Game 333: Impossible Mission

Another visitor! Stay awhile. Stay FOREVER.

Thus barks Dr. Atombender in menacing SID-powered clarity as you descend the first elevator shaft into his secret underground computer lab. Sworn to revenge upon the world for ruining his video game high score with a poorly timed power outage, you must locate and enter his control room and stop him from breaking the ICBM launch codes, ensuring the security of U.S. nuclear secrets for almost another 40 years.

You have six ingame hours, and each agent that bites it will take ten minutes to replace. 30+ lives may seem like a lot - it carried me and many others through the famously difficult Contra. But this game ranks somewhere between Jumpman and Jet Set Willy in terms of ridiculous lethality. It's not uncommon to get killed ten times in the same room and still not clear it.

Smooth C64 sprite animations are recycled from Summer Games.

Atombender's lair is a semi-randomly generated maze of elevator shafts and 32 connecting rooms. These rooms' layouts are fixed, but their positions in the lair and the behavior of the patrolling Dalek-like robots are random. Most of your time is spent here, where you must search the furniture for the passcodes you'll need to progress. They look a little plain, and a few rooms are downright ugly, but Epyx makes good use of the C64's SID chip for ambiance here, with the constant hum of mad science equipment, whirring motors, buzzing electrical discharge, and occasional orders of destroy him, my robots heard spoken over the rooms' loudspeaker systems.

The main goal is to locate the 36 punchcard fragments, but you'll also find, on occasion, one-time use tokens, which can be used at the various terminals to reset the elevators, which I rarely found useful, or to freeze the robots in place, which are incredibly useful and almost too good to use, except that sometimes a robot or two is placed and programmed as such that there's no other way to clear the room. There's definitely a puzzle element to solving these rooms along with the arcade-like platforming, and the procedurally generated elements change up the solutions from game to game to some extent.

This formula would be fine, if a bit of a letdown after so much creative level design seen in Jumpman the previous year, but Agent McFlipperson here has an awkward jumping arc that can't be controlled and is hard to predict. Pixel-perfect collision detection, normally a good thing for an arcade-style game, becomes frustrating when applied to fancy animation loops, and though his death scream is legitimately hilarious, it gets old. I frequently missed platforms simply because my feet weren't facing the ground at the precise moment of passing over them. I'd sometimes get killed by a robot behind me because my sprite's feet extended too far backward in a running animation frame and touched the robot. Tiny gaps can in theory be crossed by your stride without having to jump, but success depends on your sprite's feet spanning the gap at the precise moment of crossing. Sometimes you cross, sometimes you just fall in, and it feels completely arbitrary. There's a reason why sprites in platformers usually have compact, rectangular silhouettes, or failing that, rectangular bounding boxes or tile-based collision detection over pixel-based ones.

Another unfair difficulty factor is that you can never be completely sure of the robots' behavior until you're in their line of sight, which can be instantly lethal, and you won't know until you try. A lot of them simply patrol back and forth, firing their weapon at fixed intervals (or not), and pay no mind to you, while others may detect the player based on line of sight, or proximity, or movement, or some combination, and react by turning to facing you, accelerating, and/or shooting.

Your only defense, apart from avoidance and jumping, is the use of robot snooze tokens, which are much more limited in availability than your lives. No sonic screwdrivers here. There's the constant dilemma of whether to use a token to help pass a difficult room or to save it for an even more difficult one later. Some pieces of furniture will turn out to be impossible to search without using a token, and nothing's more disheartening than dying from a bad jump after expending one, thereby wasting both it and an agent.

A gif can't properly convey how the music puzzle works, but you must sort notes from low to high.

 

Two special rooms in the lair with chess board patterned screens theoretically grant unlimited tokens by having you solve musical puzzles, but they get unsustainably more complex with each one solved, and become basically impossible to solve without a scratch pad or a good pitch ear about eight iterations in. The token type reward you get alternates, so you must solve two puzzles for each precious snooze token you wish to obtain.

 

The game is solvable, and I managed to catch Dr. Atombender after a few days of practice, memorization, and some luck. Also with quite a bit of patience for farming tokens at the music rooms, which I solved using a variant of the merge sort and some careful note taking.

I eventually learned which rooms are the most doable, which rooms are trouble, and figured out general solutions to navigate the trouble rooms, including precise footings to jump from. I learned that it's better to squander a snooze token than risk death, but better still to clear a room deathless without using a token if you can.

My first successful run was far from perfect. I did clear most of the rooms without needing snooze tokens, thanks to my practice and a lucky avoidance of randomly bad robot configurations, but here and there I wasted lives, and a few rooms were especially bad for wasted tokens and lives. I spent a good 20 minutes farming tokens in a room near the start, and yet I still ran out as I neared the far side of Atombender's lair, forcing a time consuming refill. But as inefficiently as I played, I still finished with almost an hour to spare, and could have certainly done better than that if I had been smarter about assembling the punchcards.

Even if you know the password, you still have to assemble all nine punchcards.

 

The final challenge is to take the 36 punchcard fragments and assemble them into nine complete cards, each corresponding to a letter of Atombender's control room password. Cards can be flipped, rotated, and copied, and there's also a color-changing mechanic which only serves to waste time. You can also phone H.Q. for hints at the cost of two minutes per call. One hint option is to have them rotate two of your fragments into the correct orientation, which I decided to just do for all of my cards and then brute force them together until I found combinations that fit. The hint calls take 36 minutes to get all of the fragments into position, and I had plenty of minutes to spare.

The other option is to have them tell you whether or not a specific fragment can be completed with the rest of what you've got in your inventory, and this strikes me as nearly useless. If you've got all 36 fragments, then you can assemble cards from all of your fragments and the question is redundant. If you've got less than that, well, I estimate that you'd need 27 fragments in your inventory for there to be about a 40% chance that its three partners are located in the remaining 26. It seems far less tedious, and a more efficient use of your phone hint minutes, to wait until you have them all and use the auto-rotate hint for all of them.

Once all nine cards are attained, the control room can be entered. Hopefully you remember which room it was in!

No. No. No!

GAB rating: Average. Impossible Mission has a good overall concept, ahead of its time in ways, and some great audio aesthetics, but I didn't find it very enjoyable to play thanks to an obnoxiously unbalanced difficulty curve brought on by weird jumping mechanics and other questionable design choices.