Percent speed increase and interval

TwilightSong - Sanctuary

I've been looking at and trying to figure out the rounding of interval increases. From my testing with relentless courage it appears to me that simple rounding to the closest attack speed is not what is occurring. At a 22% speed increase the attack rate goes to 2.22 from 1.67, while if it were rounded it should have gone to a 2.0 attack rate.

I am wondering if anyone know the formula that pwi is using to round. Overall i want to find out the greatest possible speed increase i could achieve using relentless courage, at different starting attack rates.

If anyone has more information on this i would appreciate your comments

Brigid - Harshlands

Archer Attack Speed Basics by Elenacostel covers it pretty well.

To summarize though, all attack speed calculations are done in interval between attacks, rather than attacks per second. +% attack speed buffs are actually -% attack interval buffs. Thus, 22% = -22% interval between attacks, which is actually more like 28% increase in attack rate. Also, it then rounds down to the nearest 0.05 seconds.

Thus, from your example:
1.67 attack rate is 0.6 seconds between attacks.
22% reduction of that results in 0.468 seconds between attacks, which rounds to 0.45 seconds between attacks.
0.45 seconds between attacks is 2.22 attacks per second.

Elenacostel - Heavens Tear

I was never really able to exactly get the formula just right, for % increases. For the most part, applying it to the attack interval instead of the attack speed seems to bring the closest result, but there are a few examples that are rounded a bit differently that I could never explain.

salvati0n

Elenacostel - Heavens Tear wrote: »

but there are a few examples that are rounded a bit differently that I could never explain.

Really? O_o Do tell.

Elenacostel - Heavens Tear

http://pwi-forum.perfectworld.com/showthread.php?p=6471541#post6471541 - read the next few pages

http://pwi-forum.perfectworld.com/showthread.php?t=718442

Lylfo - Dreamweaver

i bet all you need now is astrelle

Ewen - Harshlands

TwilightSong - Sanctuary wrote: »

I've been looking at and trying to figure out the rounding of interval increases. From my testing with relentless courage it appears to me that simple rounding to the closest attack speed is not what is occurring. At a 22% speed increase the attack rate goes to 2.22 from 1.67, while if it were rounded it should have gone to a 2.0 attack rate.

I am wondering if anyone know the formula that pwi is using to round. Overall i want to find out the greatest possible speed increase i could achieve using relentless courage, at different starting attack rates.

If anyone has more information on this i would appreciate your comments

22% from 2.22 is ~0.49 and 2.22-0.49=1.73, I don't know where the -0.06 comes from, but I know that 22% reduction doesn't reduce the attack rate from 2.22 to 2.0 under any circumstances.

Brigid - Harshlands

Ewen - Harshlands wrote: »

22% from 2.22 is ~0.49 and 2.22-0.49=1.73, I don't know where the -0.06 comes from, but I know that 22% reduction doesn't reduce the attack rate from 2.22 to 2.0 under any circumstances.

Brigid - Harshlands wrote: »

1.67 attack rate is 0.6 seconds between attacks.
22% reduction of that results in 0.468 seconds between attacks, which rounds to 0.45 seconds between attacks.
0.45 seconds between attacks is 2.22 attacks per second.

Rounding. Already explained why her numbers work out how they do, instead of how she thinks they should. It's all due to the way the game treats attack speed buffs.

Olba - Sanctuary

Ewen - Harshlands wrote: »

22% from 2.22 is ~0.49 and 2.22-0.49=1.73, I don't know where the -0.06 comes from, but I know that 22% reduction doesn't reduce the attack rate from 2.22 to 2.0 under any circumstances.

You're wrong there. What the OP is saying that 1.22*1.67 rounds to 2.0, not 2.22. However, that's not how attack rate is counted anyway, so that's meaningless.

Also, mathematically speaking, you're overcomplicating your math. What you're doing looks like this:
2.22 * 0.22 = 0.49
2.22 - 0.49 = 1.73

Whereas you could just simplify that to be this:
2.22 - (2.22*0.22)
= 2.22 (1-0.22)
= 2.22 * 0.78
= 1.73

And yes, though you get less lines, you need memory, I don't. Requirement for memory is a complication.

What happened there was, as was explained, gotten from the way attack speed works. The attack speed boosts are all applied to the attack interval. In this case, it'd go like this:

1.67 = 1/0.6 -> 0.6 Interval
0.6 * (1-0.22) -> Interval *(1-speed boost)
= 0.6 * 0.78
= 0.468 -> Attack Interval after speed boost
=~ 0.45 -> Rounded to the nearest 0.05 interval
1/0.45 = 2.22 -> Attack Rate = 1/Attack Interval

MagicHamsta - Lost City

Elenacostel - Heavens Tear wrote: »

I was never really able to exactly get the formula just right, for % increases. For the most part, applying it to the attack interval instead of the attack duration seems to bring the closest result, but there are a few examples that are rounded a bit differently that I could never explain.

It be time to tweak the forumlae?
b:puzzled

TwilightSong - Sanctuary

Thank you all i understand it now and running my numbers with the given formula they match up nicely. I am however sad too see that a genie would have to have 136 strength in order to make attack speed increase from 2.22 to 4.0. I'll take solace that it only requires 92 to go from 2.22 to 3.33.

Partypooper - Harshlands

Thats a lot of number

MagicHamsta - Lost City

Partypooper - Harshlands wrote: »

Thats a lot of number

This be nothing.
b:chuckle

Olba - Sanctuary

TwilightSong - Sanctuary wrote: »

Thank you all i understand it now and running my numbers with the given formula they match up nicely. I am however sad too see that a genie would have to have 136 strength in order to make attack speed increase from 2.22 to 4.0. I'll take solace that it only requires 92 to go from 2.22 to 3.33.

Do note that generally speaking, skills with speed increases don't stack. So if you're planning to use a genie to reach 5 APS on a fist BM or a sin, that won't work.

Partypooper - Harshlands wrote: »

Thats a lot of number

Um... what?

MagicHamsta - Lost City wrote: »

This be nothing.
b:chuckle

Just take any university math course and you'll see a lot more of a lot scarier stuff.