Upgrades Analysis

[Borsuc]

Yes but the ratio is the interesting thing. +1 initiative isn't precise at all. If a creature has +1 init from 5, while another +1 from 30, then the one who benefits more is the one with 5 (20% increase instead of 3.3% increase)

[Asheera]

Also you might say that you need a lot of compensation for less initiative, which is actually the problem with the Minos and with the Sisters, albeit with the Sisters there IS compensation - she's more massive and therefore slower.

Well my formula favors initiative more than a simple Initiative/10 damage.

What I'm saying is that 11 Initiative can't be compared when making balance with only 10% extra damage. Since Morale is apparently "weaker" than Luck (provides a smaller bonus - but to something which, supposedly is better) then 10% extra initiative is better than 10% extra damage (ON AVERAGE, I know there are a lot of cases but when making balance or comparisons like these you have to take out an average situation, you simply can't balance for all situations that may occur)

So, I found out that extra Initiative has an effect of 33% great than extra damage. (I found this 33% number by dividing 2 (the Luck factor) by 1.5 (the Morale factor) -> result is 1.33)

This means that an initiative of 11 is actually (on average) like having 13% extra damage, NOT 10% (I used my formula which is explained in the first post)

Likewise a unit with 16 Initiative is like having 80% extra damage than one with 10 Initiative, NOT 60%!

Remember, all of those are on average situations, you can't balance for all possible scenarios.

[Jolly Joker]

Yes, you did that, but it's only half the conclusion, since Initiative affects not only the damage you do it affects the damage that can be done to you as well. You have to see it from the slower unit's point of view as well:

If you have initiative 7, you will be hit twice by a unit with initiative 15, no matter the damage you do. Which means the following (which we know from the game in practise quite well). If you have an initiative of 7 and you do 100 points of damage, you'll still die and do no damage at all if you are killed before you can act. If you have a better initiave, let's say 10, do only 20 damage, but survive because you are hit only once you will DO damage as well.

So initiative is not only adding to the damage you do, it's adding to your HPs or survivability as well. This is reflected in square-rooting the product of HPs and damage and multiplying it with simple initiative.

Which is the reason why with my method the Furies are not worse then the Sisters and are actually a tad better, while with your method Sisters are the better pick f I read that right.

[Asheera]

Yes they are a little better (by around 7%)

Also, in my formula I already weaken the slow units by applying the same "bonus" to initiative, only that it's below 1 here (since initiative is lower than 10) - so this "bonus" is actually a bonus to a weakness, thus making the unit with 7 initiative worse than only 42% (10/7) - it makes it 57% weaker overall.

[Jolly Joker]

Well, in fact we have a different method. You are looking at the changes or differences only, while I try to see the whole picture.

My main point here is that the abilities are extremely different to quantify and to get into a relation to the difference in stats. That's why you must have some idea about how to rate different abilities otherwise a comparison makes not much sense.

[Wolfsburg]

Now that we know the "raw power quoeficient" of each unit, you guys should be able to calculate hero raw troop power as well. For example taking Havez beginning troops, multiply the number of gremlins by their X quoeficient and calculate Havez raw troop power quoeficient. And then for instance comparing with Ossir's beginning raw troop power. I bet Havez owns anyone else in the game.

Plus, I am very sure you can calculate ballista's X quoeficient as well, which is an undeniable reinforcement thus further increasing Havez's edge.

This would be a pragmatic way to point out which starting troops are badly distributed.

[Asheera]

For a start I'll say that the unit specialists start with a lot more troops than other heroes - this is already an imba start, isn't it?

[Wolfsburg]

Ash... if you calculate as fast as you post, this threads first post wasn't that much of a trouble afterall

And indeed, unit specialists supposely have an imba start.

But it would be interesting to see that in numbers, in percentage.

[Asheera]

Ash... if you calculate as fast as you post, this threads first post wasn't that much of a trouble afterall

Ah no lol, I posted immediately because I knew this thing with starting armies from a lot of time ago

The point is that they start with three stacks instead of two, like the others. Sure, the others can start rarely with 3 stacks as well, but that third stack has a very low power even if it is present (not to mention most of the times it's not)

[danhvo]

If I want a rough estimate of a creature's strength, which is what these calculations are about, I just use the power rating listed with each creature. According to this rating, the Peasant is lowest at 41, and the Phoenix is highest at 8576. I'm not sure how these ratings are calculated, and I don't always agree with them or pick the upgrade with the higher rating, but I find them mostly very good.

The rating, for example, lists the Blood Furies/Sisters at 484/477, and the Minotaur Guards/Taskmasters at 474/488, which is in line with my original impression that the Furies/Sisters seem more like Tier 3 creatures than Tier 2.

[Asheera]

The power rating is extremely flawed - Nival couldn't balance it properly it seems. I'm telling you, high tier creatures are "favored" by it - meaning that they have less power than their actual effectiveness.

Do some tests yourself. A tier 7 is much stronger than X tier 1s, where X is equal to Power_of_Tier7_You_Test / Power_of_Tier1_You_Test.

That is because the tier 7s have the power too low compared to their actual strength.

BTW: This is also a useful tip for Necromancy - this means that high tier creatures are better at raising (cost less DE than their power versus the low tier ones)

[danhvo]

I disagree. Who says that the power rating should be treated linearly, and that if creature A has twice the power rating as creature B, it's supposed to be equivalent to 2 of creature B? Use whatever scale you find fits your own experience. Squared would be way too sharp a scale, and linear would be a little too flat; I think the real scale is somewhere in between.

And even if you find the ratings flawed, I seriously doubt anyone can come up with less flawed ratings. Your effort is very commendable, but I have a hard time being convinced that your calculations are better. In almost every instance where you show a different result than what the power rating says, I agree with the rating.

I do agree with your observation about Necromancy. But this follows the general principle of the (gold) cost of creatures. I know that if creature A costs twice as much gold as creature B, I'd be better off with one A than two B's (e.g. one Paladin is better than two Inquisitors). Generally, the higher the cost, the more bang for the buck you'd get. Now, if only real world economics works that way.

[John.Galt]

I didn't read much of the argument. I just have a couple things to say.

@ Asheera, sorry, didn't know you were a girl. One pretty much assumes anyone online is male...

@ JJ, I agree it is rather superficial. It is nonetheless useful. It's an estimate of the strength of a unit, trying to simplify the details as best as possible. However, I do think that a better (Although more difficult analysis) would come up with a defensive stat rating, offensive stat rating, etc. If I were taking the time to do this, which I am not, if I have any time I want to devote to heroes it goes into making my map, I would have done it this way:

Find averages for: Attack, Defense, Initiative, Hit Points, and Damage for all creatures. Then use the unit you are currently discussing's statistics to find the damage per round per weeks worth of troops that it would take from an "average creature" and damage per round per weeks worth of troops that it would deal to an "average creature".

Shooters make the analysis a bit more complicated, but for a shooter I would do it this way. Figure out the number of turns that it would take the "average" walker to reach your shooter and the damage you would deal to it in that time, etc.

Edit: The added bonus of this method allows you to compare upgrades to ALL other units, not just the other upgrade

Really, I prefer to look at each upgrade individually and analyze it during the game according to my needs etc. I don't like saying one upgrade is better than the others in all circumstances.

[Borsuc]

Use whatever scale you find fits your own experience. Squared would be way too sharp a scale, and linear would be a little too flat; I think the real scale is somewhere in between.

"Experience" from whom? Nival? Come on seriously, people always complain that math is not the only way to judge something etc... but how would YOU balance it? Don't you think anyone else will disagree because you are being entirely subjective? I mean seriously, you "think" it's somewhere in between. Math isn't perfect, any 'graph' is just an approximation, but it's waaaay better than the alternative. Not to mention the game works with math, since CPUs only know math.

I do agree with your observation about Necromancy. But this follows the general principle of the (gold) cost of creatures. I know that if creature A costs twice as much gold as creature B, I'd be better off with one A than two B's (e.g. one Paladin is better than two Inquisitors). Generally, the higher the cost, the more bang for the buck you'd get. Now, if only real world economics works that way.

No "real world" economics are fair, unlike the flawed Heroes system, so if I were to wish something here, I would say the opposite -- that Heroes needs a change, a more fair "bang for the buck" system.

[John.Galt]

In clarification, I realize I was unclear. When finding average damages per round, you should find the percentage of a weeks worth of troops your stack would lose to an average stack per round and vice versa.

Or instead of a weeks worth of troops you could set a gold amount's worth of troops, but whatever.

[danhvo]

"Experience" from whom? Nival? Come on seriously, people always complain that math is not the only way to judge something etc... but how would YOU balance it? Don't you think anyone else will disagree because you are being entirely subjective? I mean seriously, you "think" it's somewhere in between. Math isn't perfect, any 'graph' is just an approximation, but it's waaaay better than the alternative. Not to mention the game works with math, since CPUs only know math.

I am completely mystified by your post. I read it over and over and still got no clue what point you're trying to make. For example, your question, "how would YOU balance it?" Why the heck would you ask me that? Why would I want to balance anything? When did I ever complain that the game is imbalanced?

And then the statement about math. I don't get it. Are you saying that you know for a fact that the power ratings are not math (their graphs cannot be described with any equation) and didn't come from "math"?

[Borsuc]

I am completely mystified by your post. I read it over and over and still got no clue what point you're trying to make. For example, your question, "how would YOU balance it?" Why the heck would you ask me that? Why would I want to balance anything? When did I ever complain that the game is imbalanced?

Nonono. I meant that someone HAS to balance the game, in our case Nival -- how would they proceed? If you were in their place, how would you proceed if not with math? With guessing? I mean, if math cannot ever be trusted, then the numbers in the stats need to be guessed. And I'm sure if you bring 3 different people they will disagree with this and say that they favor certain tiers or creatures etc.

And then the statement about math. I don't get it. Are you saying that you know for a fact that the power ratings are not math (their graphs cannot be described with any equation) and didn't come from "math"?

No. I don't know whether they came from "math" but that is irrelevant -- the formula can be bad as well, or favoring certain tiers etc... it's not just to be "math" but to be fair -- for that I'll need to take a look at a formula or whatever, which there isn't any publicly available.

[Adicto]

This is like another power rating. Such things are useful for some game mechanics like dealing with number of neutral stacks, joining probability and so on, but they will never help players decide what is better for what in a game where the "power" of something depends on what you’re facing.

For example rot zombies are better vs dungeon thanks to their superior hit points but plage zombies are better vs stronghold because Weakening Strike lowers the rage points.

Another example: Lich masters’ damage potential comes from impressive attack & damage values, not from special abilities like that of the arcane archers, so they deal as much melee damage as ranged damage from far distance (1/2), which means better melee damage than inquisitors and even zealots if I remember correctly. That’s certainly a good thing against sturdy factions that tend to tie up your lichs forever. In the other hand archlichs need space to be able to use their death cloud, which is the source of their damage dealing power, very useful to destroy phantom forces, arcane crystals, war machines, and so on. In addition while 20 lich masters can raise a total of 1000hp (18 lich masters/10 wraiths/etc) 20 archlichs can kill 16 nightmares casting decay once (1075 damage, just like Implosion casted with expert destructive and 26 spell power). So the first was made for dealing with neutrals and early battles while the second is better suited for final battles.

[Macros the Black]

Well, that's a pretty flawed way of doing it. You're better off not calculating it like that, but looking at the different stats instead. Because some statistics are more important for some creatures.

For instance, the Blood Sister is much better than the Blood Fury if you use Bless on it. Look at the Arch Angel and the Seraph, if we're using your way of calculating it there'd be no difference, but if you just look at it you'll see the Seraph is definately aided by using Bless on it, whereas the Archangel cannot be Cursed.

[Asheera]

Of course there are situations like these, my calculations were for an "average" situation, just to get a point about the "average" power.