Fen_Hydra's Point Costing Formula

[Fen_Hydra]

Originally Posted by GreyOwl

The thing is, you can't really come up with a set number of points for each special power (unless I'm misunderstanding your system). For example, Flying is worth a lot more on a ranged figure than it is on a melee figure. Double Attack is worth more on a figure with 5 Attack than it is on a figure with 2 Attack.

What I'm trying to do for the special powers is incorporate the different stats that may affect it into how much it costs. For example,

- Double Attack:

---o 5 + ( [amount of range] / 2) + ( [amount of attack] * 1.5)
---o Do not use while the figure also has a special attack.

I try to incorporate range and attack into the formula, and after coming up with it, I then test to see it's accuracy with official and custom units.

(The second part is to avoid confusion as to whether or not you can use your normal and special, or special twice, or normal twice)

[GreyOwl]

Oh, I see. That's an interesting idea. How would you handle a special power whose value changes based on another special power? To take an old but simple example, Ghost Walk is much more valuable to a figure that has Disengage than to one that does not.

[flameslayer93]

Originally Posted by GreyOwl

Oh, I see. That's an interesting idea. How would you handle a special power whose value changes based on another special power? To take an old but simple example, Ghost Walk is much more valuable to a figure that has Disengage than to one that does not.

That's what we call Phantom Walk

Actually, I think it'd be impossible to use Ghost Walk w/out Disengage. Not too sure why though, maybe it has something to do with the "In my space" disengagement attack... I need a better way to word it.

[GreyOwl]

I know, I picked a simple example on purpose, but you get my point, right?

[Fen_Hydra]

Originally Posted by GreyOwl

Oh, I see. That's an interesting idea. How would you handle a special power whose value changes based on another special power? To take an old but simple example, Ghost Walk is much more valuable to a figure that has Disengage than to one that does not.

I'd probably do something like: (just random numbers)

Ghost Walk:
Move X 2, and add 10 more if figure has Disengage.

Or for a different case, if I notice something that may influence it, then I would add in a not that it increases by X if they also have Y power(s).

Update: Thorian Speed, Grapple Gun 25, Dragon Healing, Dragon Swoop added.

[flameslayer93]

Out of curiosity, why is Defense affecting Thorian Speed? Thorian Speed is already a power that allows attacks to be prevented.

For Dragon Healing&Dragon Swoop
You also know you can just say that you'd add 12&17 to the final numbers respectively instead of adding a number twice.

[GreyOwl]

So is there a way to use this method to cost customs that don't use official powers, i.e. original powers?

[Fen_Hydra]

Originally Posted by flameslayer93

Out of curiosity, why is Defense affecting Thorian Speed? Thorian Speed is already a power that allows attacks to be prevented.

For Dragon Healing&Dragon Swoop
You also know you can just say that you'd add 12&17 to the final numbers respectively instead of adding a number twice.

Good points... I just made a few slight modifications.

Originally Posted by GreyOwl

So is there a way to use this method to cost customs that don't use official powers, i.e. original powers?

Not really, you would just have to determine the cost not including those special powers, and then add on how much you think it would be worth and test from there.

Although I am thinking of creating custom powers, making custom cards, testing, and seeing if I can add on some custom powers to the list. (I will be taking suggestions for general power ideas to add on.)

[Kaiyu]

I'm going to give this a try (if the powers are listed) when I play Heroes of 'Scape to see if the Heroes are approx the same value while using your equations. Since I know that the Heroes at Level 1 usually lie within 80-100 points, it'll be some somewhat useful playtesting. It wont be as solid as other playtests, but it will be good for the shear number of Heroes played.

[TurtleKing99]

http://www.heroscapers.com/community...ead.php?t=1084
http://www.heroscapers.com/community...id=54&act=down

These may help; I found them a while ago. They were a good starting point, but yours will probably end up better. Good luck!

EDIT: perhaps this as well:
http://www.boardgamegeek.com/filepag.../hstksetup-exe

[lefton4ya]

Only things is you have to multiply things and not just add, and not only multiply but exponentially increase and maybe use logarithms. Take Isamu as an extreme example. If he had 5 defense OR 5 life, he would be worth a lot more than 50 points, maybe 100. If Iron Golem had double the life he would be worth more than double his points cost, although I do admit I like your taking life and multiplying by 1/3 of defense which gets close to correct. Adding disengage/stealth flying to a wyvern or dragon should have a 1.25 multiplier not add the same point for a currently 35 point and 185 point unit - but if you want to stick with adding points for powers, then basically all powers need to take life and defense (or your life equation) and also maybe attack into each power's equation. For instance, all else being equal, any movement power (disengage, slither) is worth more to someone with attack of 6 versus attack of 2, and a defensive power such as one-shield defense is worth more to someone with a higher attack.

I once did regression testing on a table of units to try and find a formula but realized I needed not only to quantify powers but designate types of powers as maybe special attacks should add points to cost, movement powers should multiply cost, and defensive powers should exponentially increase cost based on a power of life and defense. I created a spreadsheet of powers by type and the only way to get a true point cost would be to link that with current units (with revised point cost so each unit is rated B) and do non-linear regression testing which even with a program like SAS or R would take an experienced statastition/mathematition hours of work programming and massaging input equations. I do like how you put other stats in the cost of each power and this equation could be tested using regression testing and might be found to be semi-accurate. But even then, there is no guaruntee that it will find a working formula even if you use and equation that has 30 variables using addition, multiplication, logarithms, exponential and quadratic formulas all in the same equation.

Sorry to burst you bubble with the major mathematics required for unguaronteed results, but I would not accept any formula not determined using an advanced statistical model, although yours is a hell of a lot better than anything previously made so I do commend you on that.

[lefton4ya]

Ignore repeated post.