What D&D's encounter math doesn't tell you

"Is this encounter balanced?" sounds like a question with one answer. It isn't — not because the maths is vague, but because there are two official versions of it that can look at the same four monsters and the same four characters and land on different difficulties, and a separate piece of maths for statting your own monster that doesn't agree with the Monster Manual's own numbers. None of this is a bug. It's worth knowing exactly where the disagreements are, because they change what you should actually do with the number a calculator hands you.

Two methods, same fight, different answers

Take four 3rd-level characters against four CR 1 monsters — a believable Tuesday-night fight, nothing exotic.

Under the 2014 Dungeon Master's Guide method, the four monsters are worth 800 XP raw, but four monsters pushes the encounter multiplier to ×2, so the number you check against the difficulty thresholds is 1,600. Four 3rd-level characters' deadly threshold, summed, is 1,600. That fight lands exactly on deadly.

Under the revised 2024 rules, there's no multiplier — you compare the raw XP straight to a budget. Four 3rd-level characters' budgets, summed, are 600 (low) / 900 (moderate) / 1,600 (high). 800 raw XP clears the low budget and nothing past it. Same monsters, same party, and the honest answer goes from "possible TPK" to "barely notices."

Neither method is wrong. They're measuring difficulty with different instruments — one inflates for monster count and checks it against a threshold, the other folds the crowd effect into the budget number itself and checks the raw total. The trap is assuming "balanced" is a single number you can look up once and reuse across both.

The multiplier is where the surprises hide

The 2014 method's multiplier ladder is ×1 for a single monster, ×1.5 for two, ×2 for three to six, ×2.5 for seven to ten, climbing from there — and a small party (fewer than three) pushes everything up one step, a large one (six or more) pushes it down one.

The part that catches people: the multiplier applies to the whole encounter's XP, not just the monster you added. Two CR 1 monsters sit at ×1.5 — 400 raw XP becomes 600 adjusted. Add a third CR 1 monster and you don't just add its 200 XP; you also cross into the three-to-six band, so the multiplier jumps to ×2 and the adjusted total becomes 1,200. One more goblin, worth 200 XP on its own, just doubled the number you're checking against the difficulty thresholds. "Add one more" is the most dangerous sentence in encounter-building, and it's the multiplier doing it, not the monster.

The XP you judge difficulty with isn't the XP you hand out

That adjusted, multiplier-inflated number exists for one purpose: classifying how hard the fight is. The XP you actually award afterwards is the raw total, split across the party, full stop — the multiplier never touches it. Do this by hand and it's an easy place to quietly overpay a party, because the "deadly"-looking number sitting in your notes is the wrong one to divide up. If a tool's output only shows you one number, check which one it is before you use it for the reward.

Don't benchmark your own monster against the Monster Manual

Building a custom monster is a separate piece of maths: a defensive CR from hit points and AC, an offensive CR from damage and attack bonus, averaged into a final rating. It's a faithful implementation of the Dungeon Master's Guide's own method — and that method does not reproduce the Monster Manual's own numbers. Run the published stat blocks through it and most come out around one rating low.

An ogre is the clean example. 59 hit points and AC 11 gives a defensive rating of CR 1/4 — the hit point table expects a CR 1/2 monster to have roughly that much durability, and an ogre falls well under even that. 13 average damage at +6 to hit gives an offensive rating of CR 2. Average the two and the method says CR 1. The Monster Manual says CR 2. A troll, an owlbear, a bugbear all show the same pattern to varying degrees — published monsters tend to carry fewer hit points than the defensive table expects for their rating, so the defensive half drags the average down underneath where the monster actually plays.

The fix isn't to distrust the method — it's to stop using "does it match the Monster Manual's number" as the test. Use it to rank your own homebrew against itself, consistently, and it'll tell you true things: which of your two monsters hits harder than it can take, which one is over-tuned for its apparent rating. Use it to predict a specific published CR and it'll quietly undersell almost everything in the book by about one step.

What to actually do with this

Pick one difficulty method for a campaign and stay on it — switching between 2014 and 2024 mid-campaign means your sense of "this always runs about medium" stops meaning anything. Watch the monster count, not just the total XP, since that's what moves the multiplier. Award XP from the raw total, not the number you used to judge difficulty. And when a homebrew monster's calculated CR comes in under what you expected, check whether you're comparing it to a published stat block before you start adding hit points to compensate — the gap is usually the method's, not the monster's.

The encounter difficulty calculator runs both the 2014 and 2024 methods on the same fight side by side, so you can see where they disagree before you're mid-combat finding out the hard way. The CR calculator does the defensive/offensive maths above for a monster you're building from scratch, worked examples included.