Reseñas recientes de Sponge Cage

Honorable successor of the previous game, The Case of the Golden Idol (2022), which is not required to play before playing this one.

Great soundtrack that doesn't feel repetitive, although a bit shy. Great artwork and animation, they really outdid themselves in this one. Crazy to see how they can create 3D looking movements from 2D animation.

The gameplay is very similar to the first one, although a bit easier I thought. Deduction is pretty easy, the difficulty comes from wording your deductions the way the game intended. This can be frustrating at time, and I found myself cheesing it by randomly swapping words until I found the misplaced ones.

Fun game, great concept. But pretty short. Well worth its price!

game good

TL;DR: Save often to avoid game breaking bugs, you can stop playing during Act3, watch a video if you're curious about the end. Or you can skip this game altogether, you're not missing out.

I'm a big fan of the series, and there honestly wasn't a game I didn't like... until I played this one. I agree with most people who say this is probably the worst in the series. There are some very good moments, but I wouldn't recommend it.

Pros:
* Humour always on point
* Great voice acting
* Most puzzles make sense and are quite fun
* The first two acts are quite enjoyable

Cons:
* Controls: this game was clearly made to be played with controllers, don't even think about using about using your keyboard, it's absolutely not intuitive (don't even think about using your mouse, it's not even supported). Yet, the game doesn't take into account the controller deadzone, leading to infuriating moments in Monkey Kombat, and the diving competition. You need to be extra careful to take this into account, annoying.
* Game breaking bugs: completely block your progression. You need to load a save. (more info below)
* Monkey Kombat: Because of a game breaking bug, I had to do the WHOLE Monkey Kombat progression TWICE. It wouldn't be so bad if it was quick, but NO you need to 1) spend time figuring out how it works, write down all the stance combinations, figure out what moves beat what. 2) Spend time FIGHTING and BEATING other monkeys before fighting the "final" monkey to solve the "puzzle". 3) Sometimes you straight up CAN'T beat other monkeys despite doing a perfect game... So you need to spend extra time to fight MORE. 4) Sometimes because of the deadzone of the controller you accidentally play the wrong move, which means you'll most likely lose the match because of that. 5) the total time I spent on Monkey Komat is probably 1-2 HOURS! Just looking at my notes to figure out what moves to play and STILL LOSE despite doing a perfect game (this sounds really fun doesn't it??). It wouldn't be so bad if you didn't spend so much time doing the exact same boring thing...
Bringing Monkey Kombat for the final boss fight was the cherry on top. What a horrible experience to finish off this game.
* There are actually so many bugs... Some game breaking (see below), some plain annoying
* You can't alt-tab


During a single game I encountered THREE GAME BREAKING BUGS. I had to reload a previous save and redo a lot of the puzzles, this was really not fun
1. On the second island, Guybrush got stuck and wouldn't move anymore
2. On the second island, Guybrush got stuck on the raft when docking at Pegnose's house (loading a save didn't fix this one, I had to slow down the game to a halt to let the cutscene progress...)
3. On Monkey island, Herman was stuck because I had the misfortune of throwing a random object at him, and not one of the object required for the puzzle. The story couldn't progress, and I didn't even know it... I had to reload a save at the very start of Act 3 because I broke him very early unknowingly, which meant that I had to do the Monkey Kombat garbage twice... What a terrible ♥♥♥♥♥♥♥ experience
4. (There were quite a few crashes during my playthrough as well, at least 3. So that would be actually 6 times during my playthrough where I had to load an earlier save and lost progress. So much fun)

There are apparently many more game breaking bugs according to a quick Google search. But the fact that I had so many in a single playthrough is honestly shocking. Can't believe the game shipped like this...

Simple yet very effective game mechanics, perfectly utilized in all puzzles. Stunning visuals and environments. Some of which are actually jaw dropping.

Puzzles are pretty simple, yet not too simple. Getting 100% is pretty hard though, not gonna lie. Which is absolutely great, since it gives you replayability with some puzzles way more challenging than in your first playthrough (the game is intended to be played two different ways).

Sound design is very clean, although not super polished at times. When going through portals for instance, all sounds in the previous area immediately stop. Music is absolutely great though and fit the visual style perfectly, and high moments in the music match the tone and scale of high moments in the game itself (like when discovering a vista for instance).
This adds a lot to the immersion.

I'm a big fan of this kind of abstract geometry/environment and pseudo non-euclidean space. I would definitely recommend if you already find screenshots and videos on the store page visually appealing, because the game delivers.
I believe this is the main attraction of the game as the puzzles are clearly secondary and an excuse for you to explore this beautiful world.

A VR version of the game would probably be amazing and even more immersive than the game currently is.

While I enjoyed playing this game, it was very brief and easy. But I cannot recommend it.
I would give this game a 60%.

TL;DR:
- Not a programming game
- Don't expect to learn anything
- Pretty easy, but challenging if you aim for gold medals (max efficiency)
- Weird game progression: game elements introduced, then discarded; Sawtooth difficulty
- No auto saves (???) Make sure you save constantly because you can lose all your progression at any moment if you run out of money...

I was under the impression this game was going to be a programming game like"Human Resource machine" or "7 Billion Humans". But this is not technically the case.
All puzzles are basically the following: You need to sort colored shapes from a random stream into specific categories. To do this, you need to create a program using and connecting nodes together.
The twist is that the nodes you use are not the ones you would expect, and there is no efficient way to clearly identify a shape/color. It's up to you come up with the most efficient way (fewest nodes, best speed).
There is also some puzzles that don't require you to be 100% accurate and allow a certain margin of error.
This gives an interesting approach on how to solve the puzzles with a different mindset than the one you might use for programming (which I believe is actually a good thing).

I was also under the impression I would get to learn more and experiment with machine learning techniques. But once again this is not what the game is about. The game vaguely mentions machine learning techniques (both incredibly dated and fairly recent), while providing you links to learn more about it.

Most puzzles are fairly easy. Don't expect something incredibly challenging, except maybe if you aim for gold medals. In which case the puzzle becomes a balancing task to evenly distribute input streams to the output streams. This is frustrating at times when you're just one short for instance.
You can also cheese some of the puzzles (infinite loops, using nodes' "memory" as a storage instead of trashing them, intentionally "freeze" your program and still get a gold medal (1)) which does not look like this was intended at times and could introduced completely undefined behavior when reused as a single node later one.

(1) Later on in the game, instead of outputing a certain amount of shapes, you need to maintain the margin of error above a certain level for a certain amount of time to pass the level. This is extremely easy to cheese by intentionally slowing down your program.

Some elements of the game are introduced early on and are extremely efficient at sorting shapes. Unfortunately, you don't get to use these later on. What was the point of introducing them at all? They are never brought up again either, even though they might be more efficient than solutions you can come up with without it...

There is a puzzle that involves a "genetic algorithm" to solve a self-driving car problem. This is an absolute train wreck to be honest.
There are many problems with this, but the most obvious one is that you have no idea at all what's going on.
You are supposed to "train" and "evolve" your program, which is done automatically for you (except in the case of supervised learning which actually is the method that works the best in this game). But there no way to understand what's going on. This makes the learning process absolute trash. You can turn some knobs, like with the real thing, but the learning process and the outcome is seemingly random and it sometimes look like the actual learning is a complete regression.
A way to visualize all possible agents, a way to properly define fitness (this is stupid to base it on average speed when you car will break at the last moment because frontal sensors only spot cars way too late), better sensor system (no way to distinguish left/front/right sensors of the car, absolutely terrible when dodging traffic...)
For me this is the worst part of the game. Very frustrating puzzle with completely random outcomes, completely misses the point of what genetic algorithm and reinforcement learning is about.


I also didn't like the "startups" puzzles. This is pretty hard to explain exactly what is wrong with this. But basically you need to complete a puzzle as usual, except that the reward is not a medal, but in-game money.
The problem is that there is no way for you to gauge how your program is doing and what you need to do to increase the cash flow. There is the option to "patch" your program after release to potentially optimize/refine it, but this is useless since you have no way to measure what is wrong with it to begin with.
This is fine since you don't absolutely have to do them to complete the game (even to get all achievements). But I think it's just a shame, since it could have been a lot more interesting (e.g. Changing the output after release, actual telemetry of why users are unhappy).
In the end, I finished probably 2 startups puzzles since there was no real point doing it.
Side note, I also had a bug where I generated infinite money with a startup even though the user rating was awful.

Speaking of bugs, the game has no auto saves. There is a "feature" in the game where you lose all your progress if you run out of money. This happened to me right in the middle of a playthrough and had to start over from the start.
Although there are checkpoints, your previous puzzles won't be saved (which can be used as a single node to simplify your code). Meaning you basically have to start over from if this happens to you...

Very fun and challenging game. I love the concept, it's so simple and yet very effective

Once you know how to play, you could basically spend hours playing without even getting bored.
But if you think that the game is still missing a little something, the workshop is FULL of mods that could spice your game up.