Spelunky was originally released in 2008, so early in the indie boom that it was released as freeware since being able to sell a game on the internet had yet to really explode. Spelunky HD was a remake that wouldn't be released until 2012 in an effort to actually make a profit off it, and included quite a few refinements, additional secrets/areas, etc. It's probably one of the most influential games of all time since it was the first platformer roguelike (that is, randomly generated levels in a platformer)- a genre you can't walk in a digital store without tripping over these days. Even more than that, it sparked the imagination of what other genres you could apply random generation to. Diablo technically predates applying random generation to an action game by many years, but it also lost the feel of a roguelike in favor of loot gambling.
What makes Spelunky particularly incredible is that I would honestly say that it is still, 12 years and countless similar games later, hands down the best of the genre in terms of actually translating a roguelike into an action game. See, the thing about randomly generating a level is that it actually kind of sucks. If you just drop in a few rooms at random and plop down a few enemies from a pool, it will only be as entertaining as fighting those enemies is. If you hand design the room layouts, and make terrain relevant to the gameplay, then you have only created a few permutations of fighting those enemies. Add in some power ups / money, and you have only made it as interesting as how big an impact those powerups make. You might go "ah, but what if we add lock/key progression where the generator is smart enough to only put the red key before its locked door and etc!". And you will quickly find that the only reason that lock/key gameplay is interesting is because the level designer put things in interesting positions. The reality is that most games don't have mechanics that can be interesting through simple random permutations. At best you get a modern version of an arcade game where randomization makes harsh death penalties more palatable (which is frankly what the majority of action roguelikes, or roguelites, or whatever you want to call them actually are- souped up arcade games).
The thing that makes roguelike random generation work above all else is that they rely on chain reactions. Imagine a fictional game where a monster can explode and destroy terrain around it. Shopkeepers will attack you for destroying their store. As a result, this monster and shopkeepers both become more interesting when together. Now a trap that pushes things, pushing the monster into a shop, where it explodes. That's three things interacting together. Suddenly, random permutations start to have enough different combinations to become interesting. This is the foundation of roguelikes (though plenty of non-random games use it as well).
Spelunky's choice of genre being platformer allows for an incredibly natural action adaptation of these systems, since it comes with the most intuitive system there is: physics. Darts can push you and enemies into other things. Falling onto spikes can kill you- and monsters. A poorly thrown rock into an explosion can come flying back to your face. Specks of blood can cause explosives to go off. It's an incredibly smart choice, and it also highlights why attempts at other genres of roguelike are so difficult. It also just makes dying funny. Which is important in its own way. Because you'll be doing it a lot.
These types of chain reactions weren't unheard of in action games, with things like Doom allowing demons to attack each other and traps to damage them. But the necessity of players being able to read a situation in real time somewhat limits the allowed chain depth compared to a turn based game like a traditional roguelike. Spelunky works around this in a really clever way: almost every monster and trap in the game is stationary or in a set pattern until the player gets near it. It shakes it up in various ways (spiders fall from the ceiling, snakes pace, humans use their weapons only when facing you, etc), but the rule is only broken with great care. This prevents the chain system from spiraling out of control, or out of sight. In a way it even makes it a turn based game in a sense, as you usually have time to process a situation before reacting to it.
Even outside being one of the few to successfully translate chain reactions, I find it surprising that the progenitor of the genre also does a lot of things better than its own spawn. In particular, the thing I absolutely detest about most action roguelikes is that they tend to fall into a pattern where the early areas become trivial (due to the need for a skill difficulty curve), but it's still in your best interest to loot everything in them. This makes the start of every game a choice between tedium or a disadvantage. Spelunky sidesteps this in a slew of ways: levels are relatively small, runs are relatively short (30-40 minutes), early areas have intense resource costs that often require skipping parts of them, a timer spawns an invincible killer ghost if you take too long, and even when you're experienced at the game it requires a lot of focus such that it's easy to trip up even on the early areas. It's certainly not perfect about it, with some high score tactics requiring wasting a lot of time. But relative to its progeny, it does a remarkably good job of keeping the early run tedium down.
Would I recommend Spelunky HD? Uh, absolutely. Pretty much everyone should play it. The years have only made me appreciate it more after watching other games fail to emulate it. I could nitpick a bit in terms of the game perhaps feeling a bit like a way too fast slippery janky freeware platformer at times. But that speed is also pretty important for reducing the repetition and allowing risky play. I could nitpick that occasionally the randomization will screw you over in impossible ways (admittedly pretty rare, though). But that's it, just nitpicks.
Bonus: Stages of Beating It
Messing Around
What a cool game. I killed a bunch of spiders with a lady! Oh no, she fell on spikes and died.
Warping Around
The game lets you unlock shortcuts to each biome. Being some roguelike purist weirdo, I refused to use them until now because I considered them cheating. Since the goal was now to beat it, I swallowed my pride and tried them. I'm glad I did? This is still an action game, which means it requires some practice and learning of patterns. The warps make that possible, which makes my gut think that the action and roguelike marriage is maybe inherently wrong in some sense (it feels like a really hacky solution to a problem). Maybe more interestingly, the warps show off how every biome in the game is designed to be winnable no matter how poor your equipment luck is (the final area via warp is practically a different, harder game than a full run). The fact that the game doesn't feel broken even with good equipment is remarkable, since the equipment also manages to feel incredibly useful.
Beating It
Doing a full run was mostly a matter of learning one important lesson: the longer you spend in the world, the more likely you are to screw up. So it was mostly a matter of resisting greed and moving on. Not the easiest lesson to learn.
Actually Beating It
Naturally, the process required to unlock the true final area requires throwing away that lesson by forcing you to thoroughly explore certain areas to unlock it. I have some mixed feelings about this ending since it limits available strategies pretty severely, but at the same time I have to really appreciate that I was constantly honing my strategies for it throughout? When I reached the final boss for the first time, raw addiction set in and I was losing track of time- the temptation for just one more run was the strongest it has ever been.
I'd outline the progression of my strategies like this (don't read if you haven't beaten it, honestly):
- Is it even possible to make enough money for this?
- It is! I learned and honed several different methods for earning money.
- But ultimately it's really boring and slow to grind money (which means fewer total run chances), so I switched to robbing every shopkeeper I saw to get practice.
- Eventually I realized it's even more consistent to buy things and only steal when reaching the black market (excepting for important stuff).
- I made it to the final boss, and realized it's one of the few areas that isn't littered with one-shot kills. Which means doing sacrifices to get more HP would be incredibly useful for learning the boss. Suddenly, altars and the haunted castle got me INCREDIBLY excited to find while I previously didn't care.
- Victory
