How to balance Element TD [4.0pb]

[Cisz]

How to balance Element TD

A short introduction to what we do here.

(Status for version 4.0pb)

This is about pick. Samerandom is allways balanced, and allrandom never will be.

The first step to ballance is to make all towers of the same level equally strong. So zealot, quaker, tidal, laser, and impulse (all are damage towers, triples, level 1) should have the same damage/gold/second, when used correctly. In fact, impulse and laser should do a bit more dmg than the others when microed, and less then the average when left to do as they feel like. We don't even have all towers working yet, so we cannot start to balance them against each other for now.

But there is another balance problem: Balancing the builds.

Back in the days of eletd 3.0, there were some good builds (like ldwf or lwdn) and bad builds (like dfne), so when playing pick competitively, dfne was suicide.

What we are trying with the 4.0 is to get all builds decent. So whatever elements you choose, there should allways be a way to win. This does of course not mean, that choosing earth 3 times in a row is valid, it means: No matter if you play 4, 5, or 6 elements, and no matter which sorts of elements you choose, it should be possible to win. But if you go eee$$$ you are as doomed as allways.

In our current plans we have 3 strong duals and 8 strong triples, and a strong dual is supposed to be more than twice as valuable as the strong triples. These strong towers are all support towers, and they are spread out over the builds, so each build has it's share. All other towers have to be damage towers only. No slow, no armour reduction or boosting other towers, only the 3+8 can do that.

The planned result looks like this:

[pre:2kypxpeh]Strong dual - 2 points

Strong tripple - 1 point

Pts. Build

4 ldwf

4 ldwn

4 ldwe

4 ldfn

4 ldfe

4 ldne

4 lwfn

4 lwfe

4 lwne

4 lfne

4 dwfn

4 dwfe

4 dwne

4 dfne

4 wfne[/pre:2kypxpeh]

With "points" beeing a value to determine the relative strength of the towers. (Calculated on a simple calc sheet.)

[pre:2kypxpeh]Or a bit more complex (and different points):

Category Maximum points

Overall 26

Four-element builds 26

Five-element builds 26

Six-element builds 26

Worst build maximum 24 (The 4-ele-builds with 2 strong duals)[/pre:2kypxpeh]

We did this calculation with this tool.

So far (version 4.0pb), we failed. In the pb, taking all 6 elements is much stronger than getting only 4. The power of the builds in 4.0pb (as calculated on a very complex calc sheet):

4 Element builds - ~30 dmg/gold/sec

5 Element builds - ~50 dmg/gold/sec

6 Element builds - ~130 dmg/gold/sec

And since 130 is more than 30, we need to work on that.

So the plan is:

1.) Get all towers into a form that works and can be ballanced. Make sure that only the 3+8 are support, and everything else is damage.

2.) Get all damage towers equally strong. This should take into account that some require micro (they need to do more damage to be worth it) or give life or gold (like life tower or gold tower; they need to be weaker).

3.) Get the right support strengths to get all builds equal. This requires to give maxed duals (which are hard to get with many elements) enough relative power to balance out much low level support (as in a 6 ele build).

[pre:2kypxpeh]Here is a current list of all 2- and 3-element towers (4.0 pb):

ldwfne

# 1 ++ LD Trickery Mimics towers

2 + + LW Ice Single target slow and dot (like poison in 3.0)

3 + + LF Electricity Bouncing spell dmg (a bit like metal in 3.0)

4 + + LN Life Gains life with kills (but slower life gain than in 3.0)

5 + + LE Quantum Can teleport itself, lands with a bang

6 ++ DW Poison Dmg in a line (like sun in 3.0)

7 + + DF Magic Extra dmg with mana based spell.

8 + + DN Disease Summons evil flies (aoe dmg)

9 + + DE Gunpowder Long range splash tower

10 ++ WF Kindle Aoe dot (a bit like steam in 3.0)

#11 + + WN Well Speeds up other towers

12 + + WE Hydro Sometimes throws a unit in the air, dealing splash when they land

13 ++ FN Flame Single target slow and increases dmg taken

#14 + + FE Blacksmith Gives your towers a boost on dmg

15 ++ NE Mushroom Can cast sporecloud on itself for some extra speed

ldwfne

1 +++ LDW Hail Can go multitarget from time to time

2 ++ + LDF Jinx Warps creeps back on the path and damages them

3 ++ + LDN Oblivion Summons skeletons

4 ++ + LDE Laser Does more damage on closer targets

# 5 + ++ LWF Windstorm Creates moving aoe slowing and dot hurricane

6 + + + LWN Tidal Charges an aoe attack for every 10 attacks, can be discharged

# 7 + + + LWE Polar Temporarily decreases targets health

# 8 + ++ LFN Nova Dmg and slow to all creeps in range

9 + + + LFE Gold Gives bonus gold

#10 + ++ LNE Enchantment Decreases creeps armour

#11 +++ DWF Corrosion Dot and armour reduction debuff, aplied with a splash

12 ++ + DWN Drowning Chance for instant kill based on targets health

#13 ++ + DWE Muck Splash attack and aoe slow with that (much like ice in 3.0)

#14 + ++ DFN Voodoo Units get extra dmg based on dmg taken since cursed

15 + + + DFE Flamethrower Rapid fire splash tower with incinerate

#16 + ++ DNE Roots Slows all creeps in a line with dot

17 +++ WFN Impulse Does more damage the further away the target is

18 ++ + WFE Zeal Gets faster and faster the longer it fires w/o pause

19 + ++ WNE Flesh Golem Grows and grows while dealing dmg

20 +++ FNE Quaker Has a chance to do splash dmg

# - One of the stronger support towers[/pre:2kypxpeh]

[DDRKirby(ISQ)]

nice analysis, i hope this this a goal that can be realized...

[jolin012]

someday we'll reach an AP that can be played with diferent builds challanging eahch other!

SR will always be balanced and AR will never be balanced, that part we've accepted for now

[holepercent]

that would be the ideal.. fair competition for well played 4-5-6 element builds.. although it may be say harder/easier to master certain builds compared to others. but once mastered, they are (almost) equal to each other..

[Cisz]

Yes, I foresee builds that are easier for a beginner, and others that seem unlikely or even impossible, but do amazing by the hands of an experienced player.

[jolin012]

as a player, I'm getting all excited allready but the developement part of us has some to do first

[m0rph1ing]

Balancing the builds seems really hard to me. A typical 4 element build will be 3322 with the last lumber used in interest, essence, or another tier 3. A typical 5 element build will be either 33221 or 32222. The first 5 element build gets 1 strong lvl 1 dual and 2 strong lvl 2 triples more than the 4 element build at a cost of that last lumber. The second build gets 1 strong lvl 2 dual and 2 strong lvl 2 triples at the cost of a lvl 3 dual and the last lumber.

In order for these to be balanced, the usefulness of the last lumber and the lvl 3 dual has to be really high.

For 6 element builds compared to 5 element 32222 builds, the difference would be between pure/refined towers vs periodics + lvl 1 dual + 4 lvl 1 triples.

1 possibility to increase the viability of a 4 element build is to push lvl 3 duals so that a 3332 build would gain the strength of 2 additional lvl 3 duals as damage towers. Right now, the cost effectiveness of lvl 2 triples is so much better than lvl 3 duals that no one goes for lvl 3 duals for damage. The only lvl 3 duals used are the support ones, namely well and blacksmith.

Just off the top of my head, would a lvl 4 dual requiring essence help for balance? This would push the need for having 2 elements at tier 3, and by doing so, help 4 element builds. The main problem with this is that pures might be made obsolete in terms of use for essence.

I think most people agree that periodics need to be nerfed; so the 6 element build will not be as strong.

My main point is if Cisz says that strong duals are supposed to be worth twice as much as a strong triple, they really need to be buffed.

[DDRKirby(ISQ)]

lvl 4 duals requiring essence?

what about lvl 4 duals that dont require essence...

what about lvl 3 triples??

=pppp

[m0rph1ing]

Oh ya, forgot about those possibilities. With 11 picks tho, only 1 lvl 3 triple could be used in a game. Maybe 1 extra pick could work. It seems that balancing among 3333, 33222, and 222222 builds would be easier, although this would mean interest and essence would not be used in any of these builds. 13 total picks would be more interesting. It would make the game longer though, and it's already long enough as is. =/

[DDRKirby(ISQ)]

no more picks! no witchcraft + ice age!!! =p

[jolin012]

Hey, this thing is possible to balance without that, yes it sonds hard and that's why we're working hard on it, and also why we're probably going to buff the damage/sec/gold of tier 3 damage duals and the boost that tier 3 support duals give. alone with some other changes.

such a big change as adding more tiers to towers would only maybe ake it easier to balance, maybe harder. (like with tier 3 triples, WNFE WNDL and EFDL would lose all their power i guess without any closer thoughts on it.)

And, for 4.0 we're also going to stick to 11 picks. it's no good to be able to max all triples like kirby sais, or to max all eles in a 4ele build. And we would need to change the creep hp for eery level and all gold gains every level aswell, if say, the first lumber was given at start of game.

Ok, just wanted to be clear but tips and ideas are always welcome

[Karawasa]

Adding more content is not the way to balance current content. More picks or more tiers simply results in more to balance, and we have our hands full already. As jolin012 hinted at, the data suggests that we will be able to tackle this problem within the current system.

[Cisz]

More content without balance problems could come from cosmetic changes, like gfx or sound, and it could come in the form of new game modes.

[Guest Timberwolf]

I haven't read over the whole thread, but yeah there are builds in 4.0pb that seem like they *should* be good, but I don't think really are. 4 element builds should give a good player an excellent shot at winning the main game.

Prime example for me is N/D/E/F. You have a major slowing tower (thorns nature damage), and three other "damage" towers that cover three different elements (Flamethrower fire, Magic dark, Quaker earth). Also access to a damage boost with Blacksmith and an additional decent effect tower with Flame. In theory this should be a half decent build but in practice I've always had massive problems getting much out of it.

I don't think it's an issue of reworking the basics of the game, just tweaking what's there. The ideas are solid but maybe the execution is just slightly off.

[holepercent]

DFNE did take a few more attempts (for my build project) but it's still a pretty reasonable build to complete at least for me. Roots, quaker, voodoo, smith flame (mushroom) and maybe gunpowder for cover.

The main balance issue is the strength of the various 15 4-element builds. DNEW and LFNW are the easiest 2 with 2 aoe slows each. LDNE is close to unfinishable. The rest are in between with varying difficulties. DNEL and LFNE to name a few took me close to the full number of attempts. WFNE has both well and smith, appears to have decent damage but is one of the most challenging to complete.

Key balance is also that all towers out of the 3 superduals and 8 supertriples, ie all damage towers are buffable properly by well smith (trickery). Also the reworking of a few towers such as disease that are just not viable.

[electronX]

It must be pretty hard to measure build strength. For instance, take kindle. It can even do 1800 dps (position 5) as well as 4-500 (position 5 / 3).

Or, what about AoE slows? By slowing the creeps, there is more time to compete.

Anyway, I agree in one thing: it may be possible to construct a solution where each build is capable of clearing the creeps at the same point (say, exit) exactly. Say, slowing builds have less dps but have somewhat more time.

BUT.

What if you mix a slowing build with a DPS one? What if you mix slowing builds with buffing ones? (Blacksmith + Hurricane) What if you put everything together?

As long as effects stack you *must* be better. Not because of the individual towers but because of the buffs / debuffs and the ability to have many answers. Say, kindle to start, good dps to execute and some strong ranged one to finish off the evaders.

For buff towers, this can be limited: since they buff 4 towers, with 0 damage they are always balanced. However, they can not only buff duals, but also triples, everything until pures. The bonus is pretty different, while the cost is still 500.

The same goes for debuffs.

So, when having a completed 2222, or maybe earlier, it seems better to finish it to 222211. Or, as somebody said, 3322 must have the equal strength - at the same price! Increase all towers' damage by 25% for 500 gold and such.

And, yes, buy 25% of a pure for 500 gold = buy a pure for 2000 gold (it costs 20.000 atm).

I wouldn't accept this project so far.

One more thing: *all* good RTS players know that capturing enemy technology is better than just investing our own. WarCraft lacks this feature, except the undead: imagine cheap orc warriors with night elf hunters' support, healed by high elf priests. Blizzard decided to *not* allow this - I know of the undead, but worker stealing is rarely achieved.

[Cisz]

And, yes, buy 25% of a pure for 500 gold = buy a pure for 2000 gold (it costs 20.000 atm).

That's not how it works. Consider it like this:

If I have a level 1 support dual in my build, I get +25% on everything. If I have 3 lv.1 suppurt duals, I have +75%. They are so cheap, the cost is basically 0.

Now we have to set vaues for:

[pre:djxsda2q]% boost

lv1 lv2 lv3

Support dual ? ? ?

Support triple ? ? -[/pre:djxsda2q]

..in a way that all possible builds are equal. This problem allready has been solved. The theoretical values exist, and there are a lot of sollutions to this problem. Now we need to choose one of them and to properly aply this numbers to the map. This is ofc assuming that x% aoe slow = y% boost.

[echinodermata]

If I use both lv.1 well and lv.1 blacksmith on one tower, that tower gets a 56.25% damage boost right? Why do I only get 50% boost with a lv.2 blacksmith?

[Cisz]

Level two blacksmith is one of the towers that are harder to get if you are playing a 5 or 6 ele build, so their boost is set to represent a bonus to level 4 (and to a lesser extend 5) ele builds. There is no connection between 2 lv1 and 1 lv2 boost duals. We might end up with lv1 doing +12% and lv2 doing +60% or +20% and +35% for a proper ballance.

Remember that if lv1 duals are strong, 6 ele builds get strong. Too strong.

[Guest Silfarion]

Level two blacksmith is one of the towers that are harder to get if you are playing a 5 or 6 ele build, so their boost is set to represent a bonus to level 4 (and to a lesser extend 5) ele builds. There is no connection between 2 lv1 and 1 lv2 boost duals. We might end up with lv1 doing +12% and lv2 doing +60% or +20% and +35% for a proper ballance.

Remember that if lv1 duals are strong, 6 ele builds get strong. Too strong.

that does make sense. if the percentage gets higher, later on the towers (especially pure) will get ultra bonus damage. but even though that is the target of balance, it is a bit strange that the tower have weaker power. i think your last idea (from 12% to 60%) is good, but i think from 12 to 60 is a bit over