Fractal Softworks Forum

Please login or register.

Login with username, password and session length
Advanced search  

News:

Starsector 0.9.1a is out! (05/10/19); Blog post: Painting the Stars (02/07/20); Updated the Forum Rules and Guidelines (02/29/20)

Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.

Messages - TaLaR

Pages: 1 ... 135 136 [137] 138 139 140
2041
General Discussion / Re: .53a First impressions
« on: August 07, 2012, 04:02:56 AM »
Since the best (at least based on my experience) AI-controlled Hyperion build includes 4 burst pd as weapons, i really stopped recognizing enemy fighters as threat at all. Usually they are mostly gone by the time I'm done with my first/second target ship (on a heavy blaster-equipped Hyperion)...

Anyway - i think in current update balance has been seriously tipped in favor of ships vs fighters, since they got new shiny systems and serious AI improvements...

2042
General Discussion / Re: Some AI things in .53a and a few balance issues
« on: August 07, 2012, 03:52:33 AM »
Actually i'd say problem is not usefulness of burst pd in general, but comparison of light and medium versions.

OP: 8 vs 13
Charges: 2(3) vs 3(4) - with extended mags in brackets, must-have mod for ships seriously relying on burst pd.
Sustained flux /second: 70 vs 120
Sustained dps: 55 vs 82
Per shot damage: 110 vs 165

Basically it's never a good idea to use single heavy version instead of 2 light ones - they provide better pd, and better anti ship damage, if needed. Medium slots also has many other applications (at least universally useful grav beams), for light slots you pretty much don't have alternatives on large high techs, unless you want to tank most missiles on shield (Since other pd lasers are not quite as good at intercepting missiles...This might be acceptable for player-controlled ship, but AI seems to be not so good at it - try pd-less Paragon, if you don't believe).

2043
Well, actually fighting against multiple Hyperions, while piloting one yourself is quite fun:). Used modded simulator with 6 Hyperions. Even won sometimes, though this takes specialized ship design, tactics and some luck.

2044
Suggestions / Re: AI handling of phase cloak.
« on: August 07, 2012, 03:09:21 AM »
Seems there are already quite a few changes regarding phasing AI in dev build...
Well, then i'll just have to wait for next version:). Current one sure made noticeable step ahead of previous (especially group vs large ship situation).

2045
Suggestions / Re: AI handling of phase cloak.
« on: August 06, 2012, 01:40:29 PM »
I think a more dynamic use of the system is more "fun" (yay, subjective words) - wouldn't want to discourage it that much. As it stands, which approach is better is situational - and that's good.

I also prefer it this way and just want AI to be at least as adequate at using phase cloak, as it is at using shields.

There's an easier way for the Shade to take no damage vs an Onslaught, though - it could simply stay out of range :)

Sure, but if you do it this way you have Onslaught waste time and flux on phase ship that it actually can't do anything about, improving odds for other allied ships around... Which probably speaks more about importance of proper target prioritizing:)

This does bring up a good point, though - the AI could prefire a bit when it detects a ship coming out of phase, but before it's fully out.

At least currently, process of unphasing is interruptible - so it could just stop halfway and phase back (AI doesn't do this now). Might waste some flux in process compared to simply keeping cloak up, but won't get hit and also will shot in process. At least it requires active action from opponent, as opposed to pretty much overloading phase ship by simply pointing guns at it.

2046
Suggestions / Re: AI handling of phase cloak.
« on: August 06, 2012, 12:42:14 PM »
What this means is, if they use your proposed strategy, I can kill them much more quickly by baiting their first phase with a small weapon, them slamming them with heavy blasters when they unphase almost immediately after.

Actually this also has to do more with AI behavior than underlying phase mechanics - nothing prevents player from stopping unphase process in the middle and immediately phasing back. Though i guess player-as-an-opponent still could somewhat efficiently exploit this pattern, just not in such straightforward way.

It's not just the .5 seconds of no upkeep, it's that most weapons have a travel time and ships won't shoot until they think they can hit you. This actually gives you closer to 1 second of not being in danger of taking damage for the price of one activation cost, which costs the same as 1 second of phase upkeep in most cases. That's part of what makes beam weapons so effective against phase ships. If anything, the activation cost should be steeper (say, 3-5 seconds worth of upkeep) and the upkeep cost cheaper.

Yes, THAT. I actually even recorded small video of how this works out in practice - http://www.youtube.com/watch?v=mMUIxN6ZzMA&feature=youtu.be . 30 seconds of shade sitting in front and within firing range of standard Onslaught, pretty much flux-stable.

I agree, that this approach won't work quite as well against beams, but this fact just means that AI will have to check whether it is within firing range & rotation angle of beam weapons or not.

2047
Suggestions / AI handling of phase cloak.
« on: August 06, 2012, 11:12:11 AM »
Right now there seems to be huge gap between actually (close to)optimal use of phase cloak with currently defined phase mechanics and the way AI tries to use it.

First a simple question - how much does 1 second in phase cost? Obvious answer is - whatever is stated in ship stats as cloak upkeep. However, complete answer is - cloak upkeep + ship's flux vent rate.
Other important points:
-phase is entered instantly
-when you break phase, there is a short period while you are still immune & don't pay upkeep, but also do not vent flux. You also can fire. For further calculations i'll consider it to be 0.5 sec, though i don't know precise duration.

Consider following variant of shade: 125 phase in, 125 upkeep, 250 flux vent, 2500 flux capacity.

First second of long phase = 125(in) + 125(upkeep) + 250(regen loss) = 500
Not first = 125(upkeep) + 250(regen loss) = 375
Instant phase/unphase = 125(in) + 125*0(no upkeep) + 250*0.5(no regen while unphasing) = 250 for 0.5 sec immunity

This means that if ship is shot at less than once per 1.5 second it can dodge by instant phase/unphasing indefinitely (with perfect control though), more realistic estimate would be once per 2 seconds. Even human player can have good enough reaction to engage cloak right after he sees that enemy fired projectile weapon, so it's not impossible (beams are kind of special case, agreed).

Furthermore, while once per 1.5 seconds seems unrealistically rare for ship being attacked by multiple weapons, how the opponent uses these weapons is also important. AI will start firing right after unphase period. So time you get before you need to re-engage cloak = weapon fire up time (weapons like gauss cannon, but for most it's zero) + shot travel time (not quite zero even for beam weapons as it seems). With average speed projectiles at medium range this approach to phasing already starts to make sense, at long range it simply rules.

Conclusion: right now instant phase/unphase usually beats long usage of cloak both because it allows you to continue firing while phasing and stay immune longer (in terms of total *useful* immunity time). What is also important is the fact that it makes enemy waste flux, while simply staying phased lets them recharge it. Anyway - either precise mechanics behind phase needs an overhaul to make this sort of behavior inefficient or AI needs to build it's tactics around it.

PS1: tested it once more and found that unphase time can't be longer than time you spend phased. So if you have perfect reaction and can afford to phase for 0.1-0.2 sec, you can actually get even better efficiency.

PS2: Of course AI has even more obvious problems with phasing too, like staying phased under missile threat, which could easily be handled by available PD, or Doom inevitably killing itself through overload.

PS3: actually, i kind of remember where i've seen AI doing something close to described here - Paragon's Fortress shield handling. Hmm... I wonder if phase ships could be improved by letting them use variant of this logic instead (of course it needs to modified still, since *anything* is threat to them, while for Paragon it's only situations where incoming fire is too much for shield to handle normally).

Any comments on topic and the above wall of text?:)

2048
A cheap counter is 2 omens or the 2 phase ships with the emp,

I don't think even  AI-piloted default configuration will have problems with this. Even if it does something stupid enough to eat a lot of emp damage, you can't disable teleporter(or ship systems in general, except burn drive & maneuvering jets which depend on engine).

2049
General Discussion / Re: .53a First impressions
« on: August 06, 2012, 02:50:26 AM »
I feel like ramming your unshielded ship into another ship's shields at high speed should result in crazy huge amounts of kinetic damage, more than any weapon, especially depending on your ship's mass.

Yes, this would be awesome. Ramming a Paragon's shields and overloading it with an Onslaught in Burn Drive would be totally awesome. It has a counter too - just lower the shields before the impact and take the hit on the armor instead of the insane kinetic dmg to shields. :o

Or just activate fortress shield for a second...

2050
Bug Reports & Support / Re: Fortress shield and soft flux damage
« on: August 05, 2012, 09:34:26 PM »
... I just activate it if alot of weapons are going pew pew on me ...

Actually the best approach seems to be activating it right before big projectile hit & disabling right after (as there is no additional cost for pulsing it). AI can pretty much nullify efficiency of powerful low rate of fire weapons like Gauss Cannon using that (at least to point where you will run out of ammo before you can kill it).

2051
Bug Reports & Support / Re: Fortress shield and soft flux damage
« on: August 05, 2012, 09:24:26 PM »
I forgot about upkeep, right:(... So it's actually about 2000 dissipation (while we are at it - how about rounding absorb coefficient display to 1/100? 1/10 seems too rough) .

I agree that described above  behavior would make sense, if it actually did have chance to fire - but in my test it doesn't: i outrange all, but lance due to optics + targeting unit + keep my distance, and it won't fire lance since i keep shields up.

It activates Fortress shield while not firing any weapons, despite having about 200 spare capacity (i keep flux at zero for test purpose, so no bonus involved). Point about it losing because of fortress shield still seems to be valid (or at least losing much faster than it would otherwise, since i'd have to build up flux to actually overpower it).

Actually, even if we assume that 1800 is somehow borderline enough to get through it's regen under these conditions, build up rate would still be less than 600 that it gets with fortress shield.

2052
Suggestions / Re: Maneuvering Jets
« on: August 05, 2012, 12:38:27 PM »
Yeah, the way AI-controlled Conquest wastes majority of it's flux on jets is really disappointing. Somethings needs to be done about this.

Not so sure whether proposed solution is optimal, but compared to flux-cheap and extremely useful skimmers (without even mentioning Hyperion's teleport) jets eat too much flux for what they give.

2053
Bug Reports & Support / Re: Fortress shield and soft flux damage
« on: August 05, 2012, 11:28:30 AM »
Actually i'd prefer to see it even more pink ;D, since now it's kind of hard to see precisely when it's activated by enemy.

2054
Suggestions / AI needs to counter Hyperions better.
« on: August 05, 2012, 11:00:37 AM »
Phase teleport is huge game changer, and currently AI fails to adequately counter it.
1) AI raises omni shields a bit too late to block first shot after teleportation , which allows player-piloted Hyperion to pretty much ignore shields. Since freshly teleported Hyperion can raise shields quickly enough to block fire from target ships, reverse should also be possible.
2) It expects being able to raise 360 degree shields before Hyperion gets into firing range, which is usually false. Could be countered keeping shields up constantly if an enemy Hyperion is seen somewhere on the battlefield, though that's going to be big handicap against other enemy ships...
3) Ships which have neither 360 shields or omni need to cover each other and avoid staying alone.

Not sure about 2 and 3, but 1 seems feasible.

2055
Bug Reports & Support / Fortress shield and soft flux damage
« on: August 05, 2012, 10:16:59 AM »
Either i got my math wrong at some point or Paragon activates fortress shield against non threatening( less than it's regeneration) amount of soft flux damage.

I used custom Paragon with 4 HILs and 4 Gravbeams (actually it has more, but using only these at zero flux is enough ) against Elite standard layout.
That's 250x4 + 100x2x4 = 1800 raw dps on shield.
Target has 0.5 flux/damage adsorbed and 1380 regen, so theoretically it could ignore up to 2760 dps, yet it activates fortress shield and loses because of it.

Pages: 1 ... 135 136 [137] 138 139 140