... now with 35% more arrogance!

Tuesday, November 23, 2021

Gems: A Short Series

I’ve been meaning to do another movie review for a while, so obviously it’s time to start a series of posts discussing dungeon treasure stocking instead. Specifically, how to handle gems. It’s inspired by this OD&D forum post that’s been going on for that last month or so.

How do you assign gems to treasure? More specifically, how much information do you include, or should you include? Full description of each gem and quantity, minimum information necessary, or something in between?

I lean towards the minimum, with a few extra details. But before I get into that, I need break down the steps to assigning gem details. Loosely based on the “official” process, the steps are:

  1. Check if there are gems in a treasure cache or trove.
  2. Check the number of gems.
  3. Assign base gem value.
  4. Check how many gems are of higher or lower value.

If randomly assigning treasure, Steps 1 and 2 are handled as part of the treasure table in either Volume II: Monsters & Treasure (page 22) or Volume III: Underworld & Wilderness Adventures (page 7) depending on whether it’s a wilderness treasure or dungeon treasure. Step 3 involves a roll on the Gem Base Value table on page 40 of Monsters & Treasure, followed by a d6 roll for each gem or group of 5 to 10 gems as Step 4.

In the official process, you could argue that there is a fifth step, “Record the info, along with extra details like gem type or color”. But the books do not actually say that all four steps must be done at the same time, before the dungeon key is completed.

I would argue that it’s easier to do Steps 1 through 3, record the info as Step 3.5, and put off the final roll for Step 4 when each gem is appraised.

So here is my suggested gem generation process for GMs stocking treasures:

Step Zero: Pick a Dungeon Theme

Picking a theme is of course is part of the dungeon creation process, rather than the gem generation process, but what kind of dungeon you are creating and where it is should, logically, affect what kind of gems might be available. One trivial example is pearls, usually treated as if they were gems. Underwater treasures or coastal treasures might reasonably include more pearls than someplace far from an ocean. Other gem types could be more or less common depending on the region: if jade is more common in one area, dungeons in that area or connected in some way to that area might have more jade in their treasure troves than other gems.

Step One: Pick Your Gem Types

It’s arguably not immersive to describe gems to players this way:

“You find 30 gems of 100 gp value.”

It’s more immersive to describe them this way:

“You find 20 sapphires and 10 diamonds.”

But it’s crazy to roll for each gem to see what its type is, especially since the average PC probably won’t know the exact type, just the general appearance. The easier method is to assign three gem types to the dungeon as a whole, for example opaque red, murky green, and clear yellow. All gems in that dungeon will be one of those types, simplifying the next step.

Step Two: Check the Number of Gems

As you stock each treasure trove in a dungeon, you make three rolls:

  • How many gems are Gem Type 1?
  • How many gems are Gem Type 2?
  • How many gems are Gem Type 3?

You defined each gem type in Step One, so all you need to roll is a number. Gem Type 1 will be different for each dungeon. There is no table lookup for this step.

Since we can subtract a number from the roll and discard any result of 0 or less, we can fold “Check for Gems” into this step as well.

Step Three: Revealing Details to Players

For reasons I’ll explain more in a future post, each of our gem types has a predefined base value. We actually don’t need to roll any more as part of dungeon key creation. We just need to tell the players “You find 10 gems that look like this, 8 that look like this, and 2 that look like this.” And there’s only three "this"es per dungeon, so players can just keep a running total of each gem type found.

If there’s a dwarf in the party, or someone with gem appraisal skills, they can find out more details, including the exact value. Or they can have the gem appraised, or just sell it blindly without knowing the details.

I plan on detailing each of these modified steps in a separate post, with appropriate tables. I won’t be posting before next week, however.

Series Index

  • Gems Intro (this post)
  • Gems I : Types
  • Gems II: Quantities
  • Gems III: Assessments
(Will edit to add links as the posts are published.)

Creative Commons license

This work is licensed under a Creative Commons
Attribution-NonCommercial-ShareAlike 4.0

(CC BY-NC-SA 4.0) license.


  1. I made a roll-every-die gem table some years ago! Could be useful.


    1. Seems like that would be an ungodly amount of rolling, especially in cases where there are 10 to 60 gems in a treasure hoard. As you will see, I go a more minimalist route.

    2. Also, I only touched on it briefly here, but I'm not at all fond of precise gem identifiers. I prefer purely descriptive (color and other visual features.) When proper gem names are needed, I stick to the 5 or 6 best-known gem names and any fictional names linked to the campaign. I explained why in the forum thread, but will probably cover it here, too.

  2. 1. I use the 30+ named precious stone names
    2. I use the 6 "type" of stones
    3. I use 8 sizes (from pea [.015 cubic inches] to apple [8.6 cubic inches])
    Have a spreadsheet setting out all of these values. When players find stones they're told they find, for example, three almond sized dark blue stones with yellow striations. If there's a thief, dwarf or someone with expertise we do a skill check. Success over 10 and I give them the type (ornamental) if over 15 I give them the name as well.

    Longer version: all my gems come from someplace and the price varies with rarity as well as size/type. All of that's on the spreadsheet as well. My players know that the pearl you find in a harbor town will be worth quite a bit more 100 miles inland! And all I need to track is stone and size. Players meta-knowledge can give them a decent order of magnitude estimate of value.

    1. I've used gem tables from the AD&D DMG before, and that is basically why I'm aiming for the complete opposite approach from that.

  3. so supply/demand and intrinsic/extrinsic value are out the window? Not sure how that's an improvement

    1. It's an improvement in that everything I've found on gem values in the real world makes it clear they are subjective and can change for arbitrary reasons. And even if I weren't taking that into account, it's easier to generate the gem value independently and then retcon that the gem's peculiarities account for its value than doing it the other way around.