P.U.R.E. 0.8 RC3

[zwzsg]

I'm hearing a lot of resistance about the pay-now system. I guess I'll drop that- it's not really that important anyhow, compared to adding on a third resource to limit exponentiality (i.e., make the game less "spammy") and addressing the fundamental issues of Overmind's economy vs. Resistance.

Don't listen to them. Make it how you want it to be. And preferably as far as TotalA as possible. After all, are you doing Pure Annihilation (just like TA but more balanced), or your own game? And don't listen to Alantai Firestar when he says you have to be stay to TotalA for AI to works. When I asked Alantai Firestar help about making his AI work with Kernel Panic, he kept denying obvious bugs and this lead nowhere. Then KDR made a small LUA AI (600 lines, including all the empty lines), and it just worked. KDR Lua AI was also very easy to edit, to expand, and to make it use custom special commands and custom game mechanics.

So:
- Don't listen to people wanting to turn P.U.R.E. into a TA clone.
- You won't get help from C++ AI programmers.
- Lua AI are easy to make and can handle alien gameplay.

[1v0ry_k1ng]

looked through the pure package and single player. gui, missions, help o matic are totally amazing to the max. Im in love with it.
the game on the other hand is painful slow paced; ground units move slowly and build very very slowly in comparison to the time it takes them to die in combat. it just isnt much fun. I reckon you should give lua and gui and graphics etc a break, and work entirely on the gameplay- all the shell is in place for an epic mega awesome game, now you just need to fiddle with movespeeds and buildspeeds till you get a ratio that feels slow enough for intelligent thought but not so fast as to be a reflex driven click-fest

[AF]

zwzsg thankyou for lieing through your teeth and attacking unfounded without actually reading anything.

I did not say argh should drop his resource system. My objection was to his iron curtain refusing to give details which could have far reaching complications.

Now i tried to help with kernel panic and frequently got the phrase "it just doesn't work, it lags" to which i said there wasn't enough details, and you all insisted there were details, and then finally told me a lot of details and insisted they were there all along. I knew there was an issue with lag, but the feedback i got from kernel panic developers was paradoxical and they complained of issues then denied they were ever issues then said id ignored them. i tried to fix the primary issue, the lag, and made some progress, but not enough, and needed more information, consistent patterns I could recreate.

The whole situation was blatantly mismanaged and I was attacked from all sides despite ploughing hours into mystery meat bug searching, all the while i was being blamed for everything, and being flamed and attacked publicly such as what you did there.

Now here we are again at a moment where information argh has is not being shared, yet argh is complaining that AI developers are doing nothing, thus he witholds information, causing a feedback loop from which only he can break free because he has the encessary means to break the cycle and nobody else does.

I do think that it is good that there is more for AI to encompass, however, you cannot expect AI developers to simply arrive and do work of their own accord, especially when you have a habit of waltzing into that area of development, preaching what they should and shouldn't do, then pissing those developers off on a regular basis by talking down on them.

There is no incentive to work on P.U.R.E support right now, nothing but a frosty reception, and a huge gaping lack of needed information. All this at a time with the lowest ever numbers of AI developers aswell as the biggest slowdown in AI developers in springs history.

What's more attitudes such as those in the post by zwzsg, is that they harm AI development further by tarring the efforts of AI developers who have spent thousands of hours of their time, by writing a simple lua AI for a mod with far simpler dynamics. While a lua AI is appropriate in some situations, there are issues the kernel panic AI does not have to deal with, such as per unit behaviors which would be wasteful in a kernel panic AI, or multithreaded building algorithms. Zwzsg, while you may not appreciate what non-lua developers have done, do not mess around with other content projects out of spite.

[Guessmyname]

As far as I can read it, the KP vs AF flamewar basically boils down to this:

Zwzsg misunderstands AF.
AF misunderstands Zwzsg.
Insults fly,
Both are two stubborn to stop.

Shut up the pair of you and get on with life.

[Argh]

the game on the other hand is painful slow paced; ground units move slowly and build very very slowly in comparison to the time it takes them to die in combat

So, you're saying you'd prefer a system where they built maybe a bit more slowly than they do now, but also died a lot more slowly?

I'm not sure that would be an improvement, but that's just me. It'd make it more nub-friendly, which would probably be a good thing... but it would also lose a lot of the charm- everything would be slowed down a lot, range wouldn't matter as much as raw damage output, and it'd generally feel less like a fast-paced game, which I think would hurt it on the maps that it really shines on (small ones).

Does anybody else have strong feelings on that? Making a mutator to double / treble hitpoints and double / treble build-times to compensate would be easy enough, I suppose I could do that.

However, I heard enough whining about how it's so "hard" to raid with a prior version, which I thought was rather off-base... unless by "raid" we mean, "rush nubs completely out in the first 3 minutes", which the game's purposefully designed to make hard. With even basic turrets up, it should be very hard to rush. Porcing early can and should be a perfectly-valid strat, although in the longer term, it shouldn't work vs. units with large enough range advantages, and in fact that's the way it works right now- try, just for giggles, each type of unit vs. each type of turret, and you'll see that, other than a couple of units which are just plain useless for rushing against defenses (the poor Soldier Shells, in particular), that your results vary widely, depending on what you use.

Personally, I like it when things die quickly, like StarCraft and the infantry from the old C&C, and I'm not a really big fan of "you can start firing at my tank, and it will take 30 seconds to die". Real modern combat is very fast and very violent. I like the way that a rush that's not built right to counter what's in place just disintegrates, myself. It's in keeping with the OTA feel of things, where once you had some serious ground-defenses, you'd see rushes get splattered quite quickly, if you didn't build the right stuff.

On the whole AI issue... I'll take a look at the LuaAI for KP at some point.

I still think that the entire reason AI development is moribund is largely because they've been navel-gazing, and not utilizing Lua to do it properly. Nine months later... no new AI development has occurred. AF, you talk about this great reorganization Hoijoi did, and I'm sure it's great.

However, since you then claim, and have been saying for months that AI development is essentially a dead letter (RAI now plays Overmind, btw, yay), why am I then supposed to reveal nitty-gritty details of my game design in public, if it will just be horribly misunderstood, involve arguments about things, and in the end not cause AI development to actually happen?

There's a chicken-and-egg thing here- you're asking me for chickens, but I don't see any eggs. Not real excited about stepping forward, considering that I don't hear anything solid... moreover, much of this is not set in stone, I'm having this little public discussion because game-balance changes are easy to implement, for the most part- it's the mechanics that are a pain in the arse. Since I have some fundamental mechanics built, the question is, do I use them, or not? I'm inclined to try them out, just to see what the larger reaction is, than to just let the idea sit collecting dust, frankly.

At any rate, it's my belief that we really need a general framework AI, built with Lua, that's easy to patch new behaviors into. I just don't think I have the time to tackle that one. A basic cheating AI, that doesn't even pretend to give people a fair fight, but is really stupid, sure.

[zwzsg]

So, you're saying you'd prefer a system where they built maybe a bit more slowly than they do now, but also died a lot more slowly?

No, he says he want stuff to build and move faster!

[Argh]

very slowly in comparison to the time it takes them to die in combat

Things die pretty fast

[zwzsg]

Make them build and move as fast.

[clericvash]

I don't want a stupidily fast game though.

[Argh]

I think he was joking. Oh, wait, he made KP, so probably not Hey, I already did that anyhow, and I'm not interested in re-creating NanoBlobs.

[1v0ry_k1ng]

I dont understand how you managed to misinterpret that twice :p

its pretty simple:
right now things take a LONG TIME to build, a LONG TIME to move to battle and then they die VERY FAST. the result is a very slow, very small scale battle in which the moment a unit is squeezed out its sent towards the front, becuase whatever units are at the front suffer a very high rate of attrition in comparison to how long it takes to build them.

in short:
faster buildtimes by maybe 2-3x and maybe similar global cost reduction so more units are on the field (as right now the mod is very much micro over macro)

faster unit movement

slower combat

also, you should seriously consider putting in the custom formation widget, because playing a mod without it once youve experienced the joy of user definable and non-retarded group movement is agonizing

and i was serious about the 2-3x buildtime speed increase, right now I dont like fast games (BA was too fast for me, I preffer XTA) but this is just too slow to play, and too small scale considering what the engine is capable of/
to put it in the scale, fast to slow:

***FASTEST***
nanoblobs
KP
BA
XTA
NOTA
PURE
***SLOWEST***

aim for somewhere in the middle imo
loving the city creator thing btw

[Pxtl]

The problem with that is it seriously changes the game balance. Look what happened when AA boosted vehicle speeds - the porciest, most artillery-oriented Spring mod turned into vehicle spam.

[1v0ry_k1ng]

I havnt played enough to comment on the balance, but I think making the game fun to play should come first; rebalancing it wouldnt take that long, and you would get alot more testers if the game was fun (I remember when the xta clans all got together and played a big game of pure; it lasted about 40 minutes and was spent mostly microing 1 or 2 units to pass the time until the next one was finally built) although at the end one team pooled all their units together and made an unstoppable charge... of about nine tanks. it was so small scale..

[Gota]

All this is happening because efforts are nto put into balance..just all sorts of random not thought through ideas...
I rememeber asking here if pure is gonna be spammy...make it spammy...he engine allows it right?we already have a warcraft 3..that's just my personal taste though.

Also it's not like the balance is polished...you can change it right now as much as you want..but there was no prethought as to how the game should be played...at least i didnt see any posted here..
Will the game be balanced to be spammy? individual unit micro intensive? Will it be fast(relativ eto some other mod)slow.
Will units die slow or fast.
What will the flow be like?plans?something?don't take it the wrong way but by just coming with some stats and sticking them in you wont get much...

[Argh]

I dont understand how you managed to misinterpret that twice :p

I didn't misinterpret you at all, lol, you just don't understand what I said.

Moreover, none of the games I've played with people were slow and involved few units. It really sounds like you guys just didn't know how to deal with the economy- anybody who's tested this with me knows that you can and do soon have huge armies blowing each other up, on all but the smallest maps.

My understanding of what you said was perfect, and the corollary results are exactly what I said.

You can't change build-times without a corresponding change in the speed at which things die, without serious impacts on performance. You can't have a flood of units on-screen, without paying the price.

That's not a good option.

That said, I'll make a mutator that slows down the pace of combat by half, and double the rate at which things get built, so that we can see what it feels like. If it's fun, fine, if not, no biggie, it's a quick change to make. I'm not changing unit speeds.

It probably won't have a good result, but you'll be able to try it out. I'll post it here in a bit.

[Pressure Line]

but by just coming with some stats and sticking them in you wont get much...

says the person who loves calculator balanced mods

and just because you fail to grasp the point doesnt mean it isnt there.

[Google_Frog]

You can spam nano-turrets and factories. Don't limit yourself to 1 factory because the Overmind factories had the best bp/cost in the game last time I played.

Resistance is hard to increase BP because cons are bigger and more expensive. Factories are much more expensive and they have the factory constructor which is really inconvenient because they aren't queueable and they're tedious to micro.

[Argh]

Resistance is hard to increase BP because cons are bigger and more expensive. Factories are much more expensive and they have the factory constructor which is really inconvenient because they aren't queueable and they're tedious to micro.

Which is all on purpose. Resistance is supposed to get better anti-spam, and generally should be able to hold against Overmind if rushed, but must push out quickly, because they're going to lose the expansion curve in the long run. Entirely purposeful part of the design, trust me.

[Google_Frog]

Resistance is hard to increase BP because cons are bigger and more expensive. Factories are much more expensive and they have the factory constructor which is really inconvenient because they aren't queueable and they're tedious to micro.

Which is all on purpose. Resistance is supposed to get better anti-spam, and generally should be able to hold against Overmind if rushed, but must push out quickly, because they're going to lose the expansion curve in the long run. Entirely purposeful part of the design, trust me.

So are you saying they should have less BP to make them harder to spam? Wouldn't low bp make it hard to make the larger units too.

[Argh]

Not exactly. Look, the thing with Resistance is that paying for a new Advance Base is expensive, and involves some micro. Go look at the costs and timing- its 30 seconds to build an Advance Base, plus however much time it costs to build the MegaConstructor. Plus it takes 500 metal.

Compared to the cost of putting up another LandFactory (80 metal), it's far more expensive and time-consuming. And that's on purpose.

With Resistance and Overmind's economies fixed up... things are getting closer to how it should be, basically.

In an ideal world, where the gameplay actually matches the design goals... Resistance simply cannot afford to spam factories, and will generally expand linearly up until the plateau point where they can finally afford to build their second Advance Base. Overmind can and should either build multiple LandFactories, or stack BuildTowers around one, multiplying it and eating their entire income all along their curve.

And that was part of my initial balance, although a few mistakes made along the way made it non-transparent to players. People whined about turret spam, for example, because Resistance basically had free resources during most of the game. With that fixed, turret costs will probably get dropped back down eventually.

Overmind should be "spammy". Resistance should have better units, and better anti-spam defenses. Resistance should be able to defend pretty well if it porcs, but will ultimately lose on resources if it doesn't hold enough board. Overmind can't defend quite as well, but should be able to make up for this with numbers that make it hard for Resistance to gain ground. Eventually, things should get settled by siege units, air warfare, or a lucky break on the ground, if the game's being played by decent players.

Overmind's turrets are mainly about turret-creep and gradual area-denial, whereas Resistance's are purely defensive, and aren't much good past holding chokepoints. You can take out practically any number of Pillboxes with Tanks or Artillery Shells, and Resistance can use Rocket Platforms or MBTs, for example.