round but that's just silly. Holmes rounds are 10 seconds long, not the 60 second abstraction of AD&D.
I think I like the dynamics of the house rules listed above. It doesn't complicate things too much, it differentiates weapons without giving too much advantage to one type or the other, and it's mathematically sound. Any thoughts?