Poll: Worst Unit Of The (time interval) Award (IV)

[keith.lamothe]

So, I did a little math (I'm sure everyone is shocked), and the proposed new numbers (20/31/73) produce some peculiar results.
First off, by increasing the value of the harvesters, you've also increased the value of the Exoshield.  A quick set of calculations shows that, for the Mk I harvesters, it becomes worth protecting them if they get destroyed more often than every 750 seconds - about 66% more value out of the shield than before.

In all honesty: does anyone really care that much about protecting a harvester?  Even if it's producing 1000/s, so long as it can be rebuilt with basically trivial effort (if you care, you use engies to speed it up)?

I'm thinking the answer approaches zero as experience increases

My point with the exoshield change was to give it a different kind of utility in the form of shaping AI behavior in the system, as well as removing the bulk of the ongoing cost.  If 750 K is still too much for that... dunno what to say, honestly.  Will think about it.

If you're not going to III, I can't see buying II.

I thought the same, but also the same about Econ IIs, so I just tried to make them numerically similar.

Also, I wanted to confirm your '2 harvester' math, so I understand where you're coming from.  So at 12 planets with an average of 1 metal and 1 crystal each, these should roughly compare to 12 Econ II/III stations in bonus, correct?

No, I meant an average of 2 resource spots of that kind.  So if you have 13 planets (1 HW, 12 captured), you would have 26+ of either metal or crystal.  You might not have 26+ metal AND 26+ crystal, and certainly there are some ways you could pick  planets to not even come up with 13 of a specific kind (even with 4 on the HW), but the idea was that you generally would have a good chance at 26+ of one kind, particularly if you were thinking ahead with that in mind (not that you'd pick a planet solely for resources, but as a significant consideration).

Does it seem odd to anyone else that homeworld owns at least 1/3 of your harvesting resources after taking 12 planets?

It's normally been the case for rather a long time that the homeworlds have a lot of resources.  But I dunno if it's quite that far off; if you pick 4-resource-spot planets you can match the HW with 3 of them.  If you have 12 of them (admittedly unlikely), the HW is only 1/5.

For those curious, an 8*HW game will be disgusting economically with these things.  8*12*73 = 7008/s, and is only 3 HW worth of research to bring online.

Yea, I may just need to make harvester upgrade K cost be multiplied by 1+(x*(number_of_homeworlds-1)) for some x (1 would be too high), though this would be a first.

These will break 10/10 for multi-HW.

That's certainly possible, though with all the changes that have gone at least partially the player's way, it's likely that at least diffs from 9 through 10 will see an increase in AI wave and/or reinforcement size.  The k-raiding and ST-hacking changes do compensate for a lot of those player buffs too, though.

Answer: Change mineral distrubution.  Homeworlds are overloaded with harvester sites.  Reduce homeworld harvesting and get MORE harvesters 'out in the wild' to encourage planet conquering.

I agree that this is desirable to at least some degree, though I'm not sure it's as far off as you appear to think.  My thought is to change it from "12 spots per homeworld, 1-4 spots for most other planets" (there are some exceptions with higher results, I honestly don't know all the rules) to "10 spots per homeword, 2-4 spots for most other planets" (leaving those exceptions as-is).

And the finicky part is that changing this would probably cause resource spots in existing saves to move (some parts of the map are regen'd each time, not serialized).  This used to be the case, at least, but we may have added some safeguards against that last time we ran into that.  I don't think changing raid-engine/etc seeding will impact it, though, as I think that's a later stage of mapgen.

However, one of the 'boosts' for the AoE mines was they could hit mine immune things.  Does this still count?

I didn't change that part: a mine-immune ship cannot trigger an area mine, but it can be hit by the aoe component if something else triggers the mine.

Is it possible to lower rebuild times without lowering build times?

Not with current mechanics, but it could be added if necessary.  Basically have a non-standard "percent complete to start at when rebuilding starts" 

In general you need a MASS of mines for them to have any significant effect.  Though with the health boost I'm not sure of the impact.

You still need a big mass of them to do a lot of damage, but they're waaaaay cheaper now.  Basically the standard mines will be a "sure, why not?" decision when building defenses, and the interesting part is in figuring out how to use them efficiently, as well as the satisfaction in blowing things up   And area and emp mines are there if someone particularly wants to pursue that, the former for 5x as much total dps as standard mines, the latter for a variety of utility cases.

Can the same ship hit the same mine twice?

Yes, but a ship can only trigger a mine once per game-second, but can be hit by any number of the aoe effects from mines-triggered-by-other-ships.

Is it a single collision detection?

Not sure what you mean, but a ship checks its collision circle against the collision circle of each "minefield" unit (which draws a texture that has 16 icons on it, if that's what you meant).  A while ago they used to be placed individually and died on-hit, but that was more cpu/ram cost than it was worth.

I don't personally care about old games breaking.  Either I can finish it up before I do a Beta download or I can live with it.  I also don't play 3 month long games.  I finish up games in a week or two, so that's probably part of it.  That's part of being involved in the Beta testing.  It IS Beta...  *shrugs*

We never ever let a save-breaking bug stay; one of the occasional tests is loading up a game from before the first public release   But sometimes it does result in some fairly odd situations.

[TechSY730]

(lots of text)

Hmm, some of the things you point out are quite insightful.

Basically, a few things that this game that set this game apart from most RTSs:
1. Caps not only on total stuff (via energy), but also on per type stuff (ship caps)
2. Relative weakness of ships considering their ship cap (more formally, the ratio of effectiveness of a single unit / the effectiveness of a cap of units is smaller than most other RTSs (replace the effectiveness of a cap of units with the effectiveness of the practical cap of units in RTSs without per unit type caps)
3. An environment (the AI) that grows in strength very quickly

These three things give AI war a unique feel and strategy. Although many good things come about from these unique things, there are some unfortunate consequences from how they are implemented currently:

1. The AI grows in toughness basically faster than you can upgrade your units, and because the AI starts reinforcing quite a bit even at lowish AIPs, after early game all the way to you hit hitting the upper echelons of the "tech tree", small groups of ships are not very effective.
Caveat: This is not to say that well managed small groups of ships, even ships not built for raiding, can't be effective. I have taken out well defended planets on difficulty 8 with a quarter of ship caps, over several raids. However, they are no where near as effective as similarly sized groups (scaled for ship cap like restrictions) as you would see in most RTSs
2. Due to ship caps, you can't really compensate for the weakness of individual units by building more units, as you will quickly hit the cap

So, how to fix this? Well, of course, I don't have specific solutions, but I do have overall things that should be at least looked at.
1. Make smallish raids more effective
 Some ways to do this:
 a) Increase how effective an individual unit is, but keep ship caps the same. (Basically, make a cap of units more threatening) However, increase mineral and energy costs to compensate, discouraging building to caps early game
 b) Increase rewards for not putting a planet on alert
   i) There is a mantis post about this, but one idea I like is reducing how much AIP on death of AIP increasing structures if that structure dies while the planet is not on alert
   ii) Make it clearer what triggers a planet to go on alert
   iii) Make it harder for a planet to go on alert

2. Reduce how quickly the AI reinforces on low to mid AIPs
 a) To keep this from just simply making the game easier for "free", the AI can seed a little stronger at map generation

I'll add other stuff as I think of it.

EDIT: None of these things are small, so I do not expect them to be "fixed" anytime soon.

What I'm trying to say is that because so much is not stored in the save file but assumes consistent behavior of the map generator, changes to the map generator can cause old saves to "break" (like moved wormholes and stuff)

I don't personally care about old games breaking.  Either I can finish it up before I do a Beta download or I can live with it.  I also don't play 3 month long games.  I finish up games in a week or two, so that's probably part of it.  That's part of being involved in the Beta testing.  It IS Beta...  *shrugs*

Well, I tend to take very long to get through games (like I do regularly have 3 month long games), so I feel a little differently. But you point of this is a beta and as such we are not assured very much stability is well taken.

[Wanderer]

No, I meant an average of 2 resource spots of that kind.  So if you have 13 planets (1 HW, 12 captured), you would have 26+ of either metal or crystal.  You might not have 26+ metal AND 26+ crystal, and certainly there are some ways you could pick  planets to not even come up with 13 of a specific kind (even with 4 on the HW), but the idea was that you generally would have a good chance at 26+ of one kind, particularly if you were thinking ahead with that in mind (not that you'd pick a planet solely for resources, but as a significant consideration).

Ah.  I'm not sure I truly consider the two unique enough to warrant indepent consideration.  You could meld them together for equivalent K price and it would probably be the same unless you decided to be bomber or frigate heavy.  It's not like you heavily change gameplay style based on which mineral you get more of, it just slows you down as you build more converters.

Yea, I may just need to make harvester upgrade K cost be multiplied by 1+(x*(number_of_homeworlds-1)) for some x (1 would be too high), though this would be a first.

  If homeworld resources are adjusted significantly this wouldn't be as drastic an effect.

You still need a big mass of them to do a lot of damage, but they're waaaaay cheaper now.

Yes, but a ship can only trigger a mine once per game-second, but can be hit by any number of the aoe effects from mines-triggered-by-other-ships.

Is it a single collision detection?

Not sure what you mean, but a ship checks its collision circle against the collision circle of each "minefield" unit (which draws a texture that has 16 icons on it, if that's what you meant).  A while ago they used to be placed individually and died on-hit, but that was more cpu/ram cost than it was worth.

Okay, sorry, I'll clarify.  If I park a minefield off a wormhole and as the ships enter they park in place and get into combat, do they keep hitting the same minefield or do they need to move to the edge of another collision detection before they can trigger a mine again?  Basically, it sounds like 'mine lanes' are still necessary, at least for basic mines, to significantly affect a ship, particularly over MK I.  Continuous detonation is also important for things like EMP mines doing a permanent paralyzation on anything stuck in the collision area.

We never ever let a save-breaking bug stay; one of the occasional tests is loading up a game from before the first public release   But sometimes it does result in some fairly odd situations.

Fair enough, I was more talking about changes that intentionally affected older games for new mechanics and simply being incompatible.

[Wanderer]

Basically, a few things that this game that set this game apart from most RTSs:
1. Caps not only on total stuff (via energy), but also on per type stuff (ship caps)
2. Relative weakness of ships considering their ship cap (more formally, the ratio of effectiveness of a single unit / the effectiveness of a cap of units is smaller than most other RTSs (replace the effectiveness of a cap of units with the effectiveness of the practical cap of units in RTSs without per unit type caps)
3. An environment (the AI) that grows in strength very quickly

These three things give AI war a unique feel and strategy.

You missed the fourth piece.  Minor damage to the PvE enemy doesn't weaken it, it strengthens it.  IE: Raid a Warp Gate.  In most other RTS it would weaken the enemy, think hitting a vespene gas harvester.  It would slow them down/do something to get a minor advantage for your comparable loss.  Blow up an infantry producer, etc.  This game doesn't work on the econ attrition mechanic, so it's not a directly comparable discussion, but it gets the point across.  While blowing the warp gate gives you a strategic advantage, it's also strengthens your enemy due to AIP.  The best you can get out of raiding (other than Data Centers and their ilk, which are not 'local fight' raids) is a neutral result of local weakness for cost.

For the price, the fleet ball reduces the cost of the raid to perform the task.  Thus, optimal play is not to use small forces continuously, but overwhelming power once.

Caveat: This is not to say that well managed small groups of ships, even ships not built for raiding, can't be effective. I have taken out well defended planets on difficulty 8 with a quarter of ship caps, over several raids. However, they are no where near as effective as similarly sized groups (scaled for ship cap like restrictions) as you would see in most RTSs

Agreed, and so have I when preference allowed for it.  It's still far beyond optimal play.

1. Make smallish raids more effective
 Some ways to do this:
 a) Increase how effective an individual unit is, but keep ship caps the same. (Basically, make a cap of units more threatening) However, increase mineral and energy costs to compensate, discouraging building to caps early game

I have to disagree with this.  The point is balance.  Increasing cap-power just makes both you and the enemy more dangerous to each other, making it a zero sum change.  To make raids effective they have to weaken the target in some way that is advantageous compared to the cost.

b) Increase rewards for not putting a planet on alert
   i) There is a mantis post about this, but one idea I like is reducing how much AIP on death of AIP increasing structures if that structure dies while the planet is not on alert
   ii) Make it clearer what triggers a planet to go on alert
   iii) Make it harder for a planet to go on alert

This I agree with, particularly #2.  49 fleet ships of non-low caps (blade/maw come to mind) as a maximum to alert status (I think that's the value on Normal caps) is NOT enough to do a raid with when the game is balanced on ship caps, not units.  The only ships that are balanced on a per unit basis in a semi-usable (to the player without a number crunch) are starships, of which *1* ignites an alert.

[keith.lamothe]

Ah.  I'm not sure I truly consider the two unique enough to warrant indepent consideration.

Not inherently, no, but if mapgen hands you a situation where your homeworld has 8 crystal spots and the 4 first planets you want to capture have 4 crystal spots, that would be a reason to consider getting the crystal upgrades (the total benefit would exceed 4 or even 6 econ IIIs, I think), and just converting it to metal if necessary.

Okay, sorry, I'll clarify.  If I park a minefield off a wormhole and as the ships enter they park in place and get into combat, do they keep hitting the same minefield or do they need to move to the edge of another collision detection before they can trigger a mine again?

They need to move to be able to trigger a minefield, yes.  But if they hit a minefield, park for 1 second, and then move again, they will hit the same minefield as they start moving.  Solution?  Riots. Tractor Modues. Lots of little waypoints   And in practice I imagine a dense attack situation would lead to more than 16 enemy ships touching each minefield during the wormhole exiting.

If it's a big deal I could move the minefield check to happen even when not moving.  The cpu cost of the stationary fight would go up, but it still wouldn't be as much as a running fight.

[Shrugging Khan]

So, make small raids with fewer, weaker (, special?) units not cause alert on adjacent planets; make alert a clearly visible and announced state, and reduce AIP increases when destroying AIPI-inducing structures without alert.

Is that the gist of it? If yes, then I like.

[TechSY730]

Basically, a few things that this game that set this game apart from most RTSs:
1. Caps not only on total stuff (via energy), but also on per type stuff (ship caps)
2. Relative weakness of ships considering their ship cap (more formally, the ratio of effectiveness of a single unit / the effectiveness of a cap of units is smaller than most other RTSs (replace the effectiveness of a cap of units with the effectiveness of the practical cap of units in RTSs without per unit type caps)
3. An environment (the AI) that grows in strength very quickly

These three things give AI war a unique feel and strategy.

You missed the fourth piece.  Minor damage to the PvE enemy doesn't weaken it, it strengthens it.  IE: Raid a Warp Gate.  In most other RTS it would weaken the enemy, think hitting a vespene gas harvester.  It would slow them down/do something to get a minor advantage for your comparable loss.  Blow up an infantry producer, etc.  This game doesn't work on the econ attrition mechanic, so it's not a directly comparable discussion, but it gets the point across.  While blowing the warp gate gives you a strategic advantage, it's also strengthens your enemy due to AIP.  The best you can get out of raiding (other than Data Centers and their ilk, which are not 'local fight' raids) is a neutral result of local weakness for cost.

For the price, the fleet ball reduces the cost of the raid to perform the task.  Thus, optimal play is not to use small forces continuously, but overwhelming power once.

Oh duh, that's possibly the most important of all.
There isn't really anything that can be done about this without a) drastically changing how AIP increases work or b) make almost nothing increase AIP except for command stations. If this proves too easy, possibly make AIs put up a stronger fight per planet to make up for this.

[Wanderer]

They need to move to be able to trigger a minefield, yes.  But if they hit a minefield, park for 1 second, and then move again, they will hit the same minefield as they start moving.  Solution?  Riots. Tractor Modues. Lots of little waypoints   And in practice I imagine a dense attack situation would lead to more than 16 enemy ships touching each minefield during the wormhole exiting.

Heh, that's a lot of micro turning on/off the tractors, but that could be fun.  On a side note, is the collision area for a mine its exact visibility on screen, or does it have 'extended' borders compared to the visible area at maximum zoom?  I'm curious because of stacking. 

If it's a big deal I could move the minefield check to happen even when not moving.  The cpu cost of the stationary fight would go up, but it still wouldn't be as much as a running fight.

Not a big deal, I'm just trying to understand them better.

[TechSY730]

Found the post about increasing the rewards for guerrilla warfare (including better rewards for not putting planets on alert).
3226: Incentivizing Guerrilla Raids

[keith.lamothe]

is the collision area for a mine its exact visibility on screen, or does it have 'extended' borders compared to the visible area at maximum zoom?  I'm curious because of stacking.

At max zoom select one of the minefields, the circle that draws around it (not the cloaking one) is its collision circle.  Same deal with other ships.  When those overlap, blewie.  Minefields cannot be placed such that their circles overlap.  The collision check is not quite exact (no exponents were harmed in the production of this range check), though, so it can't be taken too literally.

[Hearteater]

I'd personally prefer mines hit stationary target since I always envisioned them more like Galaxy Quest homing mines.

[keith.lamothe]

On the subject of guerilla raids and "what to do while you wait for your fleet to rebuild", that fits in with some of the stuff I've been giving a lot of thought for the next expansion, so perhaps that will develop into something that will help with those.  But probably not a lot to do about it in the meantime that wouldn't do more harm than good (it used to be easier to raid without alerting, but that wasn't so great as it was).

[TechSY730]

On the subject of guerilla raids and "what to do while you wait for your fleet to rebuild", that fits in with some of the stuff I've been giving a lot of thought for the next expansion, so perhaps that will develop into something that will help with those.  But probably not a lot to do about it in the meantime that wouldn't do more harm than good (it used to be easier to raid without alerting, but that wasn't so great as it was).

Yea, that sort of thing will require significant rebalance of units and overhaul of mechanics. The kinds of which require frequent experimentation, testing, feedback, and adjusting. The kind of attention that you are currently giving (and should be giving, given the upcoming premiere) of AVWW.

So yea, we can wait for that. Right now, getting it balanced within the current setup, despite the flaws the current setup gives, seems like the best course of action.

And really, despite our griping, these issues, while not minor, are not ruining the enjoyment of the game. Overall, the game is in very good shape. Just a few warts in some implementations of details, which unfortunately, some of these warts are in very "central" parts of the game, requiring very careful removal.

[Volatar]

Yea, that sort of thing will require significant rebalance of units and overhaul of mechanics. The kinds of which require frequent experimentation, testing, feedback, and adjusting. The kind of attention that you are currently giving (and should be giving, given the upcoming premiere) of AVWW.

So yea, we can wait for that. Right now, getting it balanced within the current setup, despite the flaws the current setup gives, seems like the best course of action.

And really, despite our griping, these issues, while not minor, are not ruining the enjoyment of the game. Overall, the game is in very good shape. Just a few warts in some implementations of details, which unfortunately, some of these warts are in very "central" parts of the game, requiring very careful removal.

Couldn't have said that better myself.

[Wanderer]

On the subject of guerilla raids and "what to do while you wait for your fleet to rebuild", that fits in with some of the stuff I've been giving a lot of thought for the next expansion, so perhaps that will develop into something that will help with those.  But probably not a lot to do about it in the meantime that wouldn't do more harm than good (it used to be easier to raid without alerting, but that wasn't so great as it was).

Don't misunderstand.  Though I like options, I'm really just discussing one of the ways this game is heavily differentiated from similar games in the same general concept.  It's not really a gripe, per se, but a discussion of mechanics.  In general, this game is an inverted triangle to difficulty as you 'gain speed'. 

Part of the reason the pacing for this game seems odd, particularly to 'older hands' who are more used to all the little details in between, is because of the surge-wait optimal method of playing.  It would be best if we returned to base to pick up new ships continuously, actually, but the time that can sometimes be involved in that when you're warring 3 worlds out can be prohibitive.  So can the K for the Factory Ships.  There's little reason to raid *1* outpost, for example, when all the reinforcements will just go to a different guardpost in the same system, making the rest that much harder to raid. 

Of course, some systems require your entire fleet to 'raid', but that's more attacking the juggernautt in pieces than 'raiding'.

My concern here is more about the mapgen resource balancing, but as mentioned, that'll take some mechanical tinkering to change.  Until then I can live with the dynamic of economy being balanced back towards harvesters a while.  It'll certainly make my low-world high AI games easier.