(or, Where This Game Came From)
This is a somewhat rambling essay. Go here for a not-so-short, but more-to-the-point summary. Look on the right for a more focused post. Scan this post for something interesting. Or just read it.
I started with the idea that I wanted a war game that would play like a two-player board game. You set down the cardboard squares on the hex grid and the best tactician wins. Next, it had to be a computer game, because board games:
- Take too long to set up before hand and clean up afterwards.
- Require long, frequent interruptions to check the fine points of the rules.
- Make you waste ages waiting for your opponent to move.
- Force players to cheat, because they honestly (or dishonestly) can’t remember where a unit started its move, or whether it had already made its move, or whether it’s even allowed to move this turn, or….
- Don’t allow for hidden movement without referees and elaborate props.
Miniatures games have all these problems, plus a few others.
So it would be a computer game then. From there:
Real Time Tactics:
The Real Time Strategy genre caught my attention as a starting point. A player is able to observe and act 100% of the time. No time is wasted waiting, and waiting, and waiting for other players to move. Most RTS games involve a lot of busywork, but I could get rid of that with no problem. The closest thing to what I wanted was the game Praetorians. This had no resources but was otherwise quite RTS-ey. The tactical part of Shogun: Total War (I played the original a lot) came closer still.
A problem with both these games is that it is often hard to tell what is going on. The troops seem to slow down or even run or lose fights they should win for no reason at all. No doubt this has to do with realistic problems such as moral, fatigue, and the like. With a slower pace or more play I might have started to pick up on it. I wanted, however, to add certain complications to my game, so I thought I’d better keep the basics really simple. Therefore I abandoned poor morale, out-of-ammo, out-of-fuel, and fatigue. (For the tanks, anyway. The players would have to look after themselves.) A few simple numbers, easily available and mostly unchanging, would tell the player everything he needed to know about how two tanks will fare against each other.
Another problem with Praetorians, and every other RTS I’ve played (though not Shogun), and a few other games (such as Sid Meier’s Civilization—pick almost any Roman Numeral you want—played in simultaneous mode with a a time limit) is that important things tend to happen on unwatched parts of the board. You think you’re about to win, then suddenly find half your army has disappeared—and half your buildings/walls/towers/cities if you have such things. Or suddenly a large enemy army appears out of nowhere. (Okay, this can definitely happen in TTA&A, but in TT&A you should know, at least, that an enemy army has gone missing.) In Civ, you never know that a neighboring civilization is spending all it’s income recruiting a huge army as fast as it can until, suddenly, it pours over the boarder in overwhelming numbers.
There’s a place for hidden information, for making a player go and look to see where the enemy is. But there is also “giving” the player information in such a clever way that they don’t see it. Hence:
Giving you the information you need, when you need it, where you need it, and letting you react immediately—that’s the critical feature of any real time game. Perfection is impossible, but no player should have to ask, “Why wasn’t I told?”. If enemy tanks are moving towards one of your key positions, it should not take several minutes of study to find them when they are on your maps…somewhere.
I wanted TTA&A to play something like a board game, or turn-based game, where you could take minutes to methodically check everything on the board. But with TTA&A you should need only seconds, not minutes; real time with only the feel of turn-based. I’ve used a bunch of techniques to bring this about: fast zoom-in/zoom-out, grab the map and drag it, multiple maps (as many as a player wants, all synchronized), large-area maps that control normal maps, and a unit tray that shows the status of all your units in one place. Okay, you have to look for something to see it, but if you do look for something, you will see it—and fast.
I also enhanced Unit Doctrine, so a tank held in reserve could act on its own, running away rather than remaining motionless until its player eventually noticed it had been destroyed. This does complicate a unit’s behavior, but I’m convinced it simplifies game play by giving the player one less thing to worry about.
This C3I, the ability to know what is happening and respond, is at the Core of Tracks, Turrets, Armor, and Attitude. It is what sets it apart from most other computer tactics and strategy games and what makes it worth playing.
C3I and the Fog of War: Okay, there are things you don’t know in TTA&A. You don’t know what kind of tanks or army your opponent is going to field. Once the game starts you don’t know where the enemy is or where he is going. But I hope these are interesting puzzles that can be managed, not something that makes all thought and planning meaningless. I think there is a clear distinction here; I don’t regard it as a fine line. I believe I’ve got it right. (Though maybe it could be righter.) I’d like to know if I’m wrong, or if there are ways to make it better.
My original plan was to do a historical war game. I could set up generic rules for all sorts of pre-20th century military units and let players pick a date and select any army they wanted from that time or before.
But tanks seemed a lot easier to program. I now doubt that they were (though I haven’t done a historical game yet), particularly with the need to pick targets automatically. However, trying to explain hoplites and catapharcts to the vast majority of wargamers might not be practical, while anybody can understand a tank—a cannon, armor, wheels, and a motor.
Another option for units was warships. 20th century battleships, etc., are like tanks—guns, armor, propellers, and motors. There are minor complications, like certain guns that can’t point in certain directions, and deck vs. belt armor. In fact, I did steal one naval complication for my tanks: the use of big guns to shoot at big ships and little guns to shoot at little ones—I didn’t want the tanks to be too simple. In fact, at the game’s level of abstraction, there is almost no difference between tanks and battleships.
The real problem with naval warfare, though, was the lack of terrain. Terrain is another complication, but it’s a good one—one that makes the game more interesting. Ergo, tanks.
I did not much consider space ships. My tanks don’t model real tanks very closely for a variety of reasons. The advantage of space ships is that, as there are no real ones, any space game’s ships model them with absolute accuracy. Hence whole vessel, the weapons and defenses and everything, can be designed purely to improve the game play. But again, there’s not a lot in the way of terrain. Also, my tanks aren’t very real anyway, so I lost nothing there either.
I have two other problems with space ships. The first is that every space game I’ve seen is 2 dimensional, which is ridiculous. In fact, in games, space war is just naval war with really hokey warships. The second problem is, if you do go to 3 dimensions, how do you display fleets in 3D? The methods I use for locating tanks in TTA&A, as well as the methods I use for getting visual range, travel times, and figuring blocked shots all work in 3D. I really can track 3D motion. But how do I show it? Without, that is, confusing a space admiral who has to make correct snap decisions?
(I should add that I love most of the space games I’ve played. Their creators can focus on creating a great game without any concern for any military reality, and they do. While doing “tanks” instead of space ships, I hope I’ve grabbed some of that.)
From a completely different direction, I was inspired by Combat Mission: Afrika Korps. This tank-heavy game plays in a series of simultaneous one-minute turns. Each player spends the time they need to arrange their move, then one minute of battle time passes as the various tanks move and shoot as directed. It is very realistic, using real World War II tanks. And Afrika Korps improves over other games in the series by ditching the infantry. (Infantry tends to bring any WWII-or-later game at the battalion level and above to a screeching halt. They move too slowly to keep up with the tanks, and are soon out of the fight.)
That game teaches a lot about handling one, two, or three tanks, but for my goal of a company—or even battalion—level game with two players, it’s too slow and detailed. But it definitely showed me the possibilities of swarms of tanks shooting it out.
Civilization taught me rather a lot about exploration and hidden movement. I’ve spent hours (or maybe it’s days—I hope not weeks) exploring worlds in that game. A lot of times, after uncovering the whole map, I’d drop that game and start another. I definitely wanted exploration in TTA&A.
Exploration is a basic feature of RTS games—and a very awkward one. If it hadn’t been for Civ I would never have considered using it. It requires serious concentration by the player, and in an RTS game that concentration is usually needed elsewhere. After finding it impossible to keep up with AI players, I ended up adding doctrine (that is, AI lieutenants for human players) for exploration; some tanks will automatically explore, and all tanks can be told to do so.
I don’t like automating things. Making decisions is the human player’s job. More importantly, if the AI lieutenants make bad decisions, it rather ruins the game for players. My automation philosophy, here and in other places is: I don’t do it, but if I do, I do it perfectly. So I put a lot of work into the automatic exploration. That, coupled with some of my other command-and-control features should provide the thrill of exploration without distracting the player from the clash of the big tanks.
Terrain stays hidden until you explore it. (Actually, everything is hidden unless a friendly tank is in range to see it, but most things don’t change, so once found they stay on the display.)
Fog of War and Reconnaissance:
Enemy units, however, move, and so are visible only when in range of your tanks. This is a basic problem faced by real commanders in battle, and I considered it a critical feature of the game. It’s also something that can be done in a computer game; it’s almost impossible in a board or miniatures game. It does complicate the player’s job—and I’ve tried to simplify other parts of the game to make up for it.
I almost made special sensors available (at a price) to let tanks see farther than the standard 3 kilometers, plus special countermeasures to hide them. I think it would add a lot to the game, but I was afraid it would make it too hard to know when you weren’t seeing something because it was hidden and when you weren’t seeing it because it wasn’t there. I’m still wondering, but my feeling is that I could do more with reconnaissance.
Build Your Own Tanks:
One thing space games tend to have that I love and wanted desperately was the ability to design your own battlewagons. (Tanks, ships, weapons, vehicles, units, …— the word “battlewagons” sort of works for all of them, I think.) I loved the original Masters of Orion for a number of reasons. It was a straight Conquer the
World Universe game with no SimCity (SimPlanet?) nonsense. Most of what I liked did not apply to TTA&A, but it did let you design your own starships—and the design mattered. At a given tech level, you could build very different weapons systems that behaved very differently, had to be commanded very differently, and had very different uses.
To win at TTA&A you need a variety of tanks for exploration, recon, observation, guarding control markers, holding key ground, taking key ground, catching and destroying various types of enemy tanks, and luring enemy tanks away from from important ground.
The Armory Mode lets you look at existing tanks in detail and design new ones. Just type in the numbers you want, and the Armory adjusts the other numbers that need to change—most commonly tank weight and cost. There’s a default set of tank designs when the program is first run. You can add designs from another armory file. You can create new ones. And there’s a calculator to tell you how well a particular gun will work against another tank’s speed, size, and armor.
(I used this feature to design the current AI armies, the default player tanks, and my own, good weapons of war. This construction work is an intense game in its own right. It teaches you how to make tradeoffs—you can build the ultimate tank if you can figure out how to pay for it.)
Build Your Own Army:
Most games give you the army you are to fight, already formed—often for historical accuracy. Unfortunately, many of these give you a bunch of useless units and maybe one or two items that can actually do things. So you sit there with 30 tanks that have no purpose other than to give the other side’s helicopters, attack aircraft, and artillery target practice. Any sane military procurement would have spent the tank money on anti-aircraft and counter battery weapons. Perhaps the historical army builders were really stupid, or, more commonly, the rules and the units’ abilities don’t match the historical reality.
Some games, however, let you build your own army, purchasing each unit after careful thought. These games give you more power, more responsibility, and make you feel less like a marionette, less like you’re following a (rather dull) script—all the things a good game should try to do. I’ve always liked these games and I regard organizing an army to be as critical as maneuvering one in battle. So building your own army is an important part of TTA&A.
Command Your Own Army:
TTA&A is all about commanding a bunch of tanks—an army as computer games go. (A real “army”: 1000 tanks. Computer game “army”: 100 tanks max.) Being a tank game, the action takes place over a battlefield 20 kilometers or more across. Being a real time game, things happen at widely separated spots at the same time. Being a tactical game, the position of each tank matters.
Real Time Strategy games dodge the issue of tactics and handle it poorly too. The combat mechanics are somewhat simplistic. Melees are often one mob piling into another. Lines and columns and flanks and such don’t make much difference, so the most primitive commands work fine. Then the individual units think for themselves, so even the most primitive commands are not needed. I’ve done very well at many of these games by focusing on developing resources, building the right buildings, and cranking out as many troops as possible. I then direct these troops at the enemy in a very general way and forget about them. My usual opponents, on the other hand, carefully arrange their troops and supervise their battles.
My troops fight just as well without me, and by focusing on construction I end up putting a lot more troops in the field. I win easily. My opponents, on the other hand, have the big advantage that they, unlike me, have been playing a war game.
Which brings me back to TTA&A.
The positioning of the tanks—the tactics—matters a lot. You need to run when outnumbered, pursue when more powerful, make sure none of your tanks is blocking the fire of the others, and make sure no one tank can get blasted by enemy units too far away from your main body to be hurt. (I cover this in detail in How To Play and Controlling Lots of Tanks and in a few other places.) Tank fire and its effects are very carefully calculated. The odds work just like they would in a miniatures game with (unloaded) dice or, for that matter, in real life. I provide a number of tools that allow the player, theoretically, to figure out in advance what will happen in a melee. In practice, of course, things happen a bit too fast for precise calculations, but after a few melees a player can get some feel for how one will play out.
The Proving Ground:
There are all kinds of numbers and calculations you can do to figure out how one tank will fare against another, and with a few more calculations you can figure out how one tank platoon will fare against another in a straight fight. I even provide calculators to help. But in reality it’s hard to make sense of the numbers, and once the tanks start to move and take odd positions the numbers become just plain wrong. To get a feel for how your tanks, and their formations and tactics really work under fire, you need to try out small fights, repeating them several times with minor variations.
I played a game, Space Empires IV, for a bit that let you design your own space ships. Unlike Master of Orion the design differences were not that dramatic, but some choices could make a difference in a fight. This game let you set up melees among various types of space ships to test them out under fire, which was quite useful and which would have been even more useful had there been more design choices.
In fact, this capability got into TTA&A without me deliberately putting it there. You can run any number of player sessions on a single computer (or on two computers side by side). The tanks will do exactly as told unless given specific doctrine. It is more a matter of player realizing he can set up these test battles. The player can, in a turn-based manner, set up each set of tanks and move them into range of each other. This usually gives the best sense (“information” is not really the right word here) of which tanks are under-gunned, under-armored, or under-powered. (Okay, all tanks need bigger guns, more little guns, thicker armor, and double the horsepower. I should perhaps say, “…which tanks lack an optimal balance of firepower, armor, speed, and cost.”)
More elaborate tests are available. A player session can be set up from the start as AI, but in fact the AI can be switched on at any time. An elaborate test of tank combinations can be made by setting up a complete game, selecting the armies, moving them into a test position, then switching on the AI on one or both sides.
The AI doesn’t play like a human player. It can juggle more tanks on the one hand and isn’t really as smart on the other—if we ignore beginning human players. But it won’t tweak its decisions to favor your pet theories about which designs are best. If the emperor has no clothes, the AI will say so without hesitation.
No Artificial Humans:
I’ve tried to leave the human element out of TTA&A; you can view all these tanks as remote-controlled robots. Why?
- I’ve always found the behavior of people in battle to be quite rational; they stand and fight when it’s reasonably safe to do so and run when it isn’t. I think the player can be trusted to take care of his tanks in a similar manner. (It helps that in real tank battles most crews survived their tank’s destruction and so were a lot quicker to fight to the “death” than were other soldiers who couldn’t put quotes around “death”.)
- I think the arithmetic of battle has its own fascination. Even if one wants to understand the human element in battle it’s first necessary to understand the mechanical element. Courage, patriotism, confidence, enthusiasm, faith, obedience—these things don’t mean much to incoming artillery rounds.
- While I cheerfully do computer simulations of guns and armor and movement, I’m a bit reluctant to do the same with people. (Not that I never will, but that with TTA&A I’ve pushed my simulations far enough for one game.)
Games don’t have to be social events, but there’s nothing like lots of people all playing the same game. And using computers to keep order and keep all the players active makes it even better. I wanted a game that could handle any number of players, so I built the unlimited number of players and the team concept into TTA&A right from the start. This is also a reason for keeping the user interface simple enough to run on primitive computers; it makes it easier to get everybody a computer.
Teams: While the game itself does not attempt to model human behavior, using real humans does the job far better. Indeed, while the tanks and battlefield are all simulations, dependent in part on the imaginations of the players, the interactions of the players on a team are the real thing. Currently the basics are there: each player gets his own army and can cooperate, or not, as they see fit. The limits to team size depend only on the power of the host machine. The game runs in windows, allowing other communications software to run in parallel, so there is a lot the game does not need to do even if all a team’s members are not in the same room.
I do think the game could do more to support a command structure and cooperation. It should be possible for players to give game units and money to other players. It might also be interesting to allow certain players to be formally designated as, say, the brigade commander and battalion commanders, and allow them to assign and reassign tanks in their command to the various company commanders. An army with a feudal structure similar to the current system could be pitted against an army with a despotic structure. (If this interests anybody else, do let me know.)
Multi-Teams: A game should be played between two sides. There are multi-sided games, and I love them, but TTA&A is not really one of them. The machinations of Diplomacy would not work in a real time game. So don’t have more than two teams in a TTA&A game.
On the other hand, there’s nothing like a good free-for-all. So TTA&A fully supports this. My current guess is that there is not going to be a lot of diplomacy (lower case “D”, and no italics) in a game. But there’s always somebody who lets two other players destroy each other, then mops up. I intend this to be a game of tactical skill, not psychology. To limit that tendency any player who destroys an enemy tank gets two thirds of the price of the destroyed tank, with which they can build themselves new tanks. This causes a game to drag on for a bit, but it rewards players who fight—if they fight well.
Client-Server: I’ve known some extremely competitive people. If I wasn’t careful how I put the game together, each would be sure the other player or players had some unfair advantage. As a result, the program can be run as a player or the host. Each player session connects to the host, which actually runs the game. The host cannot distinguish among the players. (For instance, it has no idea which players are AI and which are human.) Nor does it trust the players in any way. (For instance, it decides for itself the cost of each of a player’s tank designs.) I would not trust the system for high-stakes gambling, but it will keep the players honest in a fairly unfriendly game.
TTA&A was originally to be a game between 2 human players; no AI need apply. I soon found I needed AI players if only for testing. I needed to try out different tank combinations, and it was hard to get the feel of the game when playing both sides (the downside of real-time versus turn-based). Also, extremely heavy action was needed to track down all the multithreading software bugs. So the AI had to get better.
Then I realized that anyone playing the game for the first time would probably try it by themselves, just to see what it was all about. The AI had to put on a good show just to make the game look good. Then there was bound to be times when 1, 2, or 3 people would want to do a free-for-all with 20 players. Good AI is needed again. Then there’s cooperative play—man versus machine, 3 humans against AI. Then there’s practicing up so you don’t look bad in human versus human games. And testing out various army combinations against each other on your own. And when you’re too tired to think, you can entertain yourself by watching 2 AI’s battle it out—you might learn something.
I never did get the AI to use formations. On the other hand, the AI can manage a lot of tanks more easily than a human. I left the AI a bit overly aggressive; it attacks sometimes when it should run. I felt most beginners would get annoyed at having to chase an enemy all over the map, and I don’t want to annoy beginners. A bigger problem is that the AI tends to scatter its forces. This gives a clever player the ability to mass against portions of the enemy, defeating a larger army with a smaller one. I left this tendency there because I think it teaches the beginner a useful lesson about concentration of force as well as building his confidence.
Note: when the AI figures out where you are, it will concentrate on you, with no nonsense. This is a temporary, start-of-the-game advantage. Also, this scattering may put all the AI’s units precisely where they want to be to grab control points or blow away your scouts, so it’s not a guaranteed human advantage.
The actual tanks were tricky. They needed to be very simple because a player would command a lot of them—I call TTA&A a company level game, but I’d love to have a player command a battalion or regiment or brigade. On the other hand, they needed to be complicated to make for an interesting, combined-arms game with with many radically different sorts of vehicles.
For a start, I stuck to tanks. I’d love to add infantry, wheeled vehicles, ground effect vehicles, and helicopters. Maybe even attack aircraft and artillery. But for now everything’s a tank, with a frame, an engine, a few guns, and armor.
The game tanks are modeled rather like real tanks. Add armor and they weigh more, move slower, and take less damage from enemy fire. Double the horsepower and they move faster. The bigger they are the easier they are to hit and the more a given thickness armor weighs. Generally more size give more fighting power for money spent. This is a natural tendency due to the square-cube law; the bigger battleship is usually the more powerful battleship. Of course, the law can work in reverse for tanks, particularly regarding wheels, and tracks. However, favoring size discourages players from building large armies of little tanks, thus making their armies easier to control, so I made the nature of tanks in TTA&A downplay the problems of sizie.
The guns are modeled like real guns. Smaller targets are harder to hit, just because they are smaller. At a distance, vehicles with a high power-to-weight ratio are harder to hit because their location when the armor-piercing round arrives is harder to predict. Close up, big guns have a hard time moving fast enough to keep up with small, fast tanks. And big guns do seem to switch to high explosive rounds when firing at unarmored tanks. (At Leyte Gulf in WWII, some American destroyers picked up nice, neat, 18 inch holes where a Japanese AP shell had gone through. And at one point in that war an American plane came back with a nice, neat 8 inch hole in it’s wing from a Japanese cruiser. Big AP rounds need armored targets if they are to do any damage.)
The tanks had to fire automatically. There was no way a player could direct a single tank ‘s fire without changing the nature of the game. So I put a lot of effort into making the tanks shoot intelligently. Hopefully this shows up in test battles. Tanks will try to knock of damaged tanks first, to quickly stop them from harming friendly tanks.
The tanks shoot a lot more quickly and more often than real tanks would, and they never run out of ammunition. This isn’t realistic, but anything else would give the player too much to worry about.
Tank movement is a bit crude. They move in straight lines at constant speeds and turn on a dime. This isn’t too bad for a small tank, but is a bit silly to see a 500 ton behemoth do zero to 60 instantly and do 90° turns at top speed. I’ve looked into changing this, but anything else would require a lot more programming, and it makes the player’s job a lot harder because the vehicles don’t follow the precise path he set out for them. (And following roads gets more interesting.) I am looking into this, because I might want to add jet aircraft and big warships (on the rivers and lakes) to the game someday, and they need to move in graceful curves.
The tanks had to fire automatically; I could think of no way around it. So I put a lot of effort into making the tanks shoot intelligently. Hopefully this shows up in test battles. Tanks will try to knock of damaged tanks first, to quickly