Part 3: Spawns

Post Reply
Message
Author
User avatar
DarnYak
Site Admin
Site Admin
Posts: 2364
Joined: August 12th, 2006, 2:54 pm

Part 3: Spawns

#1 Post by DarnYak »

Basics
Spawn units are probably an afterthought for most mapmakers making AoS's - they need to look nice but they're primary purpose is to be fodder for heroes to kill. This is not an entirely inaccrate thought, as if they are too important then why have heroes at all. But on the other hand, without spawns the map isn't an AoS - its just a hero arena. When making EotA, I don't remember having a distinct goal of making spawn units critically important, but I did know making sure sure they were extremely well balanced was going to be one of the most important steps in making the map, in part so that there was a level playing field to deal with the far more complex hero balance.

It's really easy to make bad spawns that really mess up what otherwise could be a fine game; this most often occurs when the creators get fancy and choose to give spawns abilities. I remember plaing some map where Banshees were one of the spawn units, and they retained their default curse skill. So the entire game you would ALWAYS be cursed, becuase there were always banshees around and they would just recast it once it wore off. I think an allied unit had slow, so the other team got to spend the entire game moving slow which is even worse. Terrible, terrible decision. At the very least, always think carefully before a spawn ability will impact player heroes.

The design of spawns themselves is naturally fairly mundane. I remember we based thier initial stats off the spawns from ToB (it's probably safe to say that ToB was a starting template for EotA if anything was). The number of spawns was driven by wanting a feel of actual armies of units fighting it out (like how Exodus feels). The numbers were proving extremely problematic at first due to lanes getting clogged, but then either Shortling or GraveI suggested that they should be shrunk and I knew that was absolutely the right way to go - WC3 units always felt bizzarrely large to me.

The other major decision was to attempt to have units be asymetrical, instead of identical units with two different models. I wanted to try it, but was prepared to just go with mirrored if balance proved impossible. It took a lot of testing but turned out not to be as hard as I initially thought. The trickiest one was I couldnt figure out why creeps would always win on the bridges. After a while, it became obvious that the units were the same size, but the furblogs were stronger, so the front melee lines would easily overpower the elves in the limited space - once the Furblogs' pathing was enlarged, the rest of balance pretty much clicked into place. My biggest regret from doing this actualy was due to the endless amount of "bug reports" because of how many players just assume the units are identical, and so once they notice its 3 on 4, or that two units don't have exactly identical stats, came to the conclusion there is an inherient imbalance or bug.

Shortling and I did spend many hours over a week or two just watching spawns fight it out during their tuning. From this, we learnt how complicated spawn flow dynamics actualy can be, especially with a little randomness thrown in. In DotA, for instance, there are mirrored spawns balanced such that the spawns generally annihilate each other roughly in time for the new spawns to take thier place, and the front lines don't really moved (at least this was the case last time I checked - a long time ago). However, if the timing is a little faster, one side will randomly get the upperhand, that push is then guarunteed to overpower every new wave, it slowly builds up with each spawn wave untill it crashes against your towers - at which point, once killed the defending side will have that advantage (even if slight), and the process continues in reverse. In EotA, it even had the extra randomness due to the randomized spawning order (note: this was done away with), but that extra randomness generally just helped speed the same process along. We concluded that such a process was desirable, and made various terrain changes to work with it (such as moving bases further apart as noted in Part 1).

The spawn waves did have one pretty fatal flaw of quicly becoming overpowering, and the losing spawn wave was more or less useless when fighting a superior force. As a result, the casters were given Flamespray/Frostsomething so that they can deal a disproportinate amount of damage to larger hordes, and weaken then by the time they get to the towers. The computer AI, unfortunately, doesn't optimally use it, but if it did it would really screw up game balance anyway. This actually did cause a major problem around in the first version or two - the damage was also hitting heroes, and so the initial versions were plagued with many many level 1 hero deaths because they were around the front line. It even lead to a great number of "omg this map is too hard" type comments - in reality, the problem could be dealt with, but it was just stupidly annoying and thus heroes became immune.

Elite Spawns
The ability to increase your forces certainly didn't start with EotA, but I like a variety of options to achieve victory and this was certainly one fitting to add. I wanted to design them such that all were viable options, and that they were specifically designed to counter one another so that team ahead in towers could aggressively drop them, but the team that's behind had a slight advantage for being on the defensive. Rock-paper-sissors is usualy the model used, but I wanted a little more variety so instead it was a pentagram setup. Shortling and I spent many, many, many tedious hours making sure the spawn setup worked right. Would make a game, one of us would drop 5 spawn towers of a specific type (ie Knights) in every single lane, the other would drop thier tested counter (ie Catapults) and we see how that plays out. Then one would get its balanced tweaked because it was underperforming in a specific scenario, and we'd have to go back and test all of its matches again. Miserable times. We also had to make sure they were equally worthwhile to drop - for instance, all catapults might quickly take down bases, but archers would hardly touch the towers. So archer types got a secondary siege attack to help kill bases more like other units.

In the beginning, additional spawns cost gold to purchase, as was standard in maps. This turned out to be a pretty bad idea, as I came to realize purchasing extra spawns with gold is virtually always the best way to spend gold, even if the price seems abnormally high. Balancing it for price is also an undesirable plan, as nobody really likes dumping their personal gold into a not that exciting mechanic. And one uncooperative teammate can set the team back.

Fairy early on the gold cost was converted to the current tower core system (although i don't remember which version that was - I would think it went in with the Obelisk objective change but I thought it was changed earlier). Putting it on its own resource was pretty helpful in a few ways. First off, it was a team resource, so noone had to feel they were making a significant personal investment only for the sake of the team. Second, it tied in very well with making it a valuable reward for objectives because it was something players can actively use, benefited the whole team, and wasn't just something simple like extra gold. The only problem is that if one side is consistently unable to obtain them, they have limited alternatives to mitigate a deficiency and come back in the game.

The Good: Overall, I think the spawn system turned out pretty excellent. I've heard others say EotA feels more like a war, which is what I was going for. Their design makes them pretty important to gameplay and more then just fodder (although they are important as that as well!).

The Bad: Air spawns have been a consistent hassle to balance. Initially they took the wrong paths and so would just do all sorts of stupid shit (dropping air on inn for instance would bypass most ground spawns and go straight to the towers). Later it was just the gross advantage of having many units able to ignore collision. Further, they always attack whatever can attack them first, so heroes tended to be a prime target making them one of the most annoying spawns to fight. I'm not sure I ever conclusively fixed their balance.

The Annoying: When originally added, Polymorphed units had normal run speed, but that would just mean they would keep walking foward, often all the way into enemy towers to die in thier sheep form. So they were given 0 movement speed and made immobile. Then Blizzard, in thier wisdom, decided they can't have 0 movement speed anymore and instead have 0.01 or something ridiculous. The result of this change is that units that would walk around a stopped unit, no longer do, and instead wait for that unit to move out of thier way, which they don't. I'm hoping to scrap polymorph in the final version, but I'd need appropriate replacements.

User avatar
DarnYak
Site Admin
Site Admin
Posts: 2364
Joined: August 12th, 2006, 2:54 pm

Re: Part 3: Spawns

#2 Post by DarnYak »

Upgrades

As I've said before, I like options for achieving victory. Doing so by building a better army seems like a completley reasonable way to do so. The truth is, it's actually a miserable way to win in an AoS. The player controls a hero, the focus should be the hero, and all routes to victory should involve use of that hero (Note, however, that this is not say spawns are unimportant). Unfortunately, this is a realization that I took quite a while to come to, as obvious as it sounds.

Consider the ramifications of global upgrades. First, it gives you an advantage in every confrontation on the map - you will now start winning lanes by default (it can be countered or mitigated, but not without considerable effort). Second, that advantage takes littarlly no skill besides gold farming and clicking the right button, and once achieved there's not even the satisfaction of controlling the spawns or anything. Third, it mitigates gains by opposing heroes - your nuke can do 50% more damage? so what, my spawns have 50% more heath! Fourth, it also makes spawns better at killing bases on thier own. The only advantage I see is sort of the compliment of #3 - in the pressence of overpowering AoE, it might be the only viable way to counter it.

Initially, EotA had nothing but basic global upgrades for the cost of gold. When I refer to global upgrades, I refer to any generic combat stat increases on spawns, even if limited to a subset of spawns (ie ranged only). If I am remembering right, the cost was initially really high and nobody purchased them. They were lowered to see more use, and suddenly people realized the best thing to do was to spam them. The cost was increased to its original cost (if not higher) and still it was a free ticket to victory. Previously I said spawn towers for gold were virtually always the best investment in any map that offers them. Here is the primary exception to that rule - if spawn upgrades exist, they are most likely even better.

Unfortunately, I believed the problem was merely that gold was used, and created yet another resource to make upgrades more independant and equalized. I also added a bunch of individual upgrades, which i concluded had to be considerably strong to compete with the global upgrades. This at least lead to a somewhat balance between teams, assuming both teams would use their upgrade resource (otherwise there's a free win for which ever side used it). Still, not at all a desirable result. In addition, the resource was (and still is) gained passively due to outpost control - the only real influence players have on it is by upgrading to higher tiers of outposts, so it's doesn't even rise to the level of being a game objective to win. In theory, it would be more reasonable to have these if they were tied to player actions like spawn towers currently are, but I hesitate to add an additional objective when there's generally enough to be worried about at any given time. Note this was actually sort of tried with Candleburg, but poorly so.

Needless to say, upgrades were finally removed until the most recent version. Part of the reason was the above balance reasons, the other part being the tedious amount of effort required to add upgrades when the new races were added, which I didn't want to do when I didn't like how upgrades were currently operating. In the most recent version, upgrades were reintroduced with a different approach. First, all globals were removed, they added nearly nothing to gameplay (except for those who enjoy winning without any real skill, who I will not cater to). Second, everything else was toned down. The goal was to make upgrades so weak that they shouldn't win you the game on your own. Instead they are focused on being largely cosmetic - every ability is something players should be able to notice happening so that they feel something has changed. This both lets players feel they've done something by upgading, and alerts enemy teams that there has been an upgrade. Passive stat increases are impossible to notice unless you take the time to click on enemy units, and so were completley removed.

Picking upgrades was actually fairly challenging. First, I wanted to avoid triggering unit abilities at all costs so I stuck to variations on default abilities. Second, many of them were passive and just unacceptable - auras, for instance, a practicly just a global upgrade. Third, I had to avoid things that would violate my spawn abilities problems mentioned above, although most could simply not target heroes at all. Fourth, some units just don't go well with abilities. So this spawn throws axes for a living. What could he do? Well, he could..throw axes faster. Yea there's one. And he can.. throw...nets! Yea there's another. And thent he axes can...explode! No that's stupid. Bounce! Alright, got all three... This is why there's probably 8 variations of berserk. Even designing some spider that just spits webs has more potential then a lot of the spawn types.

Arguably, upgrades should have stayed gone considering how minor of a role they've been deligated to. But they were a staple for so long and I felt I'd basicly promised players to bring them back that I had to find some way to re-introduce them. Were I designing the map from scratch, I wouldn't include upgrades again unless they were direclty tied to map objectives.

Command Towers

These getting thier own section I expect will suprise most people. They were intended to be part of the whole map flow, base conquering/trading and using various paths of assault. They were also one of the earliest features added when the map was in development. In a way, I'm also shocked I haven't had to remove them for completely messing up gameplay - I believe they are the single most overpowered option in the map when properly used.

There was only one breif time (I believe in early testing) where Command Towers were extensively used (aside from some games where people would ignore heroes and try to win excluslively with superior spawn directions - fun times). The strategy went like this: At the start of the game, drop command towers on every lane and send it to middle. Let that horde overrun the enemy base then continue onto the other lanes. This lead to the current level 20 requirement so that heroes should have some ability to mitigate giant pushes as well as buy time to deal with countering it. Since then I'm unaware of much use aside from abusing flying unit pathing.

However, there's quite a bit that can be done. Take defensive rallying for instance - enemy spawns generally shouldn't be able to build up a wave strong enough to break through a built up group. Sure, you're conceeding you won't be pushing in that lane, but if you've lost the front base this is an excellent way to protect the back base while you try to come back. How about blockades? For instance, on Stormwail as red team, have the back base for bottom lane rallied to the road behind blue's foward inn lane base, so that the front base is now at a disadvantage. Forced waves - just rally somewhere near the command tower, wait for spawns to build up, then order them to retreat and get new orders to attack. Different attack angles - avoid bridges/cliffs/ramps by going around them completely. There really is an incredible amount of potential to use (abuse?) these things.

I would keep these around were I to remake the map, but I would want to improve them quite a bit. First, integrating them to be built into all town halls - if they are suppose to be a core game mechanic, they need to be highly accessable. It also gives the other team an equal ability to react with their own command towers. I would, however, want some sort of limiting factor on them - something like mana so that it can leave you exposed if you overcommit or otherwise make bad decisions (although I hesitate at the idea of introducing yet another resource into EotA).

User avatar
Luftwaffles
Communist
Posts: 398
Joined: May 22nd, 2009, 11:48 pm
Realm: Azeroth (U.S. East)
Battle.net name: Ion_Killer
WereWabbit
Canucklehead
Luftwaffles
Location: Canada

Re: Part 3: Spawns

#3 Post by Luftwaffles »

Very interesting. In my youthful arrogance I'd always wondered why you insisted certain counters worked that were often in dispute (Flying vs. Cats iirc, but there were others). It's clear you tested this pretty thoroughly and thought things over- half-assed at first but it came together unbelievably in the end. Flying and polymorph will always be problems, but more with Blizzard and JASS imo.

I'm liking this dissertation on EotA, very satisfying. Keep it up!

User avatar
Discombobulator
Retired
Retired
Posts: 710
Joined: September 19th, 2006, 4:16 pm
Battle.net name: Karunecm
Contact:

Re: Part 3: Spawns

#4 Post by Discombobulator »

Why not just change polymorph to a stun?

Also, thanks for these posts. They're a very good read.

User avatar
DarnYak
Site Admin
Site Admin
Posts: 2364
Joined: August 12th, 2006, 2:54 pm

Re: Part 3: Spawns

#5 Post by DarnYak »

Stun's one thing I'm considering, but it doesn't seem quite right (although i can't explain why). Entanging roots is another option...sleep sort of works but not quite.

DarnYak

User avatar
Luftwaffles
Communist
Posts: 398
Joined: May 22nd, 2009, 11:48 pm
Realm: Azeroth (U.S. East)
Battle.net name: Ion_Killer
WereWabbit
Canucklehead
Luftwaffles
Location: Canada

Re: Part 3: Spawns

#6 Post by Luftwaffles »

I suppose a really complex way would be to give each of them one of those transform abilities (been awhile since I played but iirc something like Metamorphisis) and have it so when they get hit with a debuff or are targeted by the spell it activates it and they turn into sheep versions of themselves with all the same stats and then tell them to Hold Position (and de-activate it with a timed trigger).

But I can see why you just said "fuck it" because that's the antithesis of Occam's Razor.

User avatar
DarnYak
Site Admin
Site Admin
Posts: 2364
Joined: August 12th, 2006, 2:54 pm

Re: Part 3: Spawns

#7 Post by DarnYak »

Finished writing part 3, see the second formerly reserved post.
I suppose a really complex way would be to give each of them one of those transform abilities (been awhile since I played but iirc something like Metamorphisis) and have it so when they get hit with a debuff or are targeted by the spell it activates it and they turn into sheep versions of themselves with all the same stats and then tell them to Hold Position (and de-activate it with a timed trigger).
Yea, I've considered just forcing units to hold position if polymorphed, but that just feels like an overly complicated solution instead of just using a different spell.

DarnYak

User avatar
Luftwaffles
Communist
Posts: 398
Joined: May 22nd, 2009, 11:48 pm
Realm: Azeroth (U.S. East)
Battle.net name: Ion_Killer
WereWabbit
Canucklehead
Luftwaffles
Location: Canada

Re: Part 3: Spawns

#8 Post by Luftwaffles »

Entangle seems like the best solution, even if it's not the most palatable one. Do a magical prism for elves, frost for creeps and that one graphic with the hands coming up for undead.

I have fond memories of Polymorph, Tox raging at merc pushes trapped behind mass Polymorphed sheep. Though perhaps, too many.. with too many different people.

Post Reply