More Dungeon Treasure

Looking at the OD&D random dungeon treasure table from Vol-3:

Consider that there are 7 rows/"tiers" to that table. (Tier 1 = level 1, Tier 2 = levels 2-3, etc.). I randomly generated a sample of 1,000 treasures from each tier. Statistics are as follows:

And compiling the relative frequencies into a chart shows us this (note that the x-axis is quasi-logarithmic, using the 1-2-5 preferred numbers series):

Note how incredibly skewed the first two tiers are (i.e., dungeon levels 1-3). For example, at level 1 the median treasure is only 60 gp (in OED games we translate this to sp, but costs are all cut down in proportion, so purchasing-power is identical). 85% of the time you'll get a treasure less than 100 gp, which adds up to negligible XP for the party. But if you roll positively on the 5% gems/jewelry chance, then that treasure will be likely in the 1,000's of gp's (even possibly on the order of 40,000 gp maximum!). If you were to roll consistently and faithfully using this table then it's possible that you never get any large-value treasure in a given 1st-level dungeon (granted that by the book less than 1/3 of rooms have any treasure at all); or alternatively, you might have one outstanding treasure worth 20,000 gp sitting by some zombies or giant frogs (as happened the other day when I was doing some rolling experiments).

I think this highlights the need for the DM to intentionally place one or two "big hauls" on the first few levels of the dungeon, including gems and/or jewelry, in order to support PCs advancing levels at all before hitting lower levels of the dungeon. I think that Paul S. would further suggest that players be made aware of this from the get-go as an initial quest/mission/ante to get the action started. If one accesses the more advanced tiers, the gem/jewelry frequency increases, and therefore treasure gets a lot more evenly distributed at those levels (that is, by tier 4/level 6, about half of all random treasures have either gems or jewelry or both), which might indicate that we can then rely on the random method more for a useful amount of treasure.


  1. Heh, It reminds me of the Keep on the Borderlands group. They got fortunate in the cultists cave and made off with the Ruby eyes right off the bat. Then were hitting piddly goblin treasure for the next few weeks.

    1. Wow! Totally reversed from we'd expect. Interesting effect of the very-open layout.

    2. It does make me appreciate that "Gygaxian Naturalism" needs a nudge now and again.
      I think its why I enjoy this series of discussions so much.

  2. I dunno Dan, I don't think the tendency toward lesser treasures on the first couple levels matters a whole lot given that this table is really only for unguarded treasures. If there is a monster lairing in the room, the monster's treasure type should be used. Check out the article I wrote in & Mag 13, more on that if you like.

    1. Maybe in other non-LBB writings, but there's no such distinction on Vol-3, p. 7. Most writings indicate the monster treasure types are for outdoors only; and if they are used, and pro-rated then overall treasure is even less. (link)

    2. MM p. 5: "The use of treasure type to determine the treasure guarded by a creature in a dungeon is not generally recommended", and other indicators.

  3. I think a reasonable skew is a good thing, it gives you those "wow, score!!" moments. But maybe it's a bit too much for tier 1. Maybe 1 in 6 are good (but not quite so incredible) treasures?

    Do you have any ideas for improving the random tables? And how do you calculate the expected value of a magic item roll?

    1. I very much agree with that. Sadly, at the moment I don't have any fixes other than the DM hand-placing some "special treasures" on level 1; and also no analysis of magic at the moment (which likewise seems as thin as jewelry per the table).

  4. Another suggestion instead of deliberately posting treasures: double the odds of gems and jewels but reduce the number to 1d3. Easy to do, and evens things out a bit.

    next step to take a look at -- compare the typical loot and thus xp against the xp tab les. I found that the number of loot hoards needed increases up to a max around level 8, then drops off. Your chart shows the major leap between tier 4 and 5, which is where progression stalls out.

    I think the number of loot hordes needed to level is good for the dm to consider when designing the floors of a dungeon... Both to ensure enough rooms and to place a few key loot hordes.

    1. Gah, sorry for the awful autocorrect there from my phone. I'll post the tables i calculated when I'm on a computer.

    2. Here's the rough number of loot rooms needed to generate enough XP for a character to level. The baked in assumption is that 20% of XP is gained from monsters.


      - #s above are for 1 character harvesting 100% of available XP.
      - I suggest at least doubling the numbers above to account for missed treasure and/or henchmen and/or XP lost to character death.
      - Multiply the above numbers by the # of chars in the party.
      - Finally, to determine the size of your dungeon floor, remember only about 1/3 of rooms have treasure. So you can multiply the # of loot rooms by 3 to determine the desired size. You can get away with a smaller dungeon floor by deliberately placing a few extra loot rooms.

      Example: A party of four player characters is approaching the fifth floor of the dungeon. The GM determines that 14 loot rooms are needed to provide a minimum of XP opportunities to level up (3.5 x 4). Using the rule of thumb above, the GM doubles this figure for a map with 28 loot rooms, more or less (14 x 2 ROT). The GM decides to place twelve treasures personally first -- a double loot stash for each character, plus a large hoard consisting of quadruple the normal loot (with proportionate guardians!). The GM also plans to award each character a bonus XP award equal to a loot room for vanquishing a mighty villainous monster on this level. That leaves 12 loot rooms to distribute (28 - 12 - 4), more or less. The GM should plan on randomly generating 36 rooms to ensure sufficient treasure is available. This will provide about 14 hours of play.

  5. Hey Delta, I thought you might enjoy this. I made a spreadsheet similar to what it looks like you've made for this post, and then massaged the numbers until I was happy with the results. Basically, I found that by: a) gradually moving from copper/silver to gold/platinum coins, and b) increasing the value of gems and jewelry by level, I was able to maintain around 78% XP by treasure, 39% XP by jewelry, 22% XP by gems, and 82% non-encumbering XP (i.e. monster kills, gems and jewelry) per level. Can't wait to roll up some random dungeons using these numbers. Thanks a lot for the analysis and inspiration! https://www.dropbox.com/s/qov0kp0r07mlh2r/D%26DTreasure.xlsx?dl=0

    1. Thanks for sharing that! Any update from your random dungeon testing?