Skill Speed Alternative Formula at lvl 60

So... I was bored and sat down to poke at Skill Speed last night...
Exchanged gears a bit tho and fro and found out some fun stuff about Skill Speed...

Dunno if anyone else has wrote this but if someone did... sorry :v

Assuming you are at lvl 60:
tl;dr: GCD = 2.5-(1+Integer_Of((My_Sks-361-Integer_Of((My_Sks-354)/53))/26))*0.01


Step by step:
1) Grab your in-game Skill Speed (st_1)
- If its lower than 361, dont bother, the result is 2.5
GCD = 2.5-(1+Integer_Of((st_1-361-Integer_Of((st_1-354)/53))/26))*0.01

2) Deduct 361 from «st_1» (st_2)
GCD = 2.5-(1+Integer_Of((st_2-Integer_Of((st_1-354)/53))/26))*0.01

3) Take again the «st_1» and deduct 354 from it (st_3)
GCD = 2.5-(1+Integer_Of((st_2-Integer_Of((st_3)/53))/26))*0.01

4) Divide «st_3» into 53 and take only the integer value, ignore the decimals (st_4)
GCD = 2.5-(1+Integer_Of((st_2-st_4)/26))*0.01

5) Deduct «st_4» from «st_2» (st_5)
GCD = 2.5-(1+Integer_Of((st_5)/26))*0.01

6) Divide «st_5» into 26 and take only the integer value, ignore the decimals (st_6)
GCD = 2.5-(1+st_6)*0.01

7) Add 1 to «st_6» then multiply it by 0.01 (st_7)
GCD = 2.5-st_7

8) Deduct «st_7» from 2.5 and congrats, thats your GCD.


Other stuff:
It is pointless to try to find its value per point! Since that only gives a rough aproximation to the value in the game at the end of the day... so, something like:

2.5-(My_Sks-354)*Value_Per_Point = GCD

may tell us something... but not very helpful to build up from it :/


Ranges of Sks
2.5 GCD in the range of 354~360 sks
2.49 in 361~386 sks
2.48 in 387~413 sks
2.47 in 414~439 sks
2.46 in 440~466 sks
2.45 in 467~492 sks
and so on

So the jumps happens at
2.5 = 0 sks
2.49 = 7 sks
2.48 = 33 sks (26)
2.47 = 60 sks (27)
2.46 = 86 sks (26)
2.45 = 113 sks (27)
and so on

From the above we can see that besides the 2.49 jump, every jump happens at 26 sks, and every other jump at 27 sks, this means that 0.01 GCD = 26 sks, and 0.02 GCD = 52 + 1 sks.

The jumps, together with that +1 every 0.02 GCD is what makes calculating the value per point of sks quite pointless, since it only gives us an odd number that never really appears in front of us in the game, and doesn't help us know if the sks we have is optimized.


So lets say we have 1000 sks in game
Lets take the default out of it:
1000 - 354 = 646
Lets take out the odd 7 points for the jump to 2.49 GCD
646 - 7 = 639
From then on, lets see how many odd jumps of 27 can happen in it by dividing it into 53
639 / 53 = 12.0566
So it can happen 12 times, aka, 12 points we should deduct from that number
639 - 12 = 627
Now only the 26 jumps are in there, so lets see how many jumps that is
627 / 26 = 24.1153
So its 24 jumps, that means 0.24 GCD bedies the 0.01 odd one from the initial 7 sks points, so
2.5 - .01 - .24 = 2.25 GCD

BTW, 26 * 24 = 624, so we are wasting 3 Sks points!

Feel free to try it out o/ !!


Extra notes:
- I am assuming that the jumps happens in a constant jump of 26 and 27 sks after the 2.49 jump and having manually tested 10 of them in game with this pattern repeating in each situation.
- Due to DPS no having any piece of gear that gave me +1 sks, I was forced to try all of this as a tank... regardless the formula should be the same for all 2.5 GCDs affected by sks.
- I cannot claim that this formula also applies for GCD reduction skills like Huton, Greased Lightning, or Arrow Card. Since I am not sure if the Sks comes before of after these effects, or if Sks value per point does play a role into them.

If you want to see something interesting as well look at Empyreal Arrow while increasing Skill Speed.

It is an alternating pattern every 6 points of SPD.

0.01, 0.02, 0.01 etc

The cycle repeats every 66 points which also happens to be a 1% Reduction.

Yea, I only checked on 2.5 GCDs last night, so I assumed Empyreal Arrow would be quite the oddball for this.

Tho, right now Im wacking my head about Huton, since well, Im a NIN so I don't wanna waste in Skill Speed if I can avoid it...
Its frustrating I can't confirm my calcs so far cuz Im at work!! D:!!!

(what am I doing...)

Quote:

---

Originally Posted by Arkenne

Yea, I only checked on 2.5 GCDs last night, so I assumed Empyreal Arrow would be quite the oddball for this.

Tho, right now Im wacking my head about Huton, since well, Im a NIN so I don't wanna waste in Skill Speed if I can avoid it...
Its frustrating I can't confirm my calcs so far cuz Im at work!! D:!!!

(what am I doing...)

---

EA follows the same pattern you can just see it on a much smaller scale.

Reducing 15s by 1% VS 2.5

I'm too dead tired too try to understand exactly how your version works but here is an alternative correct SkS formula: GCD = (Rounddown((2.51-(SkS-334)/2641)*100))/100 credit to a friend of Dervy.

If I don't remember wrong Huton/GL is calculated before the rounddown, ie for GL 3 (Rounddown(((2.51-(SkS-334)/2641)/1.15)*100))/100. Though I need to test this in game as I might be wrong. But currently I'm about to pass out so that will have to wait until tomorrow.

Quote:

---

Originally Posted by Krindor

If I don't remember wrong Huton/GL is calculated before the rounddown, ie for GL 3 (Rounddown(((2.51-(SkS-334)/2641)/1.15)*100))/100.

---

Oh? that sounds interesting, may give it a shot, thanks!!

Tho looking at it again, that 334 looks odd to me... isn't 354 the base number with 0 sks from gears?

Edit: tried with my current 595 sks which in game results me in 2.04 GCD after Huton, but didn't get that number... either with 334 or 354... so I suppose this formula was for lvl 50? If so, we may need to reevaluate the 2641 constant there... or I just dont know how this works :v!!

Quote:

---

Originally Posted by Arkenne

Oh? that sounds interesting, may give it a shot, thanks!!

Tho looking at it again, that 334 looks odd to me... isn't 354 the base number with 0 sks from gears?

Edit: tried with my current 595 sks which in game results me in 2.04 GCD after Huton, but didn't get that number... either with 334 or 354... so I suppose this formula was for lvl 50? If so, we may need to reevaluate the 2641 constant there... or I just dont know how this works :v!!

---

If you reverse the steps of SS gain, it's always 26-27-26-27. The last 26/27 step takes you to 334, which is why my formula scales from 334 SS.

That extremely long formula you posted, whilst you could be onto something, can be simplified to:

2.51-((SS-334)*2641) (Courtesy of mauricedoison)

You then rounddown the values and you get the tooltip in game SS. It's worth to note, that despite our tooltip being rounded, our actual SS is not rounded at all. This value is accurate all the way up to approximately 900 SS, if I remember correctly.

Quote:

---

Originally Posted by MiniTyra

That extremely long formula you posted, whilst you could be onto something, can be simplified to:

2.51-((SS-334) / 2641) (Courtesy of mauricedoison)

---

Heeeh I see, well, yea, it was the last hour of the night so I cant deny I was onto something :v

(tho that formula there seems to be missing a division)
I'll try to correct the formula to backtrack to 334, only makes sense :v

Edit: Except that in the end it seems both formulas serves 2 different purposes :/

Yeah, it's division :P I need to correct that in my formula list... I keep forgetting.

Quote:

---

Originally Posted by MiniTyra

It's worth to note, that despite our tooltip being rounded, our actual SS is not rounded at all.

---

Have you tested this out? I've been doing a short amount of tests that shows that the actual SS is rounded. But those tests had quite the big error margin so if you've tested it and it shows that it isn't I'll take it as that.

Noticed that the formula that I wrote for GL/Huton is wrong now that I tested it as well. It is =(Rounddown(((2.51-(SkS-334)/2641)*(1-buff))*100))/100 so for GL that becomes =(Rounddown(((2.51-(SkS-334)/2641)*(1-0.15))*100))/100

I'm positive it isn't rounded. It makes no sense for SS to be rounded; it makes SS completely worthless

Quote:

---

Originally Posted by MiniTyra

I'm positive it isn't rounded. It makes no sense for SS to be rounded; it makes SS completely worthless

---

When I was testing I tried the difference between 650 (2.39) and 651 (2.38) SkS, the average difference between them was 0.015. Each SkS value was tested 3 times each time was 3 minutes and from my results it shows that just 1 SkS can make a difference. But like I said this data might be incorrect and I would like to have more people test it.