Tactical AI Kit: Go Outside.

I think this may be my third blog where I start out reminding you in all caps to ALWAYS BACK UP YOUR WORK.

I'm not going to go into detail, but long story short, Windows decided to screw itself vigorously and all I was left with was a flash drive backup I made before my trip to Turkey, with my TAIK project and a few basic small things on it.

Good news: TAIK is saved.
Bad news: I forgot to back up the documentation, and a couple hundred high-res pictures and videos of my vacation.

Here's something I want to run by any potential buyers of the TAIK (If you're out there, speak up! I'd like some idea of the demand I should be expecting so I can figure out a decent price.). My original documentation was the typical resource-style installation instructions, that involve “go to line n and insert n below n.” That usually works for most little mods, but there are a lot of changes all over the engine code and scripts for the TAIK that make that method a time consuming nuisance.

Anybody who knows the program called WinMerge will know that it's a tool from the heavens. You give it two directories or two files, and it compares the two to show you every little difference. It lets you select a difference, and merge it into the other.

So, I could ship my T3D project with no changes other than having TAIK already installed. The new, simpler installation instructions would tell the user how to use WinMerge to merge my code into theirs. You let the WinMerge find the differences between your T3D game, and the AI-only version I ship out, and then you can also use it to merge in the changes. Or, if you're absolutely positive that you haven't modified a file in question, you can overwrite it with mine. It's also a lot harder to accidentally miss something this way, and installing the kit takes under fifteen minutes.

Regarding release, I just need to write out simple documentation about setting up your weapons, players, maps, and adding new datablocks. I'm probably going to do video tutorials for map setup, because it's a lot easier to explain something through video and audio than it is through pictures and text. I'm also waiting for an email back from Derek Bronson about getting the product up in the store (unless he's not the one I should be trying to talk to, anybody with a fancy name tag willing to clarify?). If all goes well, we're still looking at a release of the messy, “the-code-makes-no-sense-but-it-works” version of the Tactical AI Kit before September!


Now, the cool part of the blog that has pictures.

Not being able to cope with being done with Torque after my TAIK project is finished, I've decided to start on an actual, you know, GAME. Mostly something to keep me busy, but if it turns out well and is fun to play, it could sell. A lot of what I'm writing about has already been implemented, some of it is part of my plans for the near future, but I thought I'd share my design ideas with you to get some feedback.

I'm quite fond of first person shooters, just not really the ones built today. Fast-paced is fun, but to force that fast pace, developers have to resort to linear maps and predictable elements (invisible checkpoints), which drags down on the replay value.

My goal is to build a first person shooter that requires some degree of thought. Instead of regenerating health and being able to take dozens of bullets without dying, the player should be physically equal to the enemy. The only thing giving you an advantage is stealth, and your ability to stay stealthy depends on the decisions you make. You take 1-2 bullets, goodbye. If you see two enemies up ahead, you shouldn't think “oh, I'll just bust out mah M203.” You should actually see them as a threat, and think about what consequences blowing them up might bring. Alerting everyone and having backup moving in to search the area is going to make it hard to survive when you're outnumbered and actually vulnerable. You might eventually resort to taking them out with some sort of silenced rifle, or wait for them to pass by if you're worried about the bodies being found. Sure, the gameplay is slowed down a bit, and it won't please a lot of the low-attention-span gamers of today, but completing a mission would actually be rewarding, because you managed to fulfill your objectives when the odds were very much against you.

This means non-linear maps, lots of night missions (night ops are always a lot of fun), and dynamic AI (cough Tactical AI Kit cough cough). This means that the missions should play differently each time, especially when the enemies are randomly spawned and wander about. I've ported most of my original work you've seen in my TAIK demos to Torque 1.5, and that's where I've started the skeleton of this project (this also allows me to fine-tune the AI so they perform in all situations). I also plan on having a few “fast assault” missions, which put you in heavy body armor, with many enemies and a nice amount of teammates by your side, as a way to reward the player for taking the stealth missions as slowly as they were.

My idea of non-linear.

I've created an experience point system. Each kill you make is rated by a number of factors depending on the quality of the kill.


“Interesting” kills rack up more experience. You can see descriptions of my last two kills on the left side of the above image. The most recent one gave me 10 points for killing the enemy, 2 points for the headshot, 30 points for him not being alert when I killed him, and 5 points for “Stranger”, which means he had never seen me in his existence. At the bottom left, above my health bar, you see the experience points I gained while playing this map (92), and all the points I've accumulated ever (8601, after messing around with this for the last few days). I've created a lot of ways to earn extra points with your kills.

Headshot - 2 points for the instant kill.
Precision - 20 points for making a headshot while prone and aiming down your sights/scope.
Spray and Pray – 5 points for shooting an enemy without aiming down your sights (which I purposefully made hard to do so that people don't do it.)
Desperate Times – 5 points for making the kill while you only have 15% of your health left.
Penetration – 10 points for shooting and killing the enemy through a thin surface
Explosive – 5 points for kills with explosives. Grenades, M203s, claymores, etc.
Messy – 10 points for directly hitting the enemy with a projectile explosive.
Close Quarters – 5 points for killing an enemy closer than 10 meters.
Long Range – Starting at 10 points at 90 meters, long range kills give you points depending on how far away your target is.
Fast Mover – If you got the Long Range bonus, you'll get an extra 10 points if your target was moving while you shot him.
Go Outside – No scopes. Congratulations, you just got one whole point for doing something that every complete douche on every game on the planet boasts about and puts on YouTube.
Clueless – Shoot a non-alert AI player. Basically, they're walking around, minding their own business, expecting nothing, and lights out. Typically achieved by players who care about stealth, and since I'm encouraging that, 30 points are earned.
Stranger – Killed an enemy who's never seen you before (only works for AI kills). 5 points. You earn this bonus by remaining undetected, regardless of how alert they are, so you can still get points for being somewhat stealthy after launching that M203 at those two patrols.
Hidden – Only works for AI kills. You kill an enemy while well camouflaged. Hidden in a patch of grass, deep in a forest, wearing black at night, etc. This can only be earned if you got the Stranger bonus, meaning that your victim never saw you, and probably wasn't going to. 15 points.
Assist – Only works for AI kills. 10 points for killing an enemy who is attacking someone on your team. This could either mean that you saved your friend's life, or you just stole his kill. Either way, here's a pat on the back.

A mission involves starting out at your insertion zone, completing the primary objectives, maybe some secondary objectives, and getting to extraction. Primary objectives can involve assassinating a specific target, taking pictures of a weapons cache, eliminating a patrol, planting explosives, disabling a radio tower, so on and so forth. The primary objectives in a mission must be completed successfully for you to win. When those are completed, you make your way to the extraction point. Before extracting, however, you can complete one or more secondary objectives. These are generally the harder ones, such as securing a hostage (meaning the captor must not be alerted by anything!), capturing intel, etc. They're optional, but if you complete one along with the primary objectives and successfully get to extraction, you get many experience points.

What the heck is all that experience good for, you ask? Well, there's bragging rights, or the realization that you've earned that “Go Outside” bonus far too much. There's also the ability to customize your weapons with attachments bought with these experience points. I've modified Dan Buckmaster's Mount Images on Images resource and worked on the inventory system to allow weapon attachments: scopes, silencers, grenade launchers. Someday, I can get some $Prefs in that let me track unlocked features, and you can use the experience to buy new weapons and attachments. In the screenshot above, I'm using an M16A2 with a red dot sight, sound suppressor, and M203 grenade launcher.

Various weapons and attachments, just three of the many combinations.

When a player starts the game with zero experience, they should be started out with basic weapons. Weapons that can do the job they need to do, but still have their disadvantages. For example, before a mission, a newbie is able to choose from a USP pistol, M16, MP5, and an M24 sniper. The USP is good as an emergency sidearm and is helpful in close quarters, and the silencer for it is unlocked at the start so it is possible for the player to try stealth out. The M16 is an accurate assault rifle, but all of its attachments are locked and the three-round burst may be a liability in close-quarters. The MP5 starts out with its silencer already unlocked, but it's relatively weak and is short-ranged. The M24 sniper rifle does not have a silencer available at all, and the bolt action prevents quick shooting, but if you need a scope you can use it.

The idea behind this is that you start out being able to do each combat task: close quarters (USP), regular mid-range assault (M16, MP5), and sniping (M24). However, each of these weapons has its disadvantages, leaving you wanting to upgrade, which gets you chasing after experience points. This is where the replay value happens, as you'll want to try out previous missions with different, better weapons.

I've been adding new features while constantly keeping multiplayer in mind, so everything is networked. Playing co-op missions on a LAN has proven itself to be a lot of fun. Putting on a ghillie suit and sniping enemies simultaneously with a buddy is more entertaining than you can imagine.

My goal: Add onto this game. Make it fun to play, reward the player for playing it well, and having a lot of replay value.

Let me know what you think! People who don't leave comments will be backtraced and reported to the cyber police (those who get this reference, probably many, are winners).