Page 117 of 163
Re: Ballistic Bro Changelogs Discussion
Posted: Thu 31 Jul , 2014 5:37 pm
by iRobot
Rymosrac wrote:Well yeah, not all exponential functions are quadratic. But pretty much all quadratic functions are exponential.
Not all 1's are 2's but pretty much all 2's are 1's.
Now you're talking, this is the math I remember!
Re: Ballistic Bro Changelogs Discussion
Posted: Thu 31 Jul , 2014 5:40 pm
by taries
Too much math in here ffs.
Re: Ballistic Bro Changelogs Discussion
Posted: Thu 31 Jul , 2014 5:44 pm
by Rymosrac
Azarael wrote:
Not sure what you're getting at.
I just meant that vanilla BW does a lot of ignoring physics and realism. For a mod that completely abandons a unified design principle and balance in favor of letting people dick around with it however they want, there's a hell of a lot of complaining about changing things from vanilla in parts of the community.
Re: Ballistic Bro Changelogs Discussion
Posted: Thu 31 Jul , 2014 5:47 pm
by Azarael
Ah, I gotcha.
Btw:
Rymosrac wrote:You forgot to close your head.
Actually, that explains a lot.
Re: Ballistic Bro Changelogs Discussion
Posted: Thu 31 Jul , 2014 6:10 pm
by Pinky
Rymosrac wrote:Well yeah, not all exponential functions are quadratic. But pretty much all quadratic functions are exponential.
Umh... not quite. You're correct when you call the number
2 in a quadratic equation the exponent, but these types of equations are commonly referred to as
power laws. The term
exponential equation is reserved for equations of the form
y = e^x, where e is the Euler number.
Azarael wrote:I'm concerned about the performance of Square() versus **. Can you offer any clarification? I can see how Square is internally handled as x * x, but, say, x ** 1.4?
Theoretically, there should be the exp() function handling this. A common pow() function is normally calculated via pow(x,y) = pow(a, y * log_a(x)), where a is usually 2 or e, depending on the implantation in the math library. So, performance could be an issue here...
Re: Ballistic Bro Changelogs Discussion
Posted: Thu 31 Jul , 2014 6:13 pm
by Azarael
Exp
native(191) static final function float Exp (float A)
Calculates e to the power of A, the exponential function.
Re: Ballistic Bro Changelogs Discussion
Posted: Thu 31 Jul , 2014 6:15 pm
by iRobot
You are no longer subscribed to this topic.
Re: Ballistic Bro Changelogs Discussion
Posted: Thu 31 Jul , 2014 6:32 pm
by Rymosrac
Pinky wrote:Umh... not quite. You're correct when you call the number 2 in a quadratic equation the exponent, but these types of equations are commonly referred to as power laws. The term exponential equation is reserved for equations of the form y = e^x, where e is the Euler number.
Damnit, Jim! I'm a robotics tech, not a mathematician!
Edit: Which I suppose obligates me to complain about any so-called power law that isn't P = IV.
Re: Ballistic Bro Changelogs Discussion
Posted: Thu 31 Jul , 2014 6:43 pm
by Pinky
Azarael wrote:Exp
native(191) static final function float Exp (float A)
Calculates e to the power of A, the exponential function.
That's the one. The exact internal implementation depends on the processor used (many mathematically functions are hard-coded within the CPU), unless your programming language emulates it via software. In any case, my guess is that exp() will be slower than x*x - if you are concerned about efficiency...
Re: Ballistic Bro Changelogs Discussion
Posted: Thu 31 Jul , 2014 6:46 pm
by Azarael
OK, so I'm given to understand that the various decay methods for the DamageScale variable, which takes values from 0 to 1, would end up looking like this if we were to avoid having a cliff at the end of the damage radius:
If so, exponential decay looks to provide a really sharp decline, so much so that it might be unworkable in gameplay terms. Hits that weren't really, really close to the enemy would deal very little damage and ones that were on target would need to one shot to be viable.
Any native function will be in C++.