Monday, January 30, 2023

Championship Lode Runner: Won!

 


It's over, and what a nasty experience this was. I've played puzzle games that were more cognitively taxing than this, but most of them don't dare you to puzzle out obscure solutions under the constant pressure of guard patrols or expect you to make long series of precisely-controlled actions, nor do they give you such an inadequate number of chances to do it all. And even though the physics and AI here are mostly deterministic, there are still so many ways things can go wrong - these levels tend to be both dexterously demanding and extremely intolerant of input errors, many have false walls which aren't revealed until you stumble into them (very likely ruining your run when you do), guards can randomly respawn in very inconvenient places, and they can even randomly drop crates in spots where it is impossible to retrieve them.

An official hint book exists, and I found myself needing it rather often, though I did try to solve each level without it. It doesn't spell out solutions to the levels, which I appreciate, but sometimes the hints are too vague, or didn't help with the parts that gave me trouble, or worse, flat out don't work.

As I mentioned in my first post on it, Championship Lode Runner does allow you to save after every level, but this won't circumvent the life limit, as loading it will cost you a life and automatically re-save it to the disk. You effectively get two tries per level, which is absurd. You can't even cheat with save states; the game will, through some kind of voodoo, detect this tampering. I wound up making a backup of the disk image after each level beaten, which it thankfully cannot detect.

Here are some highlights of the game, both good and bad.

Darkened walls are false walls (my edit)

Level 7 had me stuck for a bit. Clearly there had to be some false walls here, or else there would be no way down from the pyramids, but as in the original game they appear normal until you land on them, forcing you to discover them by trial and error, forfeiting a life whenever you trap yourself in one of the many divots. Imagine doing that with finite lives and the threat of having to restart the whole game when your save file runs out, blegh. And this is just level 7 out of 50!

Even with the false walls discovered, there's still a problem - you can't dig anywhere except the very bottom, since that requires a flat surface at least two tiles wide, so though you could drop down into the inner pyramid through the top of the outer one, there's no way out. As for the outer pyramid, you can drop onto the sides of it from above and escape through the false walls on the side, but the crates put in the divot prisons seem impossible.

Once again, it's all about guard AI manipulation. You've got to get them to explore the inner pyramid for you, collect the treasures within, and when they invariably get themselves trapped, you have to go in yourself and free them to steal their gold. Same deal with getting the trapped crates in the outer one. And I can't tell you how many times I screwed this up by freeing a guard who then just used their freedom to walk into a space that was still in my way.


I've noticed a pattern. The worst levels in the game seem to be the ones that involve false walls.

Once again you'll need trial and error to discover which of the many bricks are fake, including three inescapable pits at the top, and that's just the beginning of your troubles. The lower-right corner of the level is the only spot where it's relatively easy to trap guards, who otherwise tend to cluster and form impassable choke points in the mess of tiny ladders and single-tile brick floors at the bottom. You'll need to drop down into this pachinko-like grid many times to grab all the crates, and if you land anywhere left of more than one guard, you're dead, stuck there without a way back to the ladder on the right. Even dealing with just one guard is dicey if you're getting chased by any from the left.

The trick is to get guards to drop into the inescapable traps composed of false bricks at the top, but even that seems impossible to do reliably. Get one into the rightmost trap, and you've still got to descend to the bottom and hope that the remaining guards are clustered on the left. Get one into the middle trap and you yourself have to land more in the middle of the screen, which is deadly even with only three left. And luring one into the leftmost trap seems to be impossible except by just killing guards over and over again until one of them randomly spawns there, which can easily go wrong if your timing is slightly off even once.

Even after perma-trapping three guards, that still leaves two to chase you around the level, and you'll want to let both them follow you up the ladder so that when you drop down, you can make a mad dash to grab as many crates as possible and then scramble back to the ladder before either of them touch down somewhere in your way. There are still countless ways things can go wrong, and most of them happened before I beat this level.

Oh, and here I learned that tapping "down" when standing above a trapped guard causes you to "climb" into the guard and die. The number of times this happened is a bit embarrassing.


For a change of pace, level 19 is a fun one. It came after a completely miserable level 17, where guards tend to bunch up in the corners and coaxing them out without getting caught is nearly impossible, and a truly evil 18 with no fewer than 8 false walls, any of which will trap you if you fall in. That level even apologizes when you beat it. But 19 is challenging in a good way, with several crates that seem impossible to collect at first, but there's a solution for each of them that can be deduced logically, whether it involves guard manipulation, carefully executed digs, or often both. I wish there were more levels like this, and also that there were unlimited lives, because expecting you to solve it on whatever paltry stock you came to it with is completely unreasonable. Running out of lives needn't directly concern me thanks to savescumming, but I still find it rude.


Level 27 is another good one, a minimally designed puzzle that looks impossible until you discover the "ah-ha!" moment and doesn't waste your time with prolonged nonsense once you've figured it out.

Avoiding the guard seems impossible, but you'll quickly figure out that he'll climb up the ladder, not down, if you're both on it, sending him on a clockwise loop through the "head" of the structure and back down. As for all those crates, you just need to realize that with a well-timed drop you can ride his head and cross over to the inaccessible ladder and platforms.

 

What a perfectly hateful level this one is. Completely perfect timing and execution is absolutely required to clear out the colored zones by digging the brick(s) blocking it, entering, and moving without any delay or the slightest timing error to grab one of the crates and returning to the entrance before the bricks you blasted respawn and your exit path becomes obstructed. And yes, there are false bricks too.

So much can go wrong here; the most minor input error will probably screw up your chances of returning in time. The yellow zone is the worst as there's no way to keep the guard here completely off your tail. You've got to trap him, but if he gets in your way after freeing himself, or he gets himself killed and respawns in the way of your return trip, or you're dead and have to restart the level again minus one life.


I like this one. It comes shortly after a miserable level of corralling guards from ladder to ladder, where any timing error will split them up and make them surround you, and plenty of false bricks are thrown in to trip you up for good measure, but I like this one. One guard, two crates, and no dirty tricks - just the challenge of figuring out how you can collect the treasures from their respective crates and get out. For one of them, it's carefully (but quickly!) dig around the ladders and into the pit from a direction that lets you leave. For the other, it's carefully/quickly dig around the ladders and into the pit from a direction that lets the guard drop in, grab the crate, and leave.


The very next level is quite clever, but obnoxious in the context of limited lives, as there's no way you could possibly beat it on your first try.

Getting all of the crates seems easy at first. And it is! But when you do, the exit route reveals itself as the status bar flashes the word "HOMICIDE," and soon you'll realize that escape is impossible, as the exit ladder is just out of reach.

To escape you have to kill all of the guards before getting the last chest, so when you grab it they'll be in position for you to run over their heads and reach the exit. Unfortunately, the randomness of where they respawn could mean you have to spend a very long time killing and re-killing the guards over and over again until all of them finally go where you need them to be.


Level 46, after figuring out the false brick locations, would be a neat puzzle stage of reasonably predictable guard AI and difficult but doable digging tricks, except that whether you survive the final move or not is pure luck (or superhuman timing skills), as it involves passing over the head of a guard stuck in an animation loop at the top of a short ladder. Time it on the wrong frame - there's about a 7 in 8 chance of this - and instead of walking over his head, you die.


It all ends here with one doozy of a last level. No false walls. No guardian respawn roulette. No dirty tricks. Just puzzles within puzzles within puzzles, each and every one of them demanding impossibly precise planning and execution, and if you do any of it in the wrong order, you're almost certainly screwed.

This one took me two solid days to work out, even with the hint book, which guides you through the opening moves but then leaves you on your own.

Succeed, and the game declares you a champ. Get the highest score too, and you can personalize your copy of the game with a message, and receive a code to send away to Broderbund for a signed "Champion Lode Runner" certificate. Too late for me to get in on that, but,

GAB rating: Bad. Lode Runner is a good game with a great engine and almost limitless possibilities, and Championship Lode Runner continues to offer new ideas and innovative challenges, but there was almost no pleasure to be had with these often capriciously difficult levels, and without savescumming I almost certainly wouldn't have reached the few ones that I did enjoy.

Lode Runner Turbo wasn't going to get a great rating from me no matter what - the formula had already gotten stale well before the first game was over, and there's no new level building blocks or gameplay elements. This is a pack of levels crafted by the original game's players using the original game's level editor and does nothing that wasn't always allowed by it - in that regard, not to mention the intense frustration level, it anticipates Super Mario Bros.: The Lost Levels, except for making that game's difficulty look reasonable in comparison.

But it could have been more tolerable, with a few tweaks. First and foremost, this should have offered unlimited lives. Maybe, just for fun, it could keep track of your total death count, but there's no excuse for any lives limit, let alone one as stingy as the classic Lode Runner model that effectively grants two tries per level and boots you to the main menu when you run out. The score might as well have been lost too - it serves no purpose as it is.

Second, the level displays should have offered the player perfect information. By this I mean no surprise false walls - clearly indicate them onscreen somehow, perhaps by using a different coloring scheme. Guard respawn points should have been indicated onscreen as well, including some method of knowing or controlling where the next one will activate, because endlessly killing the same guards over and over again in the hopes of getting one to respawn where he's needed just sucked.

And lastly, even with these tweaks, some levels were just so nasty that they ought to have been toned down or even replaced. Broderbund's curating process should have been more willing to cut down on at least some of the designers' bullshit.

As I said, this wasn't going to get a great rating from me no matter what it had done differently. Tolerable does not mean good. As it was, I strongly disliked the whole experience.

Sunday, January 22, 2023

Championship Lode Runner: Guard psychology and analysis of the C code

I expended a great deal of time working on a reckoning routine that would not be trapped by such lakes and yet would retain desirable economies of motion. After much wasted effort I discovered a better solution: delete U-shaped lakes from the map. - Chris Crawford

 

Anyone who's played Lode Runner for an appreciable amount of time comes to understand that the pathfinding AI can be a bit whack. It's clearly not a greedy algorithm, but it's also one that must take some shortcuts in order to fit in the Apple II+'s 48KB of memory, and it doesn't take much for the guards to get confused and march in erratic directions that get them nowhere. For the most part you don't really need to exploit or understand this - when the guards are acting dumb, getting stuck on ladders, and walking around in circles when there's a perfectly clear and direct path to getting the runner, you just shrug and grab those crates unmolested. Otherwise, you just find ways around them and dig traps for them to stumble into when they're in your way.

This is very different in Championship Lode Runner, a set of 50 user-made, Broderbund-curated levels whose creators, much like the authors of modern Kaizo-style ROM hacks, understood the engine better than its programmer had, and with this knowledge made bastard hard levels.

So, at the suggestion of Renga in Blue author and occasional commenter Jason Dyer, I thought I'd take a post to analyze the Lode Runner AI and better understand just what's going on inside those Bungelings' subpixel-sized brains. Coder Simon Hung remade Lode Runner, porting from C code that had been officially published in a Taiwanese reference book, and his conversion was invaluable in understanding this.

Goal 1: Get the runner

The first rule, after a few trivial rules that more concern physics than AI (e.g. falling, climbing out of pits, not passing through other guards, slowing down proportionally to the number of guards), is pretty simple to understand:

If I am on the same vertical level as the runner, and there is a clear path to the runner, then move toward the runner.


Pits dug by the runner are ignored, and subsequently fallen into.


 

Natural pits, however, are regarded as obstacles.

 

Walls, curiously, are also ignored. I'll assume I can just walk through them. I can't, but if the runner is on my level, I'll try anyway.

Climb the ladder? Are you nuts? That would move me off the runner's level!

Goal 2: Approach the runner's level

Here's where things get more complicated. If I can't just beeline for the runner, either because I'm not on the same vertical level, or because something seems to be in my way, then my second and final priority is to get to an advantageous level.

To do this, I will first look down, then up, then left, and then right. Each direction is assigned a score based on how advantageous it appears, and the lowest scoring direction wins. I will pick a direction that allows me to ascend or descend, but these are the outcomes from most to least favored:

  • Get on the runner's level. In the event that there's more than one way to do this, the shortest path from me to the level (not necessarily to the runner himself!) wins.
  • Get above the runner. If there's more than one way, then the closer I wind up to the runner's vertical level, the better.
  • Get below the runner. The closer to the runner's vertical level, the better.

 

When multiple moves score the same, the first one examined wins.

Let's look at an actual level to illustrate:

 

Three guards are present here, one on the upper-left, one toward the right, on the monkey bars, and another between them, by the ladder.

 

Step 1: Look down

Only the guard on the monkey bars has the option of moving down, so this step is meaningful only to him.

Guard logic: From above this ladder, I have a clear shot all the way to the bottom.

 

Let's look at all of the places in this column where I might disembark left or right:


I will not consider disembarking left or right from a ladder into a freefall unless I am at the top or bottom of it. The runner might consider it, but I won't.

What luck, the bottom of this drop, where I can move left or right, puts me exactly on the runner's level! This column is 0 tiles away from my current position, so I give this direction an unbeatable score of 0.

Notation: "↓ =0" : "If I go down I will be level with the runner with a score of 0."

Step 2: Look up

This time, only the middle guard has the option of climbing upward, so we'll see about him:

Guard logic: I can climb this ladder to the top, and there are three places I might disembark from:


I'll already be above the runner at the first stop, so I must stop there and climb no higher. I'll be nine levels above the runner, so I give this path a score of nine.

Notation: "↑ >9" : "If I go up I will be above the runner with a score of 9."

Step 3: Look left

All of the guards can move left, so let's start with the leftmost one.

Left guard logic:

There's a ladder to my left.


 There are two possible disembarkation points on this ladder:


Both of these put me above the runner, but the bottom one is closer - six tiles above him, so I give this path a score of six.

 

Middle guard logic:

There's a ladder to my left, with two possible disembarkation points.


The closest of the two puts me 11 tiles above the runner. This is not better than going up, so I disregard this path.


Right guard logic:

Let's see how far I can climb this bar to the left.


Note that, although I will not disembark into freefall from a ladder that I've climbed, this doesn't apply to ladders that I merely maneuver onto laterally.

That's a lot of options! Two ladders, each with multiple disembarkation points, plus I could just drop down from the bars where ever I feel like it!


Climbing to the first ladder to my left and taking it down gets me the closest to the runner, but this doesn't beat the option of dropping down and landing exactly on his level. So I disregard this path.

 

Step 3: Look right

All of the guards can move right. Let's start with the leftmost one again.

Left guard logic:

To my right, I can climb a ladder or keep going and drop off.

Either way, I'll land six tiles above the runner which does not beat the option of going left. So I disregard this path.

Middle guard logic:

There's a ladder to my right, which I could climb up or down.

 

Down all the way still puts me above the runner, but closer than my previous score (3 vs. 9), so this becomes my new path.


Right guard logic:

To my right I can either climb up a ladder or down a ladder.


Neither option is as good as my initial path of just dropping down to the bottom, so I disregard this path.


The final decisions:

Left guard goes left, so he can climb down the ladder and get you. Middle guard goes right, so he can climb down the ladder and get you. Right guard drops down so he can get you.


Seems elegant, right? Well,

 

When things get weird

Let's look at some examples of ways this two-step logic can go very wrong, which Lode Runner: Championship Edition demands you understand well enough to invoke on purpose:

Example 1: Fleeing the target

 

I'm on the ladder in the lower-right corner. So why are the guards running away? Why not just go right, climb down a tile, and kill me?

Simple - the AI doesn't realize it can reach this spot! As far as pathfinding goes, it is only considering places it can disembark from ladders and/or drop from.

 

The runner's spot was never considered a reachable endpoint. The spot below him to the left is, and the path there does cross the runner, so why not go there?

Because getting above the runner is better than getting below him. Doesn't matter how much farther away you wind up. Better to be three floors above the runner than one floor below him.

And so, the guards all go to the left, where they can climb ladders to a spot three floors above the runner. But what happens when they reach the ladders?


Example 2: Ladder physics suck

Ladder trouble - it's not just for Gordon Freeman

 

So the guard decided it's going to go for the ladder and take it up to a higher vantage point, right? But then when it gets to the ladder, it just sort of gets stuck there.

The reason is simple. When the guard is at the base of the ladder, it wants to climb up to the nearest disembarkation point. But the guards recalculate their paths with every step they take, and once the guard has climbed and is no longer considered to be at the base, the base becomes a valid disembarkation point, so he climbs down, entering an infinite loop! The guards on the left ladder are attempting the same, but have created a log jam.

If you climb your ladder up or down, you will become level with a valid disembarkation point and the guards will unstuck themselves.


Example 3: A reverse false wall



Two weird things are happening here.

First, as soon as you climb the first ladder, the guards chasing you turn around and climb the other ladder. They do this because both ladders will get them to your level, but the ladder on the right is closer (to them).

Second, once the guards reach the second ladder, they just won't go over the top of the ladder, as long as you don't. What gives? Shouldn't they want to reach the top, so they can reach the bars just above you, climb over, and drop down?

They would, except for one thing. Ladders against walls are confusing to guards, because their logic for determining valid disembarkation points determines walls in relation to where their feet will be, but not where their bodies will be! Every tile on the ladder is treated as a valid disembarkation point for pathfinding purposes - as long as you don't rise above the ladder, they'll try to beeline for you to the left, thinking they can step on the tile at their feet, but they can't because there's a tile in their way. They get stuck there.

Remove the wall, and they act a bit more sensibly.

Emphasis on a bit.


Example 4: Left-to-right reading order

 

Before I climbed that ladder, the guard was chasing me to the right. After I started climbing it, he turned around and started running for the ladder to the left. We already examined why he wouldn't necessarily chase me up the ladder, but why turn around and go for the one that's farther away?

It's because neither ladder reaches me (as far as the AI is concerned), but both go to the same level above me. Left gets evaluated before right, so left wins.


This is far from a comprehensive guide on the guard AI, and does not cover the many corner cases both implicit and explicitly defined in the game logic, nor any of the peculiarities of the game physics that sometimes dictates their behavior, but I think this more or less explains how the Lode Runner guards think, and illustrates what's going on in most of their screw-up situations. For now, I'm going back to Championship Lode Runner's punishing levels.

Wednesday, January 18, 2023

Game 355: Championship Lode Runner

Hello. Have we met?

Just what I needed - another 50 levels of Lode Runner - the quintessentially classic puzzle platformer that I praised for its solid mechanics and inventive challenges but also criticized for being far too long. Essentially a retail level compilation disk, compiled from the best user-made submissions to Broderbund, Championship Lode Runner offers no new enemies, nor obstacles, traps, or treasures, and even removes the original game's level editor. What you get instead is difficulty - this one wastes no time with simple levels, presuming that only Lode Runner experts will bother.

In this first level, the guards are easily trapped in the space below, letting you grab the treasures suspended in the air from the rope above uninhibited, but then retrieving the two crates at the bottom is a problem. You can either blast one of the walls adjacent to the ladder and make a mad dash for them and return, which requires frame-perfect movement and timing and absolutely can't tolerate any guard interference, or you can manipulate them into falling down the middle and getting the crates for you, which requires precise manipulation of the AI which I still don't understand well enough to perform reliably.

Instead of a level, select you have the option to save and load games in progress, but even loading your game is punitive! Loading a file permanently removes one of your lives, and does this every time you load it. It didn't take long at all to realize I'd never beat Championship Lode Runner fairly, but the game even blocks emulator save states by booting you back to the main menu when it detects discrepancies between your disk and memory.

Well, crap. At least it can't detect WOZ backups! I think.
 

Forget my 30 minute rule - I'm saving and making a backup of the disk image every time I beat a level, to get around the penalty for loading. Especially when the levels are like this.

Cute design, but there's nowhere to dig and nowhere to run!

What becomes clear from even just the first two levels is that this level pack demands AI manipulation to get anywhere. It demands understanding things such as that standing on one part of the ladder will make the guards climb the monkey bars over to the right and drop down on you, but moving just one pixel down will make them swing to the left instead and isolate themselves from the pack, giving you some breathing room, letting you figure out a way to get all of them bunched up in a corner where they'll stay out of your way, and when you poke your head out will chase you single-file instead of surrounding you from all sides. I don't think you are expected to actually learn the AI routine in such granular detail that you can deduce the proper methods, but rather that through trial and error will uncover the specific tricks in each stage and learn where to go and when so that the guards get herded where you need them.

Now this leads me with another tricky problem - assuming I can power through and beat the game, how do I go about blogging it? Certainly not by describing every single level and its solutions in exhaustive detail - that wouldn't be fun for anyone. For the original game I gave updates roughly twice a week and showed off about 30 out of the 150 levels, giving some of them nothing more than a captioned screenshot and some others detailed accounts of their novel challenges and solutions. Even that approach felt like a chore, and stats tell me that most of my readership only bothered to read the first post and none of the subsequent updates. And progress through this game is likely to be much slower, at least in terms of how many levels I can beat per day.

As of right now, I'm on level 7, and I've been playing for roughly three hours, most of my progress made today. Some levels have definitely been more difficult than others, but five out of the six I've beaten required strange AI manipulation techniques as I've described. I'm open to suggestions on how frequently to update on my progress and/or in how much detail - updates on my progress in the original Lode Runner looked like this, for comparison. Part of me wants to just give my next and final update when I've beaten the damned thing, but that could easily take a week or two.

Thursday, January 12, 2023

Games 353-354: 3D Tank Duel & 3D Starstrike

Not every company that I introduce with an early catalog retrospective is going to be important or prolific. You have companies like Capcom, who, as I chronicled in my last post, debuted with a handful of modest titles, had a major breakout hit before the year's end, and went to be one of the world's largest and longest continually operating game developers, still producing multiple high budgeted franchises to this day. But then you have companies like Realtime Games, a minor British bedroom coder who launched with a couple of Atari game clones, and are only distinguished today by a few lesser whales produced some years later. Their most notable products may have actually been their ports of Elite and Starglider, but this blog isn't concerned with ports apart from a few exceptions, and these aren't among them.

I'm not very excited to be covering this particular chapter right now, but their output that I find interesting for its own sake is far enough removed from 1984 that it feels wrong to put their launch games off until then. So now, I cover their early games 3D Tank Duel and 3D Starstrike, two unofficial ZX Spectrum conversions of Atari's most successful 3D vector games, as discretionary whales.

 

Game 353: 3D Tank Duel

 

Apologies for the bad video quality - I used Fuse's built-in video recorder this time without testing it first and I didn't feel like replaying once I realized it introduces horrendous compression artifacts. At least the framerate is accurate.

This is Battlezone, but with a bit of color in the backgrounds, a low resolution, and a worse framerate, which, along with what feel like smaller enemy hitboxes (and a tendency for enemy tanks to randomly disappear and turn into seeker missiles) makes for a game that feels more difficult than the original Atari game. Hitting a moving target that looks like a jittery mess of jagged lines is hard enough, and not in a fun or satisfying way, and trying to discern an incoming shell composited over that mess so that it might be dodged is hopeless.

I will note that my average lifespan went up considerably once I learned to listen for the bleep sound indicating a shot was fired. It sounds more like a radar ping, and nothing like cannon fire, but once I used it as my cue to back away at an angle, as I always had in Battlezone, I could usually dodge return fire, unless I was too close to begin with, or if I backed into an obstacle.

The main differences I observed, apart from aesthetic and performance degradation, were a slower pace, slower difficulty escalation, and, weirdly, smarter AI that deliberately uses cover, though the effect mainly serves to prolong combat.

GAB rating: Below average. I can see this being somewhat fun in short bursts to a kid living in 1980's Britain without many options for getting an arcade fix, but it's pointless in an age where MAME exists.

An official, Atari-approved Spectrum port by Quicksilva came out the same year, and it preserves the arcade game's monochrome look, but I couldn't get any of the controls to work except for firing the gun.


Game 354: 3D Starstrike

 

This is actually not too shabby a conversion! No points for originality - it's a straightforward, unauthorized port with the expected downgrades in resolution, sounds, and bells & whistles, but it's pretty complete in terms of gameplay, with all of the Star Wars arcade phases presented, plus a brief reactor strike stage to replace the final bombing run.

Avoiding hits does feel more difficult than the arcade game, due to a combination of factors including dodgy hitboxes, the framerate and resolution, and lack of analog controls, but in turn you can take a lot more punishment, and each round completed gives you a decent chunk of shields back. Apart from that, the most obvious gameplay change is the addition of a laser power meter, but it recharges so quickly that it almost doesn't make a difference. I was able to finish a round on "very hard" difficulty, the equivalent of beginning on round 7.

GAB rating: Average. Once again, I can see this being a good value in its era when you couldn't just fire up MAME and emulate the real thing. The end result is more playable and more fun than Tank Duel, and I've certainly seen worse Star Wars computer games than this, but I didn't see much need to play this for more than a few minutes.

Wednesday, January 4, 2023

Games 350-352: Early Capcom

Get Capcom Arcade Stadium, which offers Vulgus, 1942, and 40+ other Capcom arcade games as DLC:
 
2nd Stadium offers a different set of games and includes SonSon for free:
 

With the holiday season over and some initial workplace chaos settling down, Data Driven Gamer enters the new year with a Capcom introductory retrospective!

Some of my favorite posts have been these retrospectives on the early games of Japanese arcade developers, but it's been awhile now - almost three years - since I've done any. Capcom, the subject of this post, is often compared with Konami, as both were arcade developers who successfully launched several console-oriented franchises in direct competition with one another. In truth, Capcom was a relative latecomer to the game, with their first title released in mid-1984, long after the industry had proven its legitimacy. 

Capcom skips that awkward phase that seemingly every Japanese studio went through, where their earliest titles are ripoffs of Breakout followed by ripoffs of Space Invaders. That bubble had long burst by Capcom's time. Instead, their first game is ripoff of Xevious.

 

Game 350: Vulgus

Latin for "common people" and speculated by some linguists to be cognate with "folk," Capcom probably wasn't all that interested in literal meaning when they came up with this title for a space blasty shoot'em up.

 
 

Not quite a carbon copy of Xevious, almost every aspect is less interesting than said predecessor, and the few things it does differently don't really advance the genre in a meaningful way. The biggest and most obvious difference here is that instead of having a secondary air-to-surface torpedo weapon, you instead get a more powerful missile with limited ammo, which is clearly intended to be used on big enemies, who otherwise absorb many shots before going down, and on specially-colored enemies who line up in a queue for a few seconds, beckoning you to gun them down in one big chain shot for bonus points.

The removal of your air-to-surface weapon effectively cuts the enemy variety in half compared to Xevious, and the replacement, which is essentially useless for regular combat does little to make things more tactically interesting. Having to weave through enemies and their fire to chase down the iconic POW icons to replenish ammo does create a risk-reward situation, but it's no more interesting than the situations in Xevious of risking your life to bomb targets of opportunity, which were more organic-feeling to begin with.

Along with the POW icons, collectable letters 'E' 'S' and 'D' will spawn as well. Collecting them causes enemies to change in appearance, and the overall difficulty does seem to increase, though I can't tell precisely what changes. Stars worth 10,000 points will spawn after collecting all three letters, which is a good chunk of the change needed to score an extra life, but I tended to make it farther if I forwent these letters and the corresponding difficulty tweaks, not to mention the risk involved in diving headlong into enemy fire to grab them.

Another unfavorable comparison are the ground visuals. Xevious had the illusion of an endlessly scrolling terrain by cleverly using a fairly compact map of forests, roads, and oceans, with the occasional landmark such as the famous Nazsca lines. Vulgus instead has three planets to fly over, but each planet is just a random set of repetitive terrain, and there's no sense of wonder on what you might see next if you make it just a little bit further than before. Each new planet you reach is colorful and visually distinct, but only for the five seconds or so that it takes you to see the entire tilemap.

The brief space interludes between planets do, at least, provide a welcome breather with easy fights compared to the intense surface combat. Except for the last one, where you encounter the titular boss.


The vulgus itself just seems to spawn mini-vulgi forever, blocking my shots and making it nearly impossible to hit it. Not much else to say here except that Andor Genesis was cooler.

GAB rating: Below average. Vulgus overall just isn't remarkable for anything except being Capcom's first game. Sometimes just seeing where a big company started is rewarding in itself, but not this time.


Game 351: SonSon

Based loosely on Journey to the West, you play Son Son the Monkey King and, in 2 player mode, Ton Ton, his pig companion, and must fight through a diverse array of creatures, monsters, and spirits to rescue their friends, while looking out for score-boosting snacks like carrots and french fries. SonSon looks like a Mario Bros-style platformer but plays more like a scrolling shooter, with platforms serving as lanes that you can quickly and freely scamper up and down from.

SonSon offers a simultaneous two-player mode, and I played it with "B."

 

And wouldn't you know it, this is a neat little game! Much more so than Vulgus, there's a clear sense of progression as new and different types of enemies spawn, with new formations and attack patterns to challenge you, and while later stages can get overwhelming with the sheer number and variety of enemies, whose attack patterns overlap in ways that make avoiding the entire combined barrage so much more difficult than any individual type. But it rarely feels totally unfair - SonSon is quick and powerful, armed with a rapidly firing staff that destroys most enemies and their projectiles with one hit, and no enemy moves so quickly or so unpredictably to make deaths feel unavoidable. Another clever touch, I must note, is that upon being killed, the mercy invulnerability customarily granted for the first few seconds of your new life takes the form of a magic flying cloud, from which you may shoot freely, but vanishes as soon as you move from it, giving you control over precisely when the effect ends and an incentive to do so.

There's a surprisingly deep scoring system here, one that dares you to lunge into danger to score those extra bonus by killing chains of enemies or snagging collectibles at the right time and/or order. Low-scoring snacks give way to high-scoring ones by collecting six in a row, POW icons transform onscreen enemies into high-scoring pickups, hidden bamboo groves can be worth thousands of points for discovering, and Capcom's trademark icon spawns and is worth thousands if you destroy the fortified checkpoints quickly enough. All too often, greed can be your downfall, and you must balance your desire for points with your will to live. With experience, though, comes stronger and more confident point-chasing.

The co-op play, which I must assume was inspired by Mario Bros., is a great feature, and while we didn't come close to finishing a loop, having double the firepower proved handy for eliminating entire waves of enemies, and added a touch of competitiveness as well, with an implied contest to see who lasts the longest and/or scores the highest, as each bonus point opportunity can only be cashed in by the player who scores it firsts. Frustratingly, though perhaps on purpose, both players can work together to destroy enemy chains, but the whole bonus goes to whoever kills the last one.

After our session, I continued to play SonSon in solo mode, eager to reach the end even if I needed save states to get there. After some hours and about four save states made using my usual rules, I was able to rack up enough lives around the midway point of the game that I could finish the last ten milestones in one segment of pretty intense gameplay, after which SonSon found a giant brass Buddha and a scroll of enlightenment or something like that.

 

GAB rating: Good. This was a surprise, but I really dug SonSon, one of the most obscure games from Capcom's early days. It controls well, plays well, and just feels well designed in so many regards, from its difficulty balance to the hidden depths of its scoring system, and is fun in both singleplayer and multiplayer. It's got personality too, with colorful, expressive sprites that anticipate the likes of Mega Man more than they recall the bland generic Xevious knockoffs from Vulgus, and cheerful, catchy oriental-themed music oscillating with the action. I suppose that with almost 20 milestones, it's a bit overlong and can become repetitive, but overall this is certainly Capcom's best game of 1984, and I recommend it without any reservations.


Game 352: 1942

"Destroy Tokyo?" I'm just one pilot, not Godzilla!

I've often felt that Capcom's games tended to be more western in style and aesthetic than the exports of their counterparts, which ranged from culturally neutral to unapologetically Japanese. Whether it was Ghosts 'n Goblins' medieval mayhem, Final Fight's gritty urban brawls, the many Disney games they produced throughout the early 90's, or Resident Evil's Alone in the Dark-cum-Night of the Living Dead rural mansion of frights, jumpscares, and viscera, their games often wore a certain stamp of familiar western identity, if that makes any sense.

So their first whale, 1942, happens to be their first game made specifically with western audiences in mind. In an otherwise bog-standard vertical shmup, the action is set in the Pacific Theater of WWII and has you singlehandedly defeat the entire Japanese fleet and win the war in your P-38 Lightning.

The game does have a set ending point, but it features a whopping 32 rounds to complete before you get there - Vulgus had only three, plus the short space interludes. Continues are permitted, and I would certainly never have bothered finishing the game otherwise, not even with save states.

Below is a recording of myself getting to the end, using continues as much as necessary. It isn't my first go, but it is my first and only serious attempt at completion - the unlimited continues and sheer length gave me little incentive to try to improve.

 

And good grief, this is overlong. The game starts to get monotonous before the second stage is done, and there are still 30 to go after that. I spent about an hour and a half shooting down the same planes over and over again, with multiple pause breaks (thankfully not recorded to video) to give my sore trigger finger (and my poor ears) a rest. The above video cuts out about 10 minutes of gameplay where no progress was made on a full set of lives - mostly from the penultimate stage - but is otherwise unedited.


To be fair, 1942 does feature some gameplay progression as enemy attack patterns become more elaborate and varied, and some scenery variety as you advance from the open oceans around Midway to the volcanic Marshall Islands, the jungles of Saipan, the desolate mountains of Iwo Jima, and the urban districts of Okinawa. But it takes too long to get there, and much to long to reach the end. I can't fault the game too harshly for having less variety than its successors, both official and spiritual, but for what's here, there should have been maybe eight levels instead of 32.

Powerups, now a standard shmup element, are seen here in a way that predecessors Vulgus and Xevious hadn't really explored, with red formations of enemies now dropping POW icons when destroyed that can confer a variety of benefits. It's modest compared to later games, but these powerups include:

  • Bonus points
  • More firepower, turning your twin cannon into a quad one.
  • Wingmen, tripling your firing spread, but also your hitbox width
  • A smart bomb that instantly destroys everything onscreen
  • Extra lives
  • Extra loops

 

Wingmen are arguably the most useful powerup, but also the most difficult to hold onto, as they themselves are destroyed by enemy fire. This can be used to your benefit, though, as they can shield you from bullets fired from the side or even kamikaze large planes.

Loops are your secondary ability, an evasive move to avoid danger, and a poor substitute for the expendable smart bombs that would come to fill this role in later games. The problem is that it's very hard to know when you're about to get hit and react fast enough to loop out of danger. A loop can also easily drop you into just as bad a position as the one you had looped out of, since it does nothing to eliminate the planes that were threatening you to begin with, and you can't fire during the maneuver. I'd typically use them to escape corner traps, though some levels also force it by pulling the occasional dick move.


Dick moves in general seem to rule the skies here, which I expect is a purposeful result of a quarter munching design philosophy that expects players to die a lot and continue until they finish the game or run out of quarters. Bullets aren't dense, but they come out fast and without warning, and in some stages they blend in with the scenery behind them. And I loved how the only way to take out large planes with any sort of efficiency is to absolutely hammer the trigger button while tailing them so closely that dodging their return fire becomes nearly impossible.

I did eventually finish all 32 stages, but the 31st throws a boss at you, a screen-filling heavy bomber that utterly annihilates your vicinity of the screen with bullets, often at far too close a range to make dodging realistic. And this is no bullet hell game where your hitbox is any smaller than the sprite - just grazing a bullet is fatal. This fight ended in many frustrating failures, most of which aren't in my video, but eventually I had a frantic, lucky run where my loops didn't land me right back into the thick of bullet sprays and where I managed to down it before running out of them.

This wasn't it.
 

Beat it and you're rewarded with an encouraging message ("YOU ARE THE BEST OF PLAYER!") and a final, comparatively easy stage with no climactic moments. And then it's over. The war is won, the atomic bomb sits in a New Mexico lab unused, and generations of nuclear paranoia are averted.

That "special bonus" comically eclipses any score you could hope for, making the whole endeavor pointless.

GAB rating: Average. There's fun to be had in 1942, but not nearly enough for so much padding. The WWII setting is a neat change of pace from the saturation of space shooters, even if it just plays like another space shooter. The visual design is appealing with its oceans, islands, and colorful but recognizable historical airplanes, though the audio design is horrible with tinny explosions and an unending loop of whistles and Morse code-like tapping noises when there ought to be bombastic action movie music or just silence. Sequels and other games in the genre would vastly improve on the formula with better variety of enemies, stages, and weapons, but for 1984 standards, this really needed to be more compact and more balanced.