Tuesday 19 June 2018

B.A.T. II - Won!

Written by Tornado

Hello, I’m Tornado, and I'm a relatively new reader of this wonderful blog. Through it, I've been able to relive games I played years ago and learn about those I missed. And so, I’m grateful for this opportunity to write this final post on B.A.T. II (aka, The Koshan Conspiracy).

At the outset of this guest post, let me express my deep thanks to Ilmari for pushing through 99.99% of the game. I played Koshan Conspiracy as a kid and never got very far. Over the years I’ve wondered what I’d missed, and so I’m grateful that he took over 33 hours of gameplay time to flesh it out here at The Adventure Gamer.

My main motive behind this post is to set the record straight regarding the game’s final puzzle, which Ilmari couldn’t quite solve.

This is where we left poor Ilmari.
Ilmari’s not alone, and in fact he made it farther than most (including me, back in the day). The Internet contains few descriptions of B.A.T. II, and these generally recognize it as an atmospheric game, but also a strange and generally impenetrable one. Other than Ilmari’s effort, I’ve only found two other attempts that made it as far as the endgame.

One of these simply says, after being Google-translated from French, “combine the lights on the left to see the end of the game.” That’s not very descriptive or helpful.

Another also beat the game, but “never figured out the logic” of the final puzzle and thus concluded, “possibly, the whole thing is just random.” In a later narrative analysis seemingly written for a college class, the same author wrote of the final puzzle, “it is very disappointing because there seems to be no logic to it.” (As I’ll show, that’s not right—but hey, speaking as a college professor myself, I approve of his choice of topic for his final paper! Hope he got a good grade.)

After playing with the puzzle, I’ve discovered that it is governed by just two straightforward rules. The game doesn’t make those rules clear to the player, but once they are understood, the solution is simple. But before getting to those rules, let’s back up a bit to the entry to Koshan Tower and describe what’s going on. This is necessary to understand how the two rules are connected to both gameplay and story logic.

You can break in to the tower during the day, but it seems more spy-like to do it at night.

At this point in the plot, the player has become king of Roma II, but the Koshan company still holds power on the planet. B.A.T. isn’t satisfied with this. Their (convoluted) plan is for the player to fake his own death in a way that implicates Koshan. Specifically, the player needs to break into the vault in the Koshan building and leave a fake murder contract in it.

As Ilmari described, the player breaks into the vault using the same method as the bank robbery earlier.

What isn’t clear is that entering the Koshan Tower also starts an invisible timer. If the timer runs out, clicking on a button in the final puzzle will always lead to instant death, even if it was the right button to press. More on that shortly.

After entering the tower, the player has two choices (other than going back outside): 1) approach the vault by going to the left, or 2) go through the door on the right. As Ilmari described, directly approaching the vault leads to the player being caught and immediately losing the game.

It might’ve been nice to have a sound effect of an alarm at this point.

To avoid that fate, the player must first go to the right and flip the switch so that it is in the down position.

Hey, it’s the big, important lever that controls the alarm system—
let’s put it in an obvious and unguarded location!

Then, the player can approach the vault, leading to this screen.

It looks like there’s a lot going on here. I’ll break it down, but a warning: It’s going to sound more complex than it actually is. I wonder if overwhelming the player with a false sense of complexity is part of the puzzle’s point.

The grid flows from left to right. The = and <> are logic gates. A = gate tests whether the two colors coming in from the left are the same. If they are, the gate outputs an orange to the right, and if they aren’t, it outputs a blue. The <> does the opposite, outputting orange if the input colors are different and blue if they are the same.

The circuit board only has one exception to this and I suspect it’s a glitch. The exception is the = gate at the bottom left of the screen. No matter what input comes in from the left, it always appears to output a blue. For example, when the inputs to this gate are the same color, it shows a blue when it should be an orange. But nevertheless, the light still acts like it is orange when input into the next logic gate. I suppose this could be a purposeful red herring, but given the other game problems described by Ilmari, I think a bug is the more likely explanation.

And ultimately, that little programming glitch doesn’t matter. Despite how complex this all looks, the response to player input is actually quite simple. The player manipulates the lights on the far right by pushing one of the sixteen buttons on the far left. As long as the timer hasn’t run out, these will have a cascading effect on the lights, according to the pattern of the logic gates. The four vertically-arranged lights at the far right are the board’s output.

If the board looks like this, and I push the button at the very top…

… then the button I pushed changes from blue to orange.
The light on the right, at the top, changes from blue to orange, too.

The goal of the puzzle is to get these output lights (the four on the far right) so that they are all orange. Once they are, the player has won The Koshan Conspiracy.

Again, the manipulation of those output lights is actually simple. Really, you can forget all the messiness with the logic gates in the center of the screen. As numbered from the top, pushing buttons #1 through #4 will always reverse the top output light; buttons #5 through #8 will reverse the second output light; #9 through #12 will reverse the third, and #13 through #16 will reverse the fourth.

There is one final wrinkle, though, and that is the changing nature of the circuit board. Every 2 seconds, the four lights at the top change, which also changes the color of the output buttons on the right. The nature of the change seems to be random, but the top lights will never change to make the output all orange (after all, that’s the players task).

If it looks like this...

… two seconds later, it might look like this.

I’m describing this in detail because I don’t think this puzzle has been documented on the Internet before. But in doing so, I fear that I might mislead the reader into thinking the puzzle is more complex than it is. And really, that’s perhaps the most daunting thing about the puzzle—it tempts the player to overthink it.

In the end, the puzzle obeys two rules:
  1. Turn all of the output lights orange.
  2. Do it within the time limit.
What is the time limit? By my watch, it’s about 3 minutes 15 seconds, as timed from when the player clicks the button to enter the tower. What’s more, I tried turning up DosBox’s cycles from 3,000 to 15,000, and the timer was the same length, so it must be based on the computer’s system clock. The every-2-seconds speed of the changing lights doesn’t seem to be influenced by DosBox cycles, either.

To be crystal clear: the timer starts when you enter the tower. So, if you enter the tower and wait in one of the other rooms for a bit—say, while you take a bathroom break, answer some text messages, or take a screenshot and a few notes for a blog post you’re writing about the game— and then go to the puzzle, you’ll get a game over when you try to push a button (see the “you’re discovered!” screenshot above).

But what if you’ve saved the game at the puzzle, and the save is beyond the time limit? That’s what you’ll find if you download the save game file posted by Ilmari. Fortunately, you’re not in a walking dead state. All you have to do is go back outside the tower. Reentering the tower will restart the timer.

It’s worth noting that the puzzle’s logic also matches the storyline logic. You’re a spy rewiring a lock that has a shifting code combination, and you only have a limited time to do it before you’re caught.

The easiest way to beat the puzzle is to wait for the output to display only one blue light, and then click one of the buttons that will turn that output light to orange.

In this screenshot, only the second output light from the top is blue. I’m about to click
one of the buttons that will turn that second output light from blue to orange. 

Once you do that, you’ve finished the game.

I like the “yeeaaahhh!” sound effect here.

In exchange for the hours invested in the game, the player gets an ending sequence that’s about a minute long.

You fake your own death.

There’s a fake funeral.

Then, some scrolling text provides a “press release on roma news” about how Koshan takes the fall for the murder.

The English in here isn’t awful, but certainly isn’t great.

With the mission finished, the player’s agent flies off into space.

A sequel we’ll probably never see. Hey, Ilmari, can we put
you down for a Kickstarter donation? What? Why not?

So, that’s the end of B.A.T. II. Clearly, it’s a mediocre game at best, as Ilmari’s PISSED rating suggests. I agree that the game’s strongest point is its Roman/sci-fi/cyberpunk setting with beautiful graphics to match. I might have given one more point in the sound/graphics category, owing to the fun European techno soundtrack (for decades, the music in the in-game fast food franchise has been stuck in my head).

If the game isn’t that good, why has the memory of it stuck with me all these years, other than because I couldn’t figure out how to progress? Why have I occasionally googled for more information about it? Why was I so eager when the game appeared on The Adventure Gamer’s upcoming list, and why was I so curious to read Ilmari’s write-up?

I think it’s because I appreciate what the game was trying to do (more than what it did do). On one level, the game is a mash-up of adventure game, CRPG, various simulators, and even basic arcade games. Taken individually, each piece is unremarkable. It isn’t hard to find a better CRPG, or adventure game, or simulator. But together, they form a world, and I think that was the designers’ goal.

As much as I love traditional Sierra and LucasArts adventure games, they often shoot for worldbuilding in story and setting rather than worldbuilding that bleeds into gameplay. Even a game as great as Fate of Atlantis can feel like a series of puzzles to solve rather than a living, breathing world. In contrast, Roma II is an environment the player gets to inhabit with its own set of rules. To be successful, the player must learn how to maneuver that world, juggling basic needs like hunger and thirst, social needs such as being liked by others, financial needs by making (or stealing) money, and so on.

I speculate that hybrid projects might lead designers toward this kind of worldbuilding-by-gameplay in a way that traditional adventure games don’t. When faced with disparate gameplay elements, designers must think about how to fit them together in a coherent way. That leads them to create an ecosystem that (hopefully) balances these pieces and melds them together. A game that fits within one genre doesn’t need to think about that.

The ultimate success of such an ambitious worldbuilding project depends, in part, upon which elements are taken from each genre. B.A.T. II seems to take subpar elements of each, combining obtuse adventure game puzzles with stats that don’t matter (a common RPG problem, so I gather from reading CRPG Addict), mixed in with clunky simulator and arcade interfaces.

In contrast, as Ilmari pointed out, the real exemplar of great hybrid-game worldbuilding is the Quest for Glory series, which is more discriminating in how it blends genres into a coherent ecosystem. The Starflight games and Star Control II also stand out to me as adventure game hybrids that do an excellent job of modeling the gameplay world. Perhaps commenters can think of others.

Enough reflection on hybrids and worldbuilding; really that could be an essay in itself, and that is not my main point. Rather, I hope this post demystifies the final puzzle for any who might later try to solve it. It also gives me a sense of closure for this game, and I hope it does for other readers too. Ilmari, thank you for the opportunity to write this, and I hope you enjoy your well-deserved break!


  1. Typo in first sentence (probably my bad; it was a late edit): "Hello, I’m Tornado, and in my offline life, and I'm a relatively new reader of this wonderful blog." ... should be "Hello, I’m Tornado, and I'm a relatively new reader of this wonderful blog."

    1. Also, "If the game isn’t that good, why has the memory of it stuck with me all these years, other than because I couldn’t figure out to progress?" is missing a "how." (I promise you, I did read through this several times before submitting!)

    2. We all make these mistakes now and then. They are fixed now!

  2. Welcome and congratulations! This is a great write-up and I am glad to see we were able to close this. (I also needed help making a "Won!" post for "Star Trek", so I'm no longer the only one!)

    1. Thanks, appreciate the encouragement! And definitely no shame in needing help with that final battle in Star Trek 25th Anniversary -- that's a toughie for sure, and I guess one similarity between that game and B.A.T. II is clunky, unnecessary flight simulators. I'm not sure adventure games and flight simulators really mesh as genres. Fortunately Star Trek: Judgment Rites, when y'all get to it, seriously tones down the simulator components.

      Glad to help put this game to bed--it's helped me scratch an "itch" regarding this game that I've had for a long time.

  3. I'm glad that you were able to get to the end, and the post is excellent.

    The unseen timer is one element of the puzzle that really bugs me, since without an explicit indication of its existence one is bound to spend too much time for solving the puzzle. Also, I don't really understand the story logic for timer resetting once you get out.

    The ending seems indeed rushed. I wonder what the four gnomish persons with the shades, carrying the coffin, are meant to be. They don't look quite human, but they definitely are not Shedish either.

    1. Thanks for the encouragement, and for yielding the floor to me to write the post, and getting it in such great-looking shape.

      Correct me if I'm wrong, but is this the only timer-based puzzle in the whole game? If so, that's a bit of a surprise to throw at the player right at the end.

      I like the timed nature of the puzzle on a story level because it makes sense that the agent would have to accomplish the task in a short amount of time. But I also think the episode would've been more psychologically compelling if the game had done something to communicate the timed nature of the puzzle to the player. Explicitly displaying a countdown timer would be one way to do that. Some kind of "warning, warning, intruder alert" message would've been another way. Or, perhaps if the player fails to solve the final puzzle in time, the death message could've said, "You're out of time and you've been caught!" rather than the ambiguous "you're discovered!" message. As it is, the puzzle is too vague to induce the kind of psychological pressure that good timing puzzles create.

      In this, I'm reminded of KQ3, which, more than most other Sierra games outside the QFG series, does a decent job of building a world that extends to rules that govern gameplay. Yes, when I played that game, it was sometimes annoying to wait for a timed event to occur. But it was also a lot of fun, for me at least, to deal with the pressure of needing to get out and back before the wizard caught me sneaking around. In Koshan, it would've been a better puzzle if they had tried to create a feeling like that on a smaller scale.

      I agree that the puzzle gets less logical when viewed in its broader context. As far as gameplay, it sure is nice that it's possible to reset the timer. But it doesn't really make sense story-wise. And in the screen when you first enter the tower (which you have a screenshot of in your "Lost!" post), there's a camera in the hallway, rather conspicuously swiveling back and forth. That doesn't make sense either--it suggests Koshan would've had video evidence showing an agent breaking into the safe, which seems like it would cast doubt on the claim that they hired the assassin.

      I wondered about the pallbearers, too. I'm guessing human because they don't seem quite Shedish. But those ears sure are big.

    2. I'm quite certain that the end puzzle is the only place with an invisible timer going on. Time does place some role in the game (for instance, you can make appointments with some characters only at certain times), but that's a very different matter.

  4. Interesting that the actual solution turned out to be relatively simple. I think the worst part about the timer is that even when you lose because of it, the game STILL doesn't acknowledge the timer's existence, making it seem like the actual failure was something else.

    1. Totally agree (see the reply I just posted to Ilmari above). The timer should've been telegraphed to the player ahead of time, or at least suggested to the player retroactively upon failing the puzzle.

  5. Ilmari: Based on Tornado's findings, do we need to reconsider any aspect of the score?

    1. I don't think so. In scoring Puzzles and Solvability -category I emphasised more the lack of any adventure gamish puzzles in the majority of the game. This hasn't really changed, although the final puzzle has proven to be not unsolvable, but more like "opaque, yet solvable". Also, the end sequence doesn't really bring anything new to the story, so Story and Setting -score doesn't change.

  6. An invisible timer! Well done Tornado at cracking this final challenge and documenting it for posterity.

    1. Thanks! I'm glad this is here for the few who might ever make it this far in this wildly obscure game.

  7. It is good to see the ending, I checked the Youtube videos for interest's sake as well. I must say that while it makes sense story wise the whole lack of indication of a hidden timer is really short-sighted. But from my experience of games in those times there were a few trying to mix genres together, it's one of those fads that everyone thought would be the path for games of the future, but ended up producing a lot of games that tried really hard yet delivered very little.

    I'm glad some-one mentioned Star Control 2, it remains one of my favourites from the 90's, I wonder if it is eligible for this blog?

    1. Since Star Control 2 is a 1992 game, it's a little bit late to add it the main playing list. Then again, if someone is willing to blog about it - and makes a good argument that it really is an adventure game - I am sure we can find a spot to it as a Missed Classic.

    2. I know Moby Games classifies SC2 as an RPG, and I think CRPG Addict is planning to cover it, but the game's adventure game pedigree is stronger than its RPG pedigree, IMO. There's no character stats or leveling up. But there is plenty of talking to characters (with excellent dialogue), an inventory, and puzzle-solving. The action bits will certainly not be to everyone's personal preference, but objectively, the ship combat engine is outstanding. And much of the ship combat can be skipped. I'd love to the game covered here if someone is willing to take it on.

  8. And kudos to Ilmari and Tornado for the effort they put into that last puzzle, with all the suggestions and theories that Ilmari tested I'm surprised he made it out sane, it was a great blog to read through!

  9. Quick note--Zach Tomaszewski, who wrote the 2003 walkthrough I linked to in my post, had written there that he would liked to be informed if someone worked out the logic for the puzzle. I had no idea if he would still be interested in it 15 years later, but nevertheless, I found his contact information and dropped him a note about it.

    Well, turns out he was interested enough--he's now updated the walkthrough to include a link to this post!

  10. This is one of the few games I ever bought for my Atari ST, mostly because it had hardware copy protection, a little dongle you connected to the modem (I think) port. It was a toss up between buying this and some other crpg (Drakkhen?), but I figured I could pirate the other one... Never got too far in this one, so thank you guys for the effort!

  11. Many thanks for this investigation into the puzzle - has helped me playing the game and making a video walkthrough for it.

    I found on my version that the puzzle timer only goes about 60 seconds before it locks out. I think it's from abandonia in dos box 0.74-2 at 3000 cycles.

    In relation to the surprise time limit, there is another sort of time limited puzzle in the game: you can buy a set of 10 deeds from Citizen Bogli Izae but only if you visit him on the first day, after that they vanish, maybe he sells them to someone else?

    I really love this game - had it on the ST originally and it was such a fun world to explore. There are also multiple solutions to many of the puzzles which is enjoyable, for instance instead of the bank heist you can just save up the money and buy the deeds.

    Great music too, one the old time classics.

    Thanks again :)


Note Regarding Spoilers and Companion Assist Points: There's a set of rules regarding spoilers and companion assist points. Please read it here before making any comments that could be considered a spoiler in any way. The short of it is that no points will be given for hints or spoilers given in advance of the reviewer requiring one. Please...try not to spoil any part of the game...unless they really obviously need the help...or they specifically request assistance.

If this is a game introduction post: This is your opportunity for readers to bet 10 CAPs (only if they already have them) that the reviewer won't be able to solve a puzzle without putting in an official Request for Assistance: remember to use ROT13 for betting. If you get it right, you will be rewarded with 50 CAPs in return.
It's also your chance to predict what the final rating will be for the game. Voters can predict whatever score they want, regardless of whether someone else has already chosen it. All score votes and puzzle bets must be placed before the next gameplay post appears. The winner will be awarded 10 CAPs.