1. A larger gulf between anti-personnel firearms that humans can carry and things that we'd expect to do serious anti-armor duty IRL. For shortness' sake, let's call those categories AT and AP.

Serious stuff like heavy artillery and other "off-board" support systems remain off to one side until reasonably-good methods can be developed to let the AI use them well.

Needless to say, in a sci-fi game, that's subject to a great deal of latitude. In general, AT weapons should be less efficient for killing advancing troops than AP stuff is, but by virtue of flaws such as longer reload times, slower shots, difficult trajectories, etc., not because the weapons can't kill small stuff.

For example, in the Imperatus arsenal, the Melter's more geared towards AT than AP; it has a slow rate of fire and while you can certainly clear a hall with one, you have to aim each shot deliberately.

The PelletBlaster, on the other hand, is AP, with a good spread and relatively fast response and reasonable performance against fleshly creatures tempered by fast falloff and practically no penetration against serious armor.

This mod is also using a lot of Lua behind the scenes to fix a lot of things that are happening in the engine or in the core game design that have never made any sense to me, like weapons tearing up terrain in ways that just don't make much sense, and much of that tinkering will continue.

I'd say that balance, in general, on the weapons is "early". It just takes time and feedback, though; once the general concept of the guns is set, it's just a matter of getting it dialed in. Cost vs. effectiveness is probably dead wrong on a lot of specific cases, but I am cautious about making wild changes until I've heard some feedback.


2. There are much wider gaps between armor categories for better differentiation. There are some new base Materials that fill the gaps in Vanilla, such as something that's a lot closer to "real" tank armor (i.e., well-nigh indestructible by most AP stuff, regardless of range).



As a result of 1 and 2, a lot of weapons may feel "OP"... until you notice what they won't kill. That's where the distinctions really start becoming apparent. Stuff that will convert a roomful of Ronin to slaw will be happily ignored by Browncoat Heavies. This is not like Vanilla at all, in that sense; I did not think that having all of the factions all equally-vulnerable was a good idea, even if MP play becomes a big deal when Steam hits; it'd be a lot better if the big arguments were about which faction was the best to tackle the campaign with or what faction counters what faction in MP, and that requires differences that matter and are worth discussing.

In general, what I've found is that, given the way the wounding system works, GibWoundLimits matter a lot less than the Material a given Actor's body part is made out of; basically, due to the nature of the engine, damage to humans, given their smallish number of pixels, tends to either be catastrophic or niggling, but is rarely in between. You can give the Ronin hundreds of GibWoundLimit points and it barely matters, because they'll still take a headshot and gib or get a limb knocked off by a bullet that passes through their frail flesh. I didn't think it'd work that way when I started this, but testing quickly revealed that hitting the limits aren't the primary cause of death. So, instead of just throwing a lot of wounds at things, I've usually upgraded their armor first, then after some testing they might get more wounds if it makes sense.

Needless to say, that means there's some tug-of-war going on every time I re-visit balance on a given weapon or guy, but it's starting to feel reasonably good in terms of sheer variety; I just think that balance still needs a lot of fine-tuning in terms of costs and availability vs. utility. Oh, and fun

1. Each AHuman / ACrab that isn't basically just a minor clone (i.e., just changed bitmaps) needs to have a Lua movement script built for it, with adjustments for that AHuman / ACrab's dimensions and specific tweaks for speed and suchlike.

Once I'm fairly certain that I don't have any more major fixes to make, I'll probably make that source available to everybody under one of the Creative Commons licenses. Right now, I'd say it's still a little early; that bug you reported and the "air-walk" stuff are probably not the last two things people will find so "awful" that they can't be bothered to play out a real campaign game and see what the mod's really about. I keep trying to point this out to people, but I guess that a lot of the "oldbies" here can't be bothered to play a serious game and just want to play Skirmish with 30K in free gold or a canned scenario they've played a million times... lol.

This step is 100% OPTIONAL... but... if you don't do that, you're going to find out, very quickly that the modded faction is non-competitive if played by the AI (and in most cases, by humans) because it's not able to move as well.

2. In general, mods that were even vaguely balanced with Vanilla (and, well, I'll be frank and say that I've never played any mods for CC that felt even vaguely balanced with the Vanilla factions) would need to check each weapon system and make sure it can still fulfill its intended role. In many cases, minor Sharpness adjustments are all that's needed, perhaps with a bullet script to prevent excessive terrain damage (see PelletScript.lua in the Base.rte/Scripts/ directory for an example).

However, with the more, erm, OP mods, they probably should remain just like they are and get adjusted via costs, so that they're roughly balanced in the Campaign Mode. It'd be pretty cool to have the Space Marines fight the Coalition and it not be an entirely one-sided slaughter for once, heh.

All that said, I'm not exactly holding my breath on conversion of mods, simply because it's not as simple as just dropping in the folders or adding some minor Lua; I'd do an example if anybody would be willing to let me convert their mod, though, just to show how the process works. Hopefully a lot of the stuff this mod addresses will get fixed by 1.0, though.

I think the problem with CC is the same problem MineCraft had, and mastered: how to take fairly open gameplay inherited from a very open-ended engine design and then put a really challenging game design on top of it to satisfy players looking for a challenge.

This irked a lot of the original players, who just wanted to "make stuff". It also made MineCraft go from being reviewed as, "interesting concept, but..." to "WOW".

I think that changes are inevitable, if DRL wants to see it really make it up the Steam charts; the game's been around a long time and just arriving on Steam isn't enough; there needs to be a solid game here.

I also feel like this is the area of the game that's gotten the least attention over the years; updates have fixed a few bugs and added features, but there's been very little sense of goals or even a design doc giving firm guidance. I don't pretend to know why that is; probably the usual muddle and fruitless tug-of-war between a bunch of guys who aren't really game designers and have no idea why it's not gelling even though people obviously like the basic idea.

You see this in some of the silly changes in the tweets: "Revolver Cannon made unbuyable".

Seriously?

After being in your tech demos for years?

Why not just make it useful? Just sayin'.

While I disagreed very strongly with how the problem of conversion from toy-to-game was solved in MineCraft (basically, making time the most precious commodity by making it take so long to get the more important minerals, without which it was literally impossible to win) they at least figured it out and gave people a mountain to climb.

Hardcore on the highest difficulty is very, very hard. The market liked that, so I can't argue that it was a bad solution, merely that it wouldn't have been my choice.

I don't think grinding is a very good solution for CC; play at the highest levels should be about knowing your characters, weapons and twitch so well that you're able to manage your armies and kill efficiently while managing to keep your Brain alive, even when the enemy has an overwhelming resource advantage, is landing new troops constantly and you're the invader.

There are few things as cool in this mod as staring at a base you and your army of disposable heroes have wrecked.

It took a bit of fixing just to get it to run; Crobotech is really old and isn't up to date with changes in files, so I had to bug-fix it just to see it.

Crobotech uses Lua on the characters, albeit in a very funky way, to change the Actor's PresetNames by spawning a clone of itself with a modified name string; I guess that changing the names to be "Lasergun Standard" or whatever was a lot more important than having working AI, lol.

The Crobotech characters air-walk just like everything else does; if you take off and are facing nearly straight up and then hit right or left, you'll observe it happening, no problem.

So, basically, there is no magical fix in this oldie-but-goodie.

That said, there were lessons worth learning. The legs getting "tucked" are an effect of the aggressive stance offsets in the jump paths; I'll have to see if the same idea can't get applied to Vanilla to improve the visual behavior, but I can pretty much guarantee at this point that air-walking:

1. Is a result of some engine code that works in a strange way. The code for deciding whether to run the walk animations isn't exposed to Lua, so far as I can tell; the engine just says, "facing nearly up, somebody hit the right / left key- WALK". Anyhow, that's been reported as a bug and we've offered what we'd like to see for solutions, but we've gotten no feedback about anything we've said, so IDK whether it'll get fixed with a simple altitude check or with a new "in-air walkpath" that gives us a pose we can put characters into. Or not fixed at all, of course, but I figure that's probably a one-liner logic check and while it's not a cheap raycast, that height data could be used for multiple things.

2. Can't be fixed by Lua, or at least can't be fixed until one of the engine folks is kind enough to help me out with the problems I've discovered with how the FootGroups actually work, as opposed to how I thought they'd work (a straightforward IK chain of MOIDS parented to the body).

Short answer is that they can get tuned down to some extent, but there is definitely a point where the AI starts having serious problems navigating again, especially on maps that are more vertically inclined. The AI simply doesn't see delta-v like people do, and I'm not even sure it can be made to do so reliably; where people might triple-jump to get enough delta and height to clear a mountain, the AI just gets stuck in the valleys. Boosting both power (significantly, in the vertical component) and fuel (marginally to significantly or even negatively, depending on the faction) seems to have fixed a lot of the problems in Vanilla, but it's still not where I'd like it to go. Just been busy with the rest of it

TBH, I've thought about two final solutions for these issues, and both are a little messy so bear with me:

1. Just declare them to have unlimited fuel, but sharply nerf their vertical thrust. So it'd feel a lot more like Vanilla, but not cripple the AI (hopefully) nor allow it to bounce as much (which may or may not improve gameplay- the AI currently flies around pretty much like human players did in Vanilla, so fair's fair).

This solution probably sounds the most radical, but really, it's the more conservative choice. Human players in Vanilla have, for all practical purposes, unlimited fuel, because they can do realtime terrain analysis with their eyes and use their delta-v in bursts that keep their total velocity within a sweet spot. Only with a few badly-tuned characters in Vanilla, like the Coalition Heavy, did I ever feel like jetpack fuel wasn't basically "free".

The AI can't get that stuff right now, it's not equipped to do that.

I could probably write it to "desire" that sweet spot, but then it'd fly around the world like players can in Vanilla, and people would moan, even though from a game perspective, it'd be "fair".

What I can't do, and I don't think is computationally practical, is to make the AI climb cliffs and mountains reliably, with weak jetpacks and limited fuel. The problem there is that while humans can look at a slope and almost instantly tell you where the top is, for AIs, it's not that easy; each character would have to perform a few dozen very long raycasts every second, which would get pretty laggy, I suspect.

So it needs one or the other to make it anyhow. Without that feature, the AI's pretty worthless; you can just park your Brain in a hill-top and dig some holes and the AI's not competitive.

So, given that one of the core goals is that the AI must be able to work well enough to give us a real game, this is one of the options that's left, and since I haven't actually bothered trying it out...

2. Keep their power at the current levels, which I personally like because you can execute hill-slides and swift pop-up attacks, but tune down the time to the minimum that will work.

I think what I'll do on that for the next build, since I have to re-tune the jetpack speeds in forward motion, etc., is I'll try giving them unlimited fuel, since that's one crazy idea I haven't bothered trying yet. I'll test it and if it totally sucks, I'll work on no. 2; I think that with the air-walk issue identified if not perfectly fixed and the rest of the major issues dealt with that it's probably time to start actually caring about tuning that stuff. It'd help to get feedback on this on a troop-by-troop basis.

Long story short, getting the jetpacks a little closer to Vanilla is definitely a goal, but it may or may not be possible to satisfy everybody on this, nor do I think it's really practical; it's one of the most crucial things the AI just can't do right and probably can never be made to do right, so I'm kind of looking for a happy middle ground where it works and the AI's moving fast and well but traditional CC players aren't too irked by evil Dummy kangaroos, etc.