La-Mulana 2

First Impressions

The Steam page reads Very Positive: 83% of 817 reviews (2026-09-17 snapshot). Counted without the purchase-path filter it is 927 of 1,084, which rounds to 86%. On the numbers alone this is a calm, solid recommendation.

Sort the negative side by helpfulness, though, and the top review — 191 votes — is not about puzzles at all. It reports a progression-breaking bug after a patch, left alone for two months. The next two are a corrupted save file and a mistranslated hint.

The top positive reviews are just as uniform. Take notes. Draw your own map. Do that and the game pays you back. What splits this review pool is not how hard the game is. It is that a player who is stuck cannot tell whether the obstacle is the design or a defect.

La-Mulana 2 in play — screenshot 1 as listed on the Steam store page (Steam store screenshot)

Putting the Mechanics into Words

The verb list is short. Walk, jump, swing the whip, read a tablet, use an item. Almost nothing has been added since La-Mulana. Text is carved on a wall; you read it; you go hit something several rooms away.

The word the positive reviews keep reaching for is "note-taking". Ten out of ten tell you to keep paper or a text editor open. Which means recording is itself one of this game's verbs. The clues are placed in the world, but arranging them into something you can look up stays your job throughout.

I call this building your own index. Plenty of games run on environmental observation, but most of them narrow down what is worth observing. This one does not. A single line on a tablet can point at a contraption dozens of rooms away.

The negative side describes the same property in different words: nothing marks where the puzzle statement ends and the scenery begins. That is an unspecified observation resolution. Both camps are looking at one design decision; they differ only on whether it is a virtue or an omission.

La-Mulana 2 in play — screenshot 3 as listed on the Steam store page (Steam store screenshot)

The Texture of Difficulty

Read the pool and being stuck here sorts into three kinds. One: the clue exists but you have no index to connect it. Two: instant-death traps and bosses that you grind down by repetition. Three: bugs and mistranslations.

Only the first is what the positive side calls bite. The line about climbing a mountain during an avalanche with a lung disease comes from a positive review, not a negative one. The sheer amount of difficulty is not actually in dispute here.

The dispute is over the second and third kinds. The top negative reviews say the instant-death traps are under-telegraphed and that there are more pits and spikes than in the first game. Then they file progression bugs and mistranslated hints under the same word: unfair.

That is the interesting part. A puzzle you cannot solve and a puzzle that is broken wear the same face from the inside. In a closed-board game like Stephen's Sausage Roll you can trust that being stuck is always your own thinking. In these ruins you cannot.

La-Mulana 2 in play — screenshot 5 as listed on the Steam store page (Steam store screenshot)

Teaching

Both camps agree the sequel is easier to get through than the original. One review calls it "La-Mulana 1 for normal people". Autosave at checkpoints, tablets recorded for you, a gentler opening. The procedural overhead really has been subtracted.

The negative side answers that the kindness lasts about two hours. It welcomes newcomers and then changes its slope. Flattening a learning curve and making the cause of a blockage visible are two different jobs. The first one is done here. The second is not.

The translation problem is invisible to anyone playing in Japanese. Several reviews describe an English hint that lost its meaning and sent them to Japanese-language forums for the answer. Localisation is part of the learning curve. Change the wording of a clue and you have changed the puzzle.

NIGORO is a Japanese studio and PLAYISM is the publisher; Japanese is the source text. So the kind of difficulty you meet may depend on which language you play in. A single 83% flattens that difference out of sight.

La-Mulana 2 in play — screenshot 7 as listed on the Steam store page (Steam store screenshot)

Place in the Lineage

Of La-Mulana I wrote that the clues are there and only the index is missing. The sequel helps you build that index a little. I think the direction is right.

The negative side counters that the shape of the ruins changed — that the areas are now a branching tree rather than a mesh that folds back on itself. Change the topology of a labyrinth and you change how an index gets built. Taste can reasonably split here.

The build-your-own-index lineage kept going. TUNIC turned the index into an object, a manual. ANIMAL WELL outsourced index-building to a community. Void Stranger folded the index into its story. This game picks none of those and stays with pen and paper.

La-Mulana 2 in play — screenshot 9 as listed on the Steam store page (Steam store screenshot)

The Reviews I Read

This piece was written from the Steam store page and the Steam review API as of 2026-09-17.

・Steam: La-Mulana 2 (the storefront shows Very Positive, 83% of 817; counted without the purchase-path filter it is 927 of 1,084 = 86%)

・I read 37 reviews in total: the top 15 positive and top 12 negative by helpfulness, plus the 10 most recent.

・Destructoid's review (no score; the writer states they did not finish, and points at the backtracking and the bugs)

・On the recent side, the progression-breaking bugs that dominated the early negative reviews have largely disappeared; what remains is a controller input bug and complaints about the labyrinth being too linear. The argument has moved with time.

・The images here use URLs exactly as returned by Steam's official screenshots API. This run could not open the image files themselves, so the captions and alt text record only where each image comes from, and make no claim about what is in the frame.

Closing

I give it 7.8, a shade below Steam's 83%. Not because of how hard it is, but because it hands you no way to tell one kind of blockage from another.

Completion times mentioned in the reviews run around 30 to 40 hours with a guide; some claim well over 100 hours going in blind. I have set the difficulty one step below the original's, because the pool agrees almost unanimously that this one is the easier of the two to get through.

Who it suits is clear. People who do not mind taking notes, who enjoy drawing their own map, and who can separate "I missed something" from "the game is broken" while stuck. That third requirement does not come up often.

Who it does not suit is equally clear: anyone who wants the game to guarantee that a clue is findable. That is not laziness, it is a fair demand to make of a design. This game decided not to meet it, and 83% is best read as the ratio of people who accepted that decision.

La-Mulana 2 in play — screenshot 11 as listed on the Steam store page (Steam store screenshot)