Talk:Accuracy Formula (UFO2000): Difference between revisions
No edit summary |
No edit summary |
||
| (3 intermediate revisions by 3 users not shown) | |||
| Line 22: | Line 22: | ||
- [[User:Bomb_Bloke|Bomb Bloke]] | - [[User:Bomb_Bloke|Bomb Bloke]] | ||
-------------- | |||
Well, these can. My problem is the ones that come afterwards. Like (2/(1/wacc^2+1/sacc^2)) <- this looks real ugly. And I was wondering about maybe delving into the statistics of hitting, but without decent formula display, I really shouldn't. Gaussians and things like that. No good. | Well, these can. My problem is the ones that come afterwards. Like (2/(1/wacc^2+1/sacc^2)) <- this looks real ugly. And I was wondering about maybe delving into the statistics of hitting, but without decent formula display, I really shouldn't. Gaussians and things like that. No good. | ||
- [[User:Arcozelo|Arcozelo]] | |||
-------------- | |||
<math>\frac{2}{\frac{1}{wacc^2} + \frac{1}{sacc^2}}</math> | |||
<math>\frac{2}{\frac{sacc^2}{wacc^2 \times sacc^2} + \frac{wacc^2}{wacc^2 \times sacc^2}}</math> | |||
<math>\frac{2}{\frac{wacc^2 + sacc^2}{wacc^2 \times sacc^2}}</math> | |||
<math>\frac{2 \times wacc^2 \times sacc^2}{wacc^2 + sacc^2}</math> | |||
2 / (1 / wacc<sup>2</sup> + 1 / sacc<sup>2</sup>) | |||
2 / (sacc<sup>2</sup> / (wacc<sup>2</sup> * sacc<sup>2</sup>) + wacc<sup>2</sup> / (wacc<sup>2</sup> * sacc<sup>2</sup>)) | |||
2 / ((wacc<sup>2</sup> + sacc<sup>2</sup>) / (wacc<sup>2</sup> * sacc<sup>2</sup>)) | |||
2 * wacc<sup>2</sup> * sacc<sup>2</sup> / (wacc<sup>2</sup> + sacc<sup>2</sup>) | |||
Don't ask me about gaussians, though. I've been out of college too long. | |||
- [[User:Bomb_Bloke|Bomb Bloke]] | |||
== Ist the formula itself incorrect.? == | |||
The lines | |||
:hp = acc * (max.health - current.health) / (max.health / 2) | |||
and | |||
:hp = 2 * acc * (max.health - current.health) / max.health | |||
are mathematically equal, but I'm really not sure if the formula | |||
itself is right. | |||
'''One Example:''' | |||
Accuracy=80 | |||
max.Health=50 | |||
current.health=25 | |||
hp = 80 * (50 - 25) / (50 / 2) = 80 * (25 / 25) = 80 (!) | |||
This can not be right. If I understood it correctly, the hp can be | |||
the half of the total accuracy in maximum. In this simple case I did | |||
use, the lost accuracy iss 100%! Not mentioned, what would happen, if | |||
the morale is also low... Negative accuracy can't be possible. | |||
Isn't it right this way? | |||
hp = acc * (max.health - current.health) / max.health /2 | |||
For the example I used: | |||
hp = 80 * (50 - 25) / 50 / 2 = 80 * 25 / 50 / 2 = 20 | |||
For the mp it's the same: | |||
80 accuracy | |||
50 morale | |||
mp = acc * (100 - current.morale) / 50 | |||
mp = 80 * (100 - 50) / 50 = 80 * 50 / 50 = 80 | |||
'''I guess this is the proper way:''' | |||
mp = acc * (100 - current.morale) / 200 | |||
mp = 80 * (100 - 50) / 200 = 80 * 50 / 200 = 20 | |||
Am I wrong, or is there a mistake in this article :-) ? | |||
BTW: I think this formula style is OK: | |||
hp = acc * (max.health - current.health) / max.health / 2 | |||
---- | |||
Looks like a mistranslation. The LaTex is fine. | |||
As you noted, the intent is to have 50% of the penalty come from health and 50% from morale. However, omitting the parentheses is technically accurate but misleading, because division is not commutative in general. | |||
The clear way to write the parenthesized versions should be: | |||
hp = (acc * (max.health - current.health) / max.health) / 2 | |||
mp = (acc * (100 - current.morale) / 100) / 2) | |||
i.e. | |||
hp = acc * (max.health - current.health) / (2*max.health) | |||
mp = acc * (100 - current.morale) / 200 | |||
-- [[User:Zaimoni|Zaimoni]], 14:28 Sept 20 2006 CDT | |||
Latest revision as of 19:27, 20 September 2006
There seems to be some kind of problem with the latex math formula generating. Either that or a problem with me. I'd appreciate some feedback about this, since it would look nice to have actual pretty looking formulas. And there are some very ugly formulas coming.
- Arcozelo
I guess MediaWiki isn't set up here. But, couldn't those formulas be simplified somewhat?
<math>hp &=& acc \times \frac{max.health - current.health}{\frac{max.health}{2}}</math>
<math>hp &=& \frac{2 \times acc \times (max.health - current.health)}{max.health}</math>
<math>mp &=& acc \times \frac{100 - current.morale}{\frac{100}{2}}</math>
<math>mp &=& \frac{acc \times (100 - current.morale)}{50}</math>
Or, even down to non-TeX format:
hp = acc * (max.health - current.health) / (max.health / 2) hp = 2 * acc * (max.health - current.health) / max.health
mp = acc * (100 - current.morale) / (100 / 2) mp = acc * (100 - current.morale) / 50
Well, these can. My problem is the ones that come afterwards. Like (2/(1/wacc^2+1/sacc^2)) <- this looks real ugly. And I was wondering about maybe delving into the statistics of hitting, but without decent formula display, I really shouldn't. Gaussians and things like that. No good.
- Arcozelo
<math>\frac{2}{\frac{1}{wacc^2} + \frac{1}{sacc^2}}</math>
<math>\frac{2}{\frac{sacc^2}{wacc^2 \times sacc^2} + \frac{wacc^2}{wacc^2 \times sacc^2}}</math>
<math>\frac{2}{\frac{wacc^2 + sacc^2}{wacc^2 \times sacc^2}}</math>
<math>\frac{2 \times wacc^2 \times sacc^2}{wacc^2 + sacc^2}</math>
2 / (1 / wacc2 + 1 / sacc2) 2 / (sacc2 / (wacc2 * sacc2) + wacc2 / (wacc2 * sacc2)) 2 / ((wacc2 + sacc2) / (wacc2 * sacc2)) 2 * wacc2 * sacc2 / (wacc2 + sacc2)
Don't ask me about gaussians, though. I've been out of college too long.
Ist the formula itself incorrect.?
The lines
- hp = acc * (max.health - current.health) / (max.health / 2)
and
- hp = 2 * acc * (max.health - current.health) / max.health
are mathematically equal, but I'm really not sure if the formula itself is right.
One Example:
Accuracy=80
max.Health=50
current.health=25
hp = 80 * (50 - 25) / (50 / 2) = 80 * (25 / 25) = 80 (!)
This can not be right. If I understood it correctly, the hp can be
the half of the total accuracy in maximum. In this simple case I did
use, the lost accuracy iss 100%! Not mentioned, what would happen, if
the morale is also low... Negative accuracy can't be possible.
Isn't it right this way?
hp = acc * (max.health - current.health) / max.health /2
For the example I used:
hp = 80 * (50 - 25) / 50 / 2 = 80 * 25 / 50 / 2 = 20
For the mp it's the same:
80 accuracy
50 morale
mp = acc * (100 - current.morale) / 50
mp = 80 * (100 - 50) / 50 = 80 * 50 / 50 = 80
I guess this is the proper way:
mp = acc * (100 - current.morale) / 200
mp = 80 * (100 - 50) / 200 = 80 * 50 / 200 = 20
Am I wrong, or is there a mistake in this article :-) ?
BTW: I think this formula style is OK:
hp = acc * (max.health - current.health) / max.health / 2
Looks like a mistranslation. The LaTex is fine.
As you noted, the intent is to have 50% of the penalty come from health and 50% from morale. However, omitting the parentheses is technically accurate but misleading, because division is not commutative in general.
The clear way to write the parenthesized versions should be:
hp = (acc * (max.health - current.health) / max.health) / 2
mp = (acc * (100 - current.morale) / 100) / 2)
i.e.
hp = acc * (max.health - current.health) / (2*max.health)
mp = acc * (100 - current.morale) / 200
-- Zaimoni, 14:28 Sept 20 2006 CDT