Your D&D Encounter Maths Is Fine. The Fight Is Still Boring.
I have built hundreds of D&D encounters that were, by every number available to me, correctly balanced. A fair share of them were also completely forgettable, and it took me an embarrassingly long time to accept that those two facts have almost nothing to do with each other.
Disclosure before I start having opinions about software. I build CharGen, an AI toolkit for tabletop RPGs. It is not an encounter builder and I am not about to pretend otherwise. Most of what follows is about tools I did not write, because the honest version of this subject hands the maths to other people and then talks about what happens next.
The maths was solved years ago
Encounter budgeting is one of the few genuinely finished problems in D&D prep. Kobold+ Fight Club still does it faster than anything else I have tried. You set party size and level, filter monsters, and watch the difficulty rating move as creatures go in and out. Its 2022 update added random modes for bosses and minions and hordes, plus a history of what you have already built, and it now carries separate 2014 and 2024 generation strategies so you can choose which era of rules you are budgeting against.
D&D Beyond Encounters covers similar ground with a combat tracker attached, which is worth quite a lot if your characters already live there. Donjon has looked dated for about a decade and remains the quickest way to answer “the players have stolen a ferry, what is in the river”, with environment filters for underdark or swamp or coast. Shieldmaiden is the one I reach for when a fight will be long and full of conditions. CritForge remembers what a party has already fought across a whole campaign. Vondy writes flavour text around a skeleton somebody else has balanced.
Between them that is a solved category. Nobody needs another CR calculator, mine included. I have watched DMs open three of those in tabs and produce a fight that was mathematically immaculate and emotionally flat, which is roughly where I spent my first couple of years behind a screen.
One real trap worth five minutes
There is a genuine bug in the current tooling and it is cheap to avoid. A difficulty label only means something relative to the version of the game your table actually plays, and the D&D Beyond Encounters beta still calculates against the 2014 Dungeon Master’s Guide rules even though the rest of the platform has moved on to 2024 content. Same party level, different assumptions underneath, no warning on screen.
So write the ruleset into your prep notes before you build anything. Then check the four things no calculator can see. Whether the cleric has any spell slots left after last week. Whether the stat block you picked is current or a deliberate legacy pull. What the terrain does to positioning, which in my experience swings a fight harder than adding a creature ever does. And how the encounter ends other than by everything on one side reaching zero hit points.
Get Nvorontsov’s stories in your inbox
Join Medium for free to get updates from this writer.
That last one changed my games more than any piece of software has. I ran five level 5 characters through a flooded observatory a few months back, one player absent, cleric nearly dry. The 2014 sums said it was comfortable. Running the same build against the 2024 strategy pushed me to cut a creature and add an escape route through the dome, and the whole thing ended on a decision rather than a stairwell grind. I am still not certain the 2024 numbers were right. I am certain the escape route was.
What none of them can do
Every tool in that list stops at the point where the numbers are correct. Which is a perfectly reasonable place to stop, if you think of an encounter as a budgeting exercise. Players do not experience a budget though. They experience a room, a thing in it that wants something, and some reason to care which way it goes.
A correctly costed manticore is a line in a spreadsheet. A manticore on a rain-soaked harbour, perched on a crane, that will break off and leave if you set the nets alight, is a scene. The stat block did not change between those two sentences. Everything else did.
This is the gap I ended up building into, so weigh the next paragraph accordingly. Once a fight is balanced somewhere else, I use my own monster generator to give the lead creature a face and a motive, then draw a board in the battlemap tool with cover and a choke point and one hazard, then crop the tokens out for the VTT. That is a step after the maths, not a replacement for it, and I would rather say so plainly than blur the line to sell you something.
The half hour I actually spend
My prep budget for a single fight is about thirty minutes and it splits roughly like this. Five minutes testing two or three shapes of the same encounter. Five minutes on the reality check above. Eight minutes designing the lead monster until it has an identity somebody could describe. Seven on the map. Five cropping tokens and writing down the one gimmick I want to remember mid-session.
Then I stop, or try to. The final twenty percent of prep adds far less than it feels like it will while you are doing it, and I have thrown away entire evenings polishing a room the party walked straight past. My own record here is poor, in fairness. I still get pulled into fiddling with lighting on a portrait nobody will look at for more than four seconds.
The uncomfortable part
I sell tools for the visual half of this, so apply whatever discount you think is fair. None of it rescues an encounter that has no point. A gorgeous battlemap of a fight that exists because the session needed a fight is still a fight that exists because the session needed a fight, and no amount of art papers over that. I have made this mistake with better art than most people have access to.
If I were starting over I would spend the first month ignoring generators entirely and writing one sentence per encounter about why those monsters are in that room and what would make them leave. Borrow the maths. Six good tools will do it for free in under three minutes. The reason is the only part of the job nobody can hand you.
If you want the tool-by-tool version with limitations and free tiers, I keep a roundup updated on my site. But the thirty minute split above is most of what I really use, and the sentence about why the monsters are there is most of what makes it work.
