BOMBANANA!'s Difficulty Curve: What the 30 Levels Actually Ask of You
Written October 21, 2026 · 7 min read · Based on the full 30-level table on the levels page
The usual assumption about a timed co-op game is that it gets harder by giving you less time. That is not what BOMBANANA! does, and the level data makes it obvious once you lay all 30 out and measure.
This is an analysis of the timer, mistake limit, module list and Chaos type for every Campaign level — the same table that drives our levels page. The conclusion is that the clock barely tightens from start to finish. What changes is what kind of thinking the game asks for, and how hard the individual modules are.
Method: measure seconds per module, not seconds
Raw timer values are misleading, because a 4-minute level with three modules is tighter than a 5-minute level with two. So the useful number is seconds of clock time per module: total time divided by the number of modules on the board.
Counting modules needs one correction that is easy to get wrong: an entry like "3 × Random Hard" isthree modules, not one. Once expanded properly, the pattern is clear.
Finding 1: the clock barely tightens
Averaged by phase, the time available per module is:
- Phase 1, "Build the language" (levels 1–7) — 98 seconds per module
- Phase 2, "Extend the packet" (levels 8–15) — 94 seconds per module
- Phase 3, "Preserve full state" (levels 16–20) — 92 seconds per module
- Phase 4, "Store sequence history" (levels 21–24) — 100 seconds per module
- Phase 5, "Protect the procedure" (levels 25–30) — 86 seconds per module
From the first level to the last, the budget per module moves by about 12 percent. The timer in the HUD goes from 2:30 to 5:00 — it doubles — but so does the number of modules, so the pressure per module is roughly constant.
That is a deliberate design choice, and it changes how you should prepare. If you are failing late levels, the problem is almost certainly not speed. Adding speed to a group that is miscommunicating makes things worse, because rush produces the ambiguous callouts that cost whole modules. The thing to fix is the callout, not the tempo.
Finding 2: the mistake limit is only brutal for two levels
Levels 1 and 2 allow one mistake. From level 3 onward, every level allowsthree, and that never changes for the remaining 28 levels.
This is worth stating plainly because it is widely misremembered as "one mistake and the run is over". That is true only in the first two levels — which are also the two with a single module and two and a half minutes, so the limit is not what makes them hard. It is a teaching device: on level 1 the game wants you to see the failure state exactly once, early, while the cost is thirty seconds.
From level 3 the limit stops being a lever. Which means that if your group is losing runs on level 20, it is not because you have three strikes instead of five — you are burning all three on the same misunderstanding.
Finding 3: the five phases train different cognitive skills
The phase names are not flavour text. Read them in order and you get the curriculum:
"Build the language" — one or two modules at a time, generous clock, introduce the basic vocabulary. You are learning to say "second from the left" and mean the same thing as your teammate.
"Extend the packet" — the modules get more varied, and the first Chaos effects appear. Now the message you pass has to carry more fields: not just "red", but which switch, which position, which moment.
"Preserve full state" — Switch and Maze arrive. Both punish partial memory: a Switch board needs all eight values held at once, and a Maze needs a route remembered rather than a fact looked up. This phase is about not dropping information between steps.
"Store sequence history" — Morse and Monkey-Says. These require holding an ordered sequence, which is a different memory task from holding a set of values. Order errors are invisible until the board rejects them.
"Protect the procedure" — all five Chaos types are live, and the modules are Medium and Hard. The skill here is procedural discipline: the order you do things in, and not pressing anything until the whole picture is confirmed.
That progression — vocabulary, then payload, then state, then sequence, then procedure — is a much better description of the difficulty curve than a timer graph.
The seven points where it actually tightens
Even though the average is flat, there are specific levels where seconds-per-module drops sharply. These are the ones worth preparing for:
- Level 3 — 75s per module, down from 150s. The first real jump: two modules instead of one.
- Level 6 — 70s, three modules, first Pressure Chaos.
- Level 7 — 60s, four random easy modules. The tightest early level, and the first with a full board.
- Level 12 — 80s, and Random Medium appears for the first time.
- Level 15 — 60s with four modules, mixing Easy and Medium.
- Level 19 — 80s with Maze added to a three-module board.
- Level 26 — 80s, and the clock actually drops from 5:00 back to 4:00 while Hard modules enter.
- Level 30 — 75s for four Hard modules. The tightest board in the game.
Note what happens at level 26: the timer goes down after five levels at 5:00. That is the single clearest signal that the designers were not thinking in terms of "add time", and it is also the point where many groups first hit a wall.
What this means for how you practise
Three practical conclusions, all of which follow from the data rather than from opinion:
Do not train for speed. The per-module budget is nearly constant, so a group that is accurate will finish comfortably and a group that rushes will fail. Time saved by talking faster is smaller than the time lost to one misheard colour.
Train the specific skill each phase introduces. If you are stuck at level 16–20, your problem is state-holding, not reflexes. Practise reading all eight Switch values before flipping anything. If you are stuck at 21–24, it is sequence memory — practise repeating an order back before acting on it.
Expect the wall at 26, not 30. The difficulty jump at 26 combines three things at once: Hard modules, a shortened clock, and all five Chaos types. Most groups that describe the game as "suddenly impossible" are describing that level.
The data, if you want to check it
Every number above comes from the table on the levels page, which lists the timer, mistake limit, module composition and Chaos type for all 30 Campaign levels, grouped into the five phases. If you want the per-module arithmetic rather than the phase averages, that table has the raw values and you can divide them yourself.
For the modules themselves — what each one asks you to read and in what order — themodule overview is organised by module rather than by level, and theprintable tables are what your reader will actually be holding.