Warlords and Merchants Discussion

[Xenon]

It's an AOS type map, so the unit waves can snowball into a mob and unload a ton of shots at a time. Some player weapons could be doubled etc. and some of them are supposed to fire several shots at once or split into smaller shots.

There's a tot of talk about that 1/32 seconds but I've seen no convincing evidence. It just seems to execute twice and then wait 1/16 sec, or maybe do the second execute at a different part of the game loop (after pathing or so forth) but I doubt it.

The graphical movement can be made to blend so that's a non-issue. The problem is that it has to move each projectile multiple times per game loop to check for collisions between them. Spamming search areas probably is the main CPU hog regardless of the actual time interval, but I don't know if it's possible to make a math-based vector system that's faster.

[Ricky_Honejasi]

From what I remember, the 1/32 seconds can be actually achieved with a while loop and using a wait of 1/32 seconds inside of it. I didn't test or anything but it can be worth trying.

However, I don't think it's worthwhile to focus so much on having dodgable/collision shots and such. Sure it looks better achievement-wise but I never seen it as a real necessity while bogging down the performance.

If you still insist, maybe one thing you might randomly try is to always fetch the X,Y of every REAL unit on the map then compare it with the fetched X,Y of any shot instead of using the Search Area. Another aspect might be to pre-split the map into 40+ logical areas so you minimize greatly your comparisons (only compare Ships of Area #23 with Shots of Area of #23, etc.).

As for me making a battleship map, I didn't plan much in like forever concerning that. Plus many of my side plans (if I ever poke a new map one day) would involve considerable bank file use for more unique maps.

Unfortunately, still limited time and still care about Disgaea 4 for now. A new real map always take considerable time to get any real first version too.

EDIT : Maybe the big exception would be if you guys truly want to play something different to the point that you don't care for W&M updates for quite a while.

[IskatuMesk]

I think that if W&M was to considerably progress past this point you'd want to hand it off for major balancing overhauls. It does not seem like much else can be achieved without a lot of time and, yes, I am starting to grow very tired of it.

[Xenon]

That's the 1/32 thing I've heard of and tried a long time ago. Like I said, executes twice then waits 1/16 from what I could tell.

However, I don't think it's worthwhile to focus so much on having dodgable/collision shots and such.

What the- that was always a major part of the map concept, even from Warcraft 3! It was inspired by the fancy triggered shots in Pheonix-IV's take on the Waht map.

Also, the ships are supposed to have "destructible" sections with armor and weapons protecting a core like Battleships Forever. For that to make any sense it needs projectile collision. The way the different types of shots collide with each other was even intended to be a counter system.

I've considered a type of castle-building map with collision shots, where unit numbers can be reasonably limited and shots don't need to hit each other. Not sure how to do some of the specifics of the gameplay or shooting through fortifications, though.

Mesk> I don't think Ricky will like it that you brought that up again.

[IskatuMesk]

Like it or not, it's best to consider where you want to devote your resources with the circumstances you are in. If you don't want to hand it off, that's fine. If you want to make another map but are restricted by time, I think I would rather see him try something new at this point.

Anyways, at this point I don't even really want that responsibility personally. I'm just losing interest in this and would rather have something else worth playing for a change seeing as all the generic maps are awful.

[Ricky_Honejasi]

*big sigh* No, I still intend to keep the W&M map in my hands.

Anyway, one of my plans if I poke a new map would be the following :

Essentially, it's some kind of defense map where the main focus is to defend against customized units by you and other players in the game.

In solo or at a game's end (instead of a direct defeat), you can customize offensive enemy units and defense custom units that will be kept into your bank files and used in your next game.

Thus, in a way you have to prepare for the unexpected as lots of potential deadly cheese can come out of your friend's evil creations!

For the customized defense units, it would be more additional units on top of some melee race. However, they would be pivotal since it would be a big ticket to defeat a lot of cheese. If you lack any customized units, all unmodified "slots" will be given the default pre-customized unit for that slot.

One of the potential plans would be to allow anywhere from ultra-cheap units (easy to get) to "unlimited" units (can take forever to get Mesk-like powerful units). Unmodifiable supply restrictions does force a slot-like system even if I can modify mineral/gas/time by triggers.

In the short term (probably first real versions) :

- Able to customize generic stats of enemy and defense units. Probably on a very limited set.
- Have some of level scheme in customized enemy units so you don't face off super foes on round 1.

Later on (maybe, depending on effort required):

- Having more sets
- Weapon modification (stats mostly).
- Being able to name your unit. Possibly having some sort of unit description as well.
- Able to pick behaviors to put on enemy and defense units.
- Able to pick model/wireframe/etc. of enemy or defense units.
- Able to pick the "row" of the unit. (Ex : If melee, you want to pick "Front row" while support would be "Back row")

Premium ideas :

- Vast expansion of addable custom behaviors to allow a lot of cool stuff.
- Ability to set some kind of general mini-AI (mostly trigger-driven)
(Ex : "Dumb Charge", "Split into 2 and attack sideways", "Annoy mode")
- Ability to have some set of logic triggers for the unit
(Ex : "[On damage] , [Blink] [Randomly]" or "[Reaches player base], [Stim]" or "[See unit] [in range of [10] - [Yamato]")
- Maybe managing to fake ability use for enemy units via equipped items.


(including more advanced stuff such as life/shield/energy cost for weapons or being able to life/shield/energy steal)

[IskatuMesk]

I kind of like the sound of that, actually. Depending how easy/flexible it is to make units. But I can see why it would take a lot of time to enter the stage of usability. Heavily mechanic-driven things tend to be like that.

Doesn't relying on banks open you up for a lot of cheating, though?

[Ricky_Honejasi]

Doesn't relying on banks open you up for a lot of cheating, though?

Depends. Blizzard did add that validate thingy for banks. At worst, the bank files would keep all the stats and I can just force my map to recalculate the level based on the stats and such.

It's certain I won't keep data such as unit level, mineral, gas and time costs into bank files (thus calculated instead) since that can allow cheating for anyone devoted enough to fake the bank files.

I would say the bigger problem is having a good limit of how much a player can upload bank files size-wise without getting timeouts at game start. In addition, it implies trying have some algorithm to compress kept data to allow the most units to be kept in player's bank files.

[IskatuMesk]

A few gigs should be sufficient I think.

/e

Meanwhile, I hear Jim_Raynor is getting really good.



Laconius put up a valiant fight, the texts report, but it wasn't enough.

[Xenon]

Okay, here's that unit. Radar is hardcoded to ignore cloaked units (ugh) so I had it briefly decloak units periodically in addition to detection.

Should be good to sweep the map in the endgame, but not to spread over the map in midgame. They'll fall apart if a Hellion ball looks at them funny. It should be easy to obtain them, combo or elite tier 1-2 or so.

If you want to change the radar range, remember to also change the detect range, radar range actor radius and pulse search area effect, as well as the tooltip for radar.


As for maps... I don't know. I like W&M and being able to make units for it, but some variety would be nice too. Maybe I'll keep poking around to see if projectiles can be made more efficient...

[Ricky_Honejasi]

Okay, here's that unit.

Alright, ill look at it sooner or later.

Although back at the Bank stuff, I managed to do some tests although I clearly need 8+ players for any final tests. Preferably 12 players. So I intend to drag everyone at FFS at some peak time into the test for like 5000 integers each.

5000 integers/player with 12 players is probably the minimum to ensure the map is worthwhile. I really want to ensure about 10 custom fully-decked enemy units/player and 6 custom fully-decked defense units/player.

There is the limitations of bank but there are a few extra possible things fortunately :

- Have some secondary map where you are forced to be solo but able to use your main bank and 1-2 "mule" banks that can hold like 10-20x more without desync. The idea is if in a main game everyone is stuck to relatively low limits (ex : 5000-8000 integers) then in a solo game, you can potentially use like 100000 integers global for many more units. You could swap in and out active customized units for the real map.

- Internalize at least some bank units into the map. For example, if we all love X unit of a specific player then I could get his bank file, put the data into the map and then clean up the spot in his bank file.

EDIT : Since I aim to try to put lots of integers into long of strings. I might have come up with an extra way to compress said numbers. A number is at base 10 (decimal) while base 16 (hexadecimal) exists.

However, how about I start trying to use the most different types of characters possible to have like a base 100 to 300 something? That can be enough to say that extra space or two (ex : "53242" into "@^*" ).

Obviously, I would have to test how far I can reliably push such idea.

[IskatuMesk]

Out of curiousity, why do you need such an insane number of integers? Can't you organize the data like this...

Let's say we have a Zergling with Charge, Blink, Lifesteal. We have a section of integers that the map refers to a table to build from... with X representing the Zergling's integers used to build that info.

Ability X - Charge
Ability X - Blink
Ability X - Lifesteal
XX - Lifesteal %
XXXX - Health
XXXX - Shields
XX - Armor
X - Model Entry
X - Actor entry
XX - Movespeed

I mean there's probably a lot of stuff I am missing but 5k seems to be an extremely large number for just 10 units. I'm just a dumbfuck when it comes to programming stuff so excuse me if I am being dumb as fuck.

I'm also assuming you need an integer for each value of a unit's life but again that is a baseless assumption. If you only need one integer for the entire health pool then that makes less sense to me why you need so many.

[Ricky_Honejasi]

Okay if I start making a big list :

HP, HP regen, Life Armor, SP, SP regen , Shield Armor, MP, MP regen, Move speed, base unit type, actor, model, maybe abilities, "life" steal (life/shield/energy), "life" cost (life/shield/energy)

Allowing up to 4 weapons or so, damage per weapon, range per weapon, attack speed per weapon, possibly of various effects, etc.

Probably 10+ behaviors.

Unit name, if I allow this for like 30 characters, it could be the equivalent of 15-25 integers by themselves

Unit description (although choppable), if I allow this for like 500 characters or so.

Multiply all of this by like 20 (12 enemy units minimum + 8 defense units). Preferably want to allow 25 enemy and 16 defense units.

Also factoring everything could be worsen by a factor of 3 if banks just plain hates me.

Note : You also have to consider that bigger integers takes more space than smaller integers (ex : 500000 takes more than 20) while my "integers" test implies around 3-4 digits of average from my tests.

I prefer to have very pessimist counts considering it could be very annoying due to timeout issues. Also due that it can be very easy to use limited space very rapidly despite the impressions to have enough.

[IskatuMesk]

It sounds like you may have to settle for something much simpler and instead use those bank entries to reference to prebuilt abilities and such of varying scales/"costs" if this doesn't work for you.

[Ricky_Honejasi]

We will see. But I tend to prefer to prepare for worst case scenarios honestly. Especially if I expand the map's features over time (which will require more bank space per unit).