I don't know how I feel about innate bonuses for different movement parts. I was about to say I don't like any of them (that includes flight's 10% nat.dodge), but I can't think of a good balance change if that were to be removed.
But then I gave it another three seconds and realized that since I can't think of a way to remove the flight bonus fairly, adding a bonus to the opposite end of the scale seems like a great idea.
And I thought biomatter would be all for that bonus and this change from the beginning
Increasing the strength and width of the movespeed dodge bonus? I dunno.
I think that would make dodging via flight way too effective. Not on its own, necessarily, but once you combine it with parts like a Cloaking Device and a good set of Maneuvering Thrusters...
To reiterate: The bonus doesn't stack, it's just the bonus for using treads at all, yes? And on second thought, maybe you don't want recoil reduction on wheels anyway. The bad guys have wheels, too. Some days it's more useful to deny your foe a bonus than it is to get one yourself, and all that... ;-)
Right, the idea is if you are using any treads at all there is a simple one-off -X modifier, no stacking. (Probably to be -10.) Tactically speaking, this means a flying Cogmind with ballistic cannons could attach one tread among its flight units, and when it's time to make a stand drop to the ground, activate its one tread to overcome the recoil effect, and become a stationary gun platform with a better chance to hit
Actually, I don't believe there are any hostile robots that are both wheeled and use multiple heavy weapons with recoil. Wheels are generally found on non-combat robots (and even most combat robots, until later in the game, are only equipped with a single weapon, in which case recoil has no negative effect).
And it interesting balancing a game where the enemy is capable of doing all the same things you are, so yeah we have to be careful with where bonuses are applied and all.
A small sidenote: I know you intentionally want 'shoot your way through the game' to be Hard Mode, but when you're selling your game based on the Cool Effects Of Explosions, you might want to make sure a rookie who excitedly installs all those machineguns (it's how you play the game, right? The trailer showed it off!) and keeps installing machineguns is able to do at least reasonably well with the combat-first approach, even if the difficulty ramps up 'more quickly' in the later few levels - so they get to actually play the game while learning how to better a) fight, and b) realise that maybe there are better methods than fighting.
This is actually the case right now. You can focus nearly entirely on weapon slots and do quite well for the first half of the game; it's the latter half where that approach becomes much more difficult.
What's keeping players from progressing are other things like figuring out strategies for dealing with different types of enemies, a standard learning curve for any roguelike (though this one happens to be quite different from other roguelikes, making it somewhat more difficult due to the lack of familiarity--tactics that work in other roguelikes don't work in Cogmind).
In fact, in my opinion the best approach to the early game is actually to be very confrontational and destroy hostiles
before they can form groups and swarm you. What I've noticed is that many beginners do the exact opposite! They have trouble facing a couple robots (due to unfamiliarity with the mechanics and strategy), start to flee, then encounter more enemies and are easily overwhelmed... it snowballs into an increasingly difficult situation. They either get really good at fleeing, and start reaching further areas, or switch to an aggressive enough strategy that they can overcome hostile groups and proceed that way.
Hard mode should totally be 'army of robots', anyway. It's almost certain to be the most hilarious. ;-)
This is totally the case--both methods are fun, but the sheer destruction and chaos that results from big battles often leads to some crazy results.