Defender capsules
A single capsule launches 1 drone each doing 5 physical damage per shot. They last 45 seconds and will follow you. Cost: 21 RU/capsule.
Clearly this is intended as an early game supporting role. The problem is that a single drone is eclipsed by the SMG with piercing ammo, not to mention grenades. If you go to the trouble of fully automated defender capsule production, you're still limited in your ability to deploy them thanks to 'follower robot count' which pegs you at three defender capsules before you need blue science packs. And that's assuming you want to do any follower count research at all (which is dubious).
Distractor capsules
A single capsule launches 3 drones each doing 5 laser damage per shot. They last 30 seconds and do not move. Cost: 117 RU/capsule.
Massively OP. Forget the tank, this is all you need. Their damage type (laser) scales well against medium and big biters. They're not limited by follower robot count, which means much less time and resources in research. They wont follow you around which means they're sort of a single use thing, but with some practice you can kill one biter nest per distractor capsule, pretty much regardless of how dense the biters/nests are. It's very easy to finish a sandbox game on a variety of settings only using distractors. Only at the very extreme deathworlds might you "need" destroyers.
Destroyer capsules
A single capsule launches 5 drones each doing 10 laser damage per shot. They last 2 minutes and will follow you. Cost: 536 RU/capsule.
Each ach destroyer capsule consumes four distractors in the recipe. But they only last 2 minutes. So in that much time you really need to move to four different locations to make them break even. But you're also limited to the follower robot count, and by this stage of the game you will need to have dumped massive amounts of resources in to follower count to get it up around 20 (a single capsule launches If you're fighting small, isolated bases in the mid-late game, distractors are better because the destroyers won't survive the trip to the other bases. If you're fighting thick nests with hundreds of biters, some of the destroyers will be killed in combat unless you're crazy high in follower robot count. So you might as well use the cheaper distractors. This in turn means you should never really bother with follower robot count research.
Suggestions
- Replace follower robot count research with a component for modular armour - a drone control module. Have basic and advanced drone control modules which increase your follower count by different amounts per armour slot. This should allow a better balancing of early and late game drone use. With Basic Modular Armour, you might be able to control, say, 8 distractor capsules. This would also give some more purpose to the Basic Modular Armour which is a bit redundant at the moment.
- Currently the drone recipes use 4 defenders per distractor and 4 distractors per destroyer as a way to scale resource cost and manufacturing complexity. Why not make use of flying robot frames? Defenders could use one and be small, zippy little drones while destroyers could be a single, larger drone using multiple flying robot frames. This would place drones as a mid to late game weapon, avoiding redundancy problems in the face of the SMG and grenades.
- Make distractors follow you, but differentiate from destroyers by making them a more cost-efficient laser-based drone platform you build in the mid game (lighter, weaker). Later you switch to destroyers because they do more damage at a cost of much more resources per drone, and provide a better DPS per drone control module. Also hint to the player that destroyers are doing more damage with a heavier sound effect, twin lasers or something.
- With regard to ROF and damage research, the current model encourages an all or nothing approach. Switch from the "exponential returns for exponential cost" model to a "diminishing returns for linear cost" model. This will make the first few levels of research more tempting. Follow up by unifying the bonus across all relevant weapon systems (eg: bullet damage and ROF research affects pisol, smg, tank gun, turrets and drones). My feeling is that the research tree is excessively padded out with these weapon upgrade techs. I remember a thread that summerised research costs per category and it revealed the weapon upgrade techs were a huge time and resource sink...
- I think biter nests need to become stronger in line with evolutionfactor, but that's a bit off-topic...