why the 5e treasure tables aren’t very good (and how they got that way)

I’ve been running D&D 5e dungeon crawls lately (testing my Inspiration app) and one thing jumps out: by-the-book treasure isn’t nearly varied enough.

treasureguysI’m not talking about magic items: sure, I’m a little sick of potions of giant strength, but D&D has a generosity with its magic item list that I very much appreciate (more than 300 items in the DMG!). I’m talking monetary treasure.

This problem doesn’t come up much in my regular D&D campaigns, because neither I nor my co-DMs use the loot tables in the 5e DMG. After all, there’s not much to spend money on in non-houseruled 5e, so you can make up literally any amount of money and it makes no difference to game balance.

Even without much to buy, discovering treasure still is (or should be) fun. As a DM, I’ve always liked rolling on treasure tables, and as a player I like uncovering exciting caches of gold and jewels. But when I stick to the letter of the law, as I’m doing in this playtest, the defects of the 5e treasure tables become apparent.

In the course of a recent dungeon delve down to level 3 of a random dungeon, I’ve written down the treasure hoards we’ve found. Here are a few of them:

On level 1, we found: 2500 CP, 900 SP, 90 GP, 6 50-gp gems.
On level 2, we found: 2100 CP, 1100 SP, 80 GP, 6 50-gp gems.
On level 3, we found: 2100 CP, 1100 SP 60 GP, 5 10-gp gems.

When I listed them out like this in my notes, it became clear to me that, apart from magic items, every treasure I found was going to look exactly the same. Around 2000 copper. Around 1000 silver. Around 100 gold. Maybe 6 or so handful of gems or art objects. Every. single. time.

There are a few main problems with cash rewards in 5e:

Lumpy progression.

There’s not a new chart every level. You roll on the exact same treasure chart for all encounters from CR 0 to 4, and then graduate to the next chart up, which you use for six more levels, and so on. Each treasure table generates a coin value ten times the last one – a huge jump. It’s too big a jump, too infrequently. If the devs had found a way to divide this increase evenly over every level, then each level could have been worth 50% more than the last: a noticeable increase each level, and one that gives meaning to choices about challenging an easier or harder foe. (Right now, four bugbears have the same expected treasure trove as four kobolds. Level 3 of the dungeon has the same treasure as level 1. Why venture deeper?)

Samey treasure.

Every treasure hoard of a given tier group has exactly the same coinage mix. For all of CR 0-4, it’s impossible to roll up a treasure hoard containing electrum or platinum, and it’s impossible to roll one WITHOUT about 2000 copper, 1000 SP, and 100 GP. There’s a bit of variation when it comes to the gems and art objects rolls, though you will never find a solitary gem or art object (the smallest number of either you can find is 2d4, and the average is 7).

You’ll also never find a cache of gems or a magic item without some attendant cash. Every randomly rolled treasure must contain money. After the Lady of the Lake raises her arm from the water to give Excalibur to Arthur, she also tosses 42,000 gp and 28,000 pp ashore.

There are more interesting variations in the “individual treasure” tables, but unfortunately that’s all pocket change.

No big scores.

There’s little roulette-wheel drama in 5e monetary treasure: no sense that you could stumble on a big score (except by DM fiat). The only meaningful variation is in whether you get gems or art objects, and how many of each. On each treasure table, the best gem or art object result occurs about 20-30% of the time, so it’s not particularly unusual; and apart from the lowest tier of play, the best possible gems/jewelry roll can increase total treasure value by about 10%, so it can almost be ignored.

The ninja rule.

you know what? The minor point that I mentioned before, that you can’t roll up a single gem or art object, is actually a big problem for me. A single well-described piece of jewelry or idol, or a fist-sized emerald with a strange flaw, could grab someone’s imagination. However, gems and art objects generally come in packets of seven. Rolling up and describing seven art objects is an exercise in tedium for the DM and player.

DM: … encrusted with rubies; and, finally, a goblet etched with a baronial shield and the motto EVER TRIUMPHANT.
PLAYER: ….. ok, so… seven art objects worth 2500 each, get it. (instantly forgets every detail)

Since gems and art objects are just interestingly-described cash, there’s no point in presenting them in numbers that repel description. I might make up one or two interesting art objects that give flavor to my setting, but not seven.

Why is 5e like this?

I’m not sure exactly what was going on with treasure tables during the development of 5e, but I think that most of the problems stem from a single design decision: making the treasure tables require only one percentile roll to generate the whole treasure.

Take a look at one of the treasure tables, the one for CRs 11-16.

Screen Shot 2020-06-23 at 4.41.41 PM

After a single d100 roll, you know almost everything about the treasure: what types of coinage you get; the number and denomination of gems and jewelry; and what magic item tables to roll on. Compare that with, say, the 3rd edition level-11 table, which required 3 d100 rolls: one for coin type, one for art/gems, and one for magic items.

Screen Shot 2020-06-23 at 4.50.46 PM

I think the decision to go down to one d100 roll in 5e had to be a deliberate choice. But it was one with some bad consequences.

No roll for coin type

In the 3e chart, you roll to see if you get silver, gold, or platinum. (It’s a little weird that you can’t get a mix of coins, but whatever.) Each of the possible coin types had a slightly higher average value, so rolling a higher-denomination coin was better.

In the 5e chart, you always just get gold and platinum, no roll required, and no variation between treasures.

Combining the art/gems and treasure roll

Look how big the 5e 11-16 treasure is! There are 33 rows. It takes up almost a full print page. And if you look closely, it’s not super varied. The 32 rows of the “gems or art objects” column are just alternating the same four entries over and over again, eight times. The 28 rows of the Magic Items columns are just repeating the same seven entries four times each.

In fact, this table looks like it was derived by listing out all the permutations of two much tidier tables. These tables don’t exist in print but I’m confident they looked like this:

gems and art objects: roll 1d20:
1: nothing
2-6: 2d4 (5) 250 gp art objects
7-11: 2d4 (5) 750 gp art objects
12-15: 3d6 (10) 500 gp gems
16-20: 3d6 (10) 1,000 gp gems

magic items: roll 1d20:
1-3: nothing
4-6: Roll 1d4 times on Magic Item Table A and 1d6 times on Magic Item Table B
7-12: Roll 1d6 times on Magic Item Table C
13-15: Roll 1d6 times on Magic Item Table D
16 Roll once on Magic Item Table E
17 Roll once on Magic Item Table F and 1d4 times on Magic Item Table G
18-19 Roll 1d4 times on Magic Item Table H
20 Roll once on Magic Item Table I

If you map the permutations of these two charts onto one d100 chart, it perfectly recreates the existing 5e table, bar rounding error.

The big problem is, the two charts, while they require an extra roll, are more economical of space. If the two tables were printed side-by-side, as was done in 3e, you’d only use 8 rows instead of 33.

page count matters

Combining every treasure and art/gems roll into a single roll matters a lot because it balloons the space requirements for a single treasure table. And, since book pages cost real money to print in an already fat book, you just can’t include a lot of these huge tables.

It takes more than half a page, on average, for each of 5e’s four treasure hoard tables. If you had one such table per Challenge Rating from, say, 0 to 21, that would mean devoting 12+ book pages just to treasure charts. That would mean cutting 10 pages of other content.

On the other hand, if they’d used separate d100 rolls for art/gems and magic items, as in 3e, they could probably have fit 22 such charts in around 3 pages.

So, once they’d decided to present giant treasure tables that required only a single roll, there was a real-world economic incentive to cut down the number of tables to four, one per tier.

Does limiting treasure tables to one d100 roll even make sense?

I’ve posited that the developers made a decision to limit monetary treasure generation to a single percentile roll. Is that even a valuable goal?

Not in my opinion.

First of all, once you make your percentile roll, there are still secondary rolls required to determine how many coins, gems, and art objects you find: 4d6x1000 gp, 3d6 500-gp gems, etc. The averages are given: “4d6 x 1000 (14,000) GP”. But you have to roll, or otherwise vary, these amounts. You can’t just take 14,000 every time.

According to the DMG: “Over the course of a typical campaign, a party finds treasure hoards amounting to seven rolls on the Challenge 0-4 table, eighteen tolls on the Challenge 5-10 table, twelve rolls on the Challenge 11-16 table, and eight rolls on the Challenge 17+ table.” Imagine if a DM didn’t roll, and used average treasure for all of these monetary rewards. From around levels 5 to 10, the party would find eighteen consecutive treasures with exactly 700 CP, 7000 SP, 2100 GP, and 105 PP. The treasure coinage is already noticeably lacking variety even if you do roll; if you don’t, it would be comical. The same is true for gems and art: you’d always find either 5 or 10 of each.

So, assuming we are rolling for coinage and gems/art, we’re always making three or four extra rolls in addition to our percentile roll.

And that’s not even counting magic items! Look at the 5e CR 11-16 chart above. There is a lot of percentile rolling here! Some of the entries call for rolling up 1d4 magic items from one table, and 1d6 from another. Each of these d4 and d6 rolls requires further percentile rolls, so, on average, this table entry requires an average of eight rolls, most of them percentile rolls. Taking the table as a whole, you make an average of three and a half extra rolls to determine magic items.

In total, a DM is actually making around nine rolls to determine each CR 11-16 treasure. Saving that one extra percentile roll didn’t make much of a difference.

looking backwards to go forwards

So now that I’ve exhaustively detailed my problems with 5e’s monetary treasure tables: can we do better?

I think so. I’d like to scrap the four 5e treasure tables and make (at least) twenty new ones. I think the in-game economics of 5e are fine as far as they go, I just think the tables could be presented better and the lumpy parts smoothed out.

In my next post I’ll look back at previous editions to see how they did treasure, and we’ll see if we can find the best model for how we’d like 5e treasure tables to look.

3 Responses to “why the 5e treasure tables aren’t very good (and how they got that way)”

  1. Ios says:

    Very good post! looking forward to your solution

  2. Randy says:

    This is so true! I find it enjoyable as a DM to give out treasure to my players. I love it when they try to figure out where things came from or I can tie an interesting piece into the story. I have had to “manually” spice up treasure to make it interesting and it becomes a challenge to come up with something new and interesting each time. To your point, I would like to see new treasure tables, too.

  3. Darrow says:

    Sorry to post this on another article, but could you help clarify/ expand on your damage budgeting section of ‘5e monsters on a business card’ post? I’m writing up a bestiary and have mostly been using your methods but that one part keeps confusing me.

Leave a Reply