The 1e AD&D DMG random treasure generator is extremely stingy on magic items:

Taking G1 there's 28.8 average (accounting for D#23, and counting the javelins as 1) in 58 areas, which is about 50% per room.

In the DMG the random treasure generator gives you 3% chance of a magic treasure in a treasure found by itself, which is 1/20th of rooms found, and 2 chances of 13% in a room with a monster and treasure which is in 3/20ths of each room found. That's a grand total of 4.5% of rooms having a magic item on average, That's less than 1/10th of G1. (checked OSRIC, and it's even worse with 10% for with monster instead of 13%)

Now there is some chance of getting NPC parties with magic items, but that depends on level and is difficult to figure, most will not be hostile to the PCs though (although evil PCs may decide to kill everyone for their loot). Taking a mid level like 7, that's 2-5 (3.5) characters with 70% of 3, 50% of 2, and 10% of 1, some of those have chances for 2, taking that into account it's 4.04 magic items each, and a number of 4th level henchmen equalling 4-7 (5.5) with 40% of 2, and 20% of 1 with the chances of 2 that comes to 1.3 each, total between the party averaging 21.29 magic items (whew that's quite a bit!). If we take into account the chance that an NPC party will be encountered.. is around 8% per monster encounter, so 1.7 magic items average per encounter, and .42.5 per room brining our total up to 47%. That's actually fairly close. But if you consider hostile parties it'd be more like 10-15% depending upon how hostile the NPCs are, and also upon the level.

If we go by first level it's 10% chance of only the 2-5 1st levels having 1 or 2 (1.35), or 40% per NPC party of one item, there's also only a 3% chance of NPCs per monster encounter on the first level, so that only ads 1.2% per encounter, which is only 0.03% per room, a negligible difference.

That at least explains why my party going through the random dungeon has almost no magic items, and more than half have come from the very few hostile NPC parties, my party also has a 16 Cha character, so hostile NPCs are rarer than they would be with a party with average charisma, only two or three over 7 levels, and one of those TPKed them. It seems like a bad design that the random dungeon would punish high Cha parties so much.

I think to properly compare we really need to compare the number of occupied rooms though - in the random dungeon that's 5 in 20 has a monster, 1 in 20 has a treasure by itself, and 1 in 20 has a trick or trap. So 7 in 20, which brings our baseline up to 12.8% Before NPCs. Looking through g1 it looks like 7 of the first 25 are empty except for incidentals, and likewise 8 of the 33 in the dungeon, so 43 occupied areas with 28.8 magic treasures or 67%.

Now the question is how to fix this?

I know some people use lair treasure with the random dungeon too when the numbers encountered qualify, and I've tried that, but I haven't noticed much of any difference, as most don't qualify for lairs, and not many lairs have magic treasure. So I'm going to ignore that for the moment.

The easiest way would probably be to just put a multiplier on the number of magic items found when they are found. Only thing is it would have to be about 5x the current number. It seems like it would be a bit extreme to come across caches of 5 magic items though every time. Could change that to 2d4 magic items. That seems more reasonable. I think I'd rather see the average number found around 3 though - 2d3-1 or d10/2 would work. We need to up the chance to be found by about 1.6 times though. My first thought is to give a +1 to the roll per level, which I like as it gets rid of some of the copper at higher levels, which seems like it needs to be adjusted out anyway, and there does seem to be more magic in higher level modules. At a comparison level 7 would then be 10% for 1/7th of treasure rooms and 40% (20% x2) for 3/7th of rooms with monster and treasure, which gives 18.5% x3 = 56%.

I need to do that with a low level module, I'll take B2 since I have it handy though T1 or U1 would probably be more appropriate being actual AD&D modules. We have 56 rooms of interest, I've had some trouble figuring number of magic items. this is a 3rd edit, it's actually 32, or 66 if you count all the amulets of protection from good and protection from turning. That comes to 56% or 117%.

Our % for an average of lv 2 in the random dungeon with +1% per level and 3 average magic items when they are found comes out to 40%. Still not bad, might need a slight boost in each case though. Adding a flat +3% gives 49% for our lv 2, and 64% for lv 7, lv 8 might be a better comparison for G1, so that would be 68% - really close.

1.) Roll 2x on Table V.F for each chamber or room. (less empty rooms, possibility of 2 monsters)

2.) Table V.F. if a Chamber on a roll of 12 check for Pools on Table VIII.A.

3.) Add +1% (or more?) to the roll on the Treasure Table V.G. for each level of the dungeon. (less copper more magic the deeper you are)

4.) When getting the result 'Magic Item' for treasure on Table V.G., use table I. Map or Magic pg 120 or Magic II.B (results in a lot more consumables)

5.) When getting result Trick/Trap on V.F Chamber or Room contents roll d3, 1 = use table VII as normal, 2 = use Appendix G Traps instead, 3 = use Appendix H Tricks instead.

My DMG I hand numbered appendix H for randomness and I got some pretty cool & weird things from that.

You might also want to do something with random monster encounters to make them occasionally something different from the standard tables. Like you might find a room with a whale in it. This would probably require a lot more work though. Perhaps just a list of all listed creatures in the MM with #'s 1-100, with a 1 in 10 chance it'll be from that table instead of the DMG table. The advanced version would be to put all creatures on the dungeon tables based on their XP value, and perhaps give it a 1 in 2 or less chance to use that table instead so all the encounters aren't weird.

I also recommend ignoring the part about NPC parties being of 7th to 12th level on dungeon level 6 as I wiped a party that way with 12th level evil NPCs on the 6th level. Just use the normal determination of party level.