Before I start I want to say I like supply drones, as a concept. Destructible ramp ups for economy are cool, and makes you think twice about whacking them on an unsafe well. Requiring players to invest in drones to boom is cool too, there is the risk that you have less military when the opponent comes to raid you, with the payoff of having a stronger econ in the mid/late game. These things own, the stasis beam after scon 2 owns, the art on the drones owns.

BUT!!

And this is a big but, the maths on them are just insane.

For those that haven't experimented much with them, supply drones follow Servos usual stack behaviour for buffs. In most cases, buffs that multiply are applied one after the other to get the final buffed stat. Only notable exceptions to this are multipliers on the base value, which apply before other multipliers and flat bonuses. Anyway, that 1.3x action properties is worth a lot more than it sounds.

In a typical RTS a worker unit gathers at x amount per second, or carries a flat amount back to a drop off point. To scale it up you create more and more workers, in some games having as much as half your pop made up of workers. This means that 2 workers is 100% more than 1 worker, in a linear fashion. I don't think that is the kind of economy BXP have in mind for this but it is important to remember while contemplating the power of supply drones.

When a supply drone beams a well, it takes the action properties and ends up with a new bloom/sec. I do not know exactly what variables the beam changes but putting drones on the well will show the outcome of the buff, for now I'm going to assume it modifies bloom/sec directly rather than indirectly via yield and speed for sanity's sake. So 3 bloom/sec x 1.3 = 4 bloom/sec

Now you add a second supply drone. 3 bloom/sec x 1.3 = 4 bloom/sec. 4 bloom/sec x 1.3 = 5.2 bloom/sec. Or 3 * (1.3^2) = 5.2

If the bonus were additive instead this would look like 3*1.6 = 4.8. As you can see, stacking drones is already paying off compared to spreading them.

Consider 10 drones:

3 * (1.3^10) = 41.36

20!

3 * (1.3^20) = 570.15 bloom/sec

But this is all napkin math, these equations could be laughably wrong (if anyone at BXP wants to explain exactly how it works please do) But what I'm trying to say here is that stacking drones on a single well is better than spreading them out. And so, it is not your typical RTS worker unit that simply scales linearly with numbers. Supply drones improve wells exponentially. You can test this out yourself on mission 1 with the infinite well. If you stack enough of them you'll find the max bloom you can bank and the maximum output of a well. These are reached before you can make the maximum amount of supply drones. More practically you can make between 10-20 drones in a real game and eat every safe well within a few minutes. I encourage you all to try it.

Anyway with that out of the way, onto actual feedback for the devs that isn't 'Draco is talking out his arse about maths'

I appreciate that you tried to keep the economy smallish compared to, say, AOE or SC, even with the ludicrous buff the supply drones give to wells the mechanics of the game doesn't promote 'always make workers'. I'm down with always make workers in most games but it'd be too much for a hero RTS like this. I tend to get pop cap up to the amount I need to make the supply drones I plan to have and queue them up. Basically, workers don't take up all my mental bandwidth in the early game and this lets me use the lead servo. All smiles here.

However, when the drones gather on a well, time no longer has meaning. RTS games tend to have timings, when someone going for a certain plan will be able to execute it. Likewise, these plans usually have timings where they are vulnerable and can be countered. For example you might aim to have a proxy barracks and several men 'o war to attack enemy workers in AOE2, if the other player doesn't see it and you can reach your goal it may pay off and do a lot of damage. If they stop you before it happened you are left with the sunk cost + regret. Servo kind of had this going on, though early imbalances have tended to make 1 strat really good, there were still multiple timings you could try to delay with harassment. This patch I see 1 timing, critical mass of supply drones and the draining of your local wells.

No matter what strat I try to do, all aoe behemoths, 2 blobs of building tech'd gladiators, going servo heavy etc, there are only 3 stages to the plan

-Get wells, get up pop cap, train 10-20 supply drones

-as first well dries up, go scon 2. when scon 2 finishes your second well is half empty, go scon 3 unless the tech you want is available at 2

-build your entire army in about 30 seconds out of an obscene amount of factories, or roll your face across the keyboard at the lab and do what you wanna do. bloom is no longer an object.

First, an ideal example, playing against AI

Some times to look out for

2 minutes: first well dries up

1 minute 40: scon 2 starts

2 minutes 20: scon 3 starts

3 minutes: 7 factories, behemoth aoe tech, 30 behemoth drones

Spoilers: The one that mattered was the first, without sucking that well dry I would not have reached scon 2 and 3 around 2 minutes. Because drones stack so well the next well is gone in 30 seconds, 15 of that being because I had idle drones. The 3rd is gone 15 seconds after that, then the 4th, and a 5th. At 3 and a half minutes I have enough bloom for my army, and all supporting tech I want. The longest part is moving the behemoth drones.

Now you would be correct to wonder, what about a real match? What about if someone tries to stop you before it gets too cray? Thankfully, I have another example, a match with Graham

I set out in this match to Give Some Feedback on gladiator drone building damage tech upgrade But the same early game opening applies, mass the drones, suck up the wells, bang out your plan as fast as you can click.

I actually make some big mistakes here, I leave 2 wells unsucked in the safer base, and neglect to pull my drones back when Graham comes knocking on the door.

Again the timing ends up about 2 minutes to ramp up, 1 minute to get the plan rolling, then a few minutes to pull it off. Graham opts to spread his drones out, 5 a well, this ends up giving him a decent income that allows him to reach scon 3 by the time my gladiators are rolling out. However, while this would normally be a real great timing, concentrating the drones has given me the income to do that AND build a game ending army.

This match is tragic because Graham isn't doing anything that would normally be Wrong in Servo. He has an econ going, steadily propelling him to scon 3 on time, he uses his servos well, opting to ignore mine mid fight and attack into the base, not allowing me to delay him any longer. He walks into my base vs weak opposition, my own servos being totally out of position. 19 supply drones are taken out by a skill shot before I can move them. I try to, too late, and they all die. My economy is RUINED. I might as well not even have any wells at this point, I'm all in. If my attack doesn't work I lose the game, RIP, Rip and tear my guts, complete ownage.

But I have 4k banked at this point, about a dozen factories, the building damage tech, and my whole army. It splits into 2 and attacks straight into the drop ships. Despite my best efforts Graham completely destroys one of these forces, but was not able to save the other drop ship. I tell all my factories to make more galdiators and the 2 pronged attack continues unabated. I finally hit scon 3 just to distract Graham's servos. Finally when a wave of gladiators arrives when servo abilities were on cooldown the dropship falls over.

Something i did not notice in all this was Graham had dropped some airfields, he was going to bomb the heck out of my base while his servos pushed in. This would have been real nasty, but then the gladiators happened. The only thing he might have done to bring this back would be to make aoe behemoths, but I do not know if he had the income at the time to pay like 2k bloom for a decent force.

It seems a bit much to expect raids on the economy in Quagmire within 2 minutes. But because supply drones work like this you have to be in there killing the drones before 2 minutes or the supply drones pay off big time and raids on economy take on less meaning.

Quagmire has dual dropships and short distances between bases, what about another map? I only have 1 example, another match with Graham on Gates.

On other maps the time before your supply drones make it rain is less extreme, but it is still only about 3-4 minutes. On this map you have 3 paths to the enemy, all 3 blocked by a speed bump gate of some kind. In one corner is a line of blocks of shield, susceptible to EMP. On the other are similar lines of armoured blocks. Through the middle is a wall of tier 2 and 3 creeps. You aren't getting to the enemy in 3 minutes to stop this. Graham breaks through to what was an undefended booming base to a dozen factories making aoe behemoths. A few volleys and his whole army dies, a servo blows up in the cross fire, and I chase down the remaining 2. While doing this I'm teching up so these are not beelined behemoths, they're buff as hell.

This match is an open and shut case, sadly, I just had too many behemoths for Graham to stop me rolling through the gate he opened.

Last update the meta was soldier drones with firmware in scon 2 because it was fastest to reach, this patch the meta is anything you want, so long as it begins with a boom. If you don't boom the booming player eats you, no way to stop it without booming yourself. Any strategy you want to pull off will be ready to go around 3 minutes on quagmire and 5 on other maps without dual dropships. While the nomad map, Sandstorm, is a neat concept why would I not drop the dropship immediately and queue up supply drones? They don't even have to be on a well while they're building up, what matters is having a nice blob of them so you can drain every safe well you have, needing only 15-30 seconds each well to suck them dry.

So, having done these 2 matches and finding human players aren't enough on their own to stop a good boom I picture a match with wondible in this meta, with level capped profiles and good gear. What would happen? I think it'd be something like

-Both of us boom and build gladiators, drop a servo with firmwares and try to do some damage

-If we both survive that we both get behemoths with aoe, bomb drones etc and fight an aoe war

-One of us has our economy ruined before we can suck up half our map and it snowballs from there

-If the guy with worse economy somehow survives it becomes sudden death between servos, a long tug of war with 0 income and nothing banked anymore, probably going to whoever planned ahead and got upgrades for the servos

It doesn't sound that great to me. These are all assumptions though but can it really play out another way?

Something I have found odd is that during the dev streams this doesn't seem to happen much, even though you are playing with the same supply drones as the ones on public test. Graham and Jacob tend to have fewer supply drones, when they make more than 5 they tend to put them on different wells, which brings in a decent income relative to pre-supply drone economy. I learned about stacking drones from Jacob though, during one of these streams, so he knows what they can do. He told me one of the developers has a strategy much like mine. Dave told me 2 weeks ago they are able to counter this but I really have no idea how you would. Before raids become a problem I have more bloom than I can spend.

The old economy had some problems, but I miss the feeling of having aggressive creep spawns reclaiming the map while being a potent resource. When wells went dry you had to risk creeping to get more bloom, some maps made this too hard but IMO the mechanic was too fun to remove entirely. I like the supply drones but the way they make me play feels like I have cheats enabled.

ATM I think I'd prefer:

-Creeping as a big part of economy again, with creeping on time and creating opportunities where it is safe to creep being important for good timings

-Smarter distribution of creep spawns, to avoid giant clumps of creeps, while trying to drop them in places that are vulnerable and strategically important, wells for example

-supply drones to stay how they are except the beam buff caps at 5 per well, so they provide a nice boost but not get out of hand and make time meaningless. Would make the player think more about the drones throughout the match, and would make it more likely that some supply drones are near a big fight to use stasis beam in a timely manner

-After about 6-8 minutes a booming player starts seeing their wells wink out one by one, starting about 10-12 minutes for a player focusing entirely on military, but the military focused player should have the chance to do enough damage to delay the booming player, narrowing the time gap in a real match.

To save people having to quote this whole thing, here are some more specific questions

For other players: What do you think of supply drones and the econ changes?

For the developers: What do you want from supply drones? Is there some ideal economy you are aiming for? What happened to creep spawns, were you finding them not fun enough?