Pistoleer Archive
Thread: Porposed Fix for Speed
Well.. It still has problems, but that's somewhat due to the weapons being so **edit** fast in the game in general. The numbers can of course be adjusted for it to give a lesser benefit, but you cant make it all that much less orthe modsjust wont be effective enough. You can get 1.2 speed pistols, and if you want to make it so that you wouldnt reach the cap with pistols.. Who wants to get master pistoleer for just a 30% increase to speed?
The thing is - currently with just pistoleer you dont even get to a +100 mod, so at the +79 mark you'd start capping out with 1.8 speed or something around that pistols. It's still a LOT more acceptable than the current system, cause you can get those over 2.0 delay pistols with higher damage to get full benefit of the speed bonus like other professions can, so pistols would be more at an equal footing there.
You're not understanding the problem with the current speed implementation.
Mathematically: An equation with a decreasing percentage in the divisor increases exponentially.
That's why everyone ends up shooting at the cap no matter what - because the equation is broken. It doesn't matter if you have a 7 speed rifle or a 2 speed pistol, if your speed skill is high enough, you shoot once per second. Which means that the pistol user sees a 100% increase in damage, but the rifleman sees a 600% increase in damage.
Another side of the issue:
Starting from zero skill it takes 50 points to hit 50% reduction in delay and double your damage output.
From 50 it only takes 25 more points of skill to double your damage again.
From 75, it only takes 12.5 points of skill to double your damage.
If you're at 90, it only take FIVE points of skill to double your damage, and you're now doing (if you're not hitting the cap) over TEN TIMES as much damage - the equation is logarithmic. It was a bad design choice and needs to be fixed.
Noules,
1/(1+(1/speed)) is not linear, it's log, albeit it's quasi-linear in the range for speed 0 to 50. In other word, as speed skill increase, the net speed gain decrease.
A linear graph would be (raw speed) * (1-speed) which is the way it currently is. NOTE: for clarity sake, I'm expressing speed as a percentage, so just take the skill # and divide by 100; eg pistol speed skill 65=speed .65
The issues are
1) dabbler can get speed over 100
2) since pistol weapon speed is already low, a pistoleer net speed will cap out early. The slowest unsliced master-crafted pistol I seen on my galaxy Corbantis is the DX2 at 2.1. Hencethe cap for pistol speed skill is 52. Any skill mod after this is wasted. If I use special, then it'll cap at a little higher skill #, but the same principle apply.
So 2 thing we're asking for is
1) some type of capfor the net speed. For example, at speed skill 95, you can reduce a 20 second weapon plus special delaydown to 1 sec, which is crazy. There should be a cap for the reduction. I'm proposing 4x, so the fastest one can shoot for this 20 sec wpn/special is5 sec. This imply a skill cap at 75. If people need a little more speed, then 5x (skill cap 80) suffice. Anything higher is just too powerful.
2) some "thing" which won't waste the higher pistol speed skill #. Since pistol speed is already low, pistoleer will hit the net speed way too soon. Give us something so these "over cap" are not wasted. One thought is to give us some mean to increase damage by sacrificing speed (either by allowing smuggler/weaponsmith to tune the weapon for increased damage at a slower speed). Another thoughis to increase the damage done for any skill over the cap, subject to #1 above. For example, I'm a master pistoleerwith pistol speed skill 74. I auto-attack using my 2.1 FWG5. The equation for net speed is 2.1 * (1-.74) = 0.54 net speed which waste our higher skill. So if the net speed cap at 1 second, then increase damage by 27%; eg 2.1 * (1-.47) = 1 second time .74 (pistol speed) - .47 (net speed cap) = 27% damage increase.
You're close, but you're missing a piece of the puzzle that makes all the difference!
Yes, "Speed * (1 - speed skill / 100)" is a linear equation. At 0 skill your speed is unchanged, at 50 skill your speed is half, and at 100 skill your speed is zero.
BUT the resultant speed from that equation is then used in the DENOMINATOR of another equation: "DPS = damage / speed". And when the denominator of an equation decreases linearly towards zero, the value of the fraction increases exponentially to infinity.
When you take Noules's new equation and plug it into the DPS calculations, you get a linear increase in DPS: 20 skill results in 120% DPS. 40 skill results in 140% DPS, all the way up to, and even past, 100: 100 skill results in 200% DPS, 125 skill results in 225% DPS.
For a more in depth analysis, pop back to the top of this thread and scan down till you come to my first long post. It's all spelled out clearly in there - with the effects, pros, and cons of both the original and proposed new equations.
YetiIronFist:
I understand the difference between the 2. Noules equation is a speed reduction inverse to speed skill. For DPS, one divide by speed, thus Noules equation become linear for DPS.
Likewise, the current speed equation is a linear speed reduction to speed skill. Thus for DPS, it become negative inverse to speed skill(eg DPS approaches infinity as speed skill increases).
So the question is what's more important? A linear speed decrease or a linear DPS increase?
For me, DPS is just one of many way to determine effectiveness. Heck, DPS is not enough, we need to throw in accuracy and HAM cost to have a better sense of effectiveness (and yes, I'm collecting data for this analysis).
So for me, I prefer a linear speed decrease to a linear DPS increase. DPS is a calculated, secondary stat. speed is a primary stat.
And I do understand that Noules equation does not require an artificial cap since the speed decrease is an inverse equation with 0 as the limit, one don't need to worry about infinity as speed skill increase whereas one need to worry about infinity with the current equation.
So it really boil down to is what's more important, linear speed or linear DPS?
Speed by itself means nothing.
Linear, or reasonably non-linear, effectiveness (in this case DPS) and balance (in this case not destroying class, weapon, and specials balance) is what is important.
If skill gain were linear, if mob toughness were linear, then it make sense for DPS gain to be linear. But that's not how thing are. Skill gain are geometric as are advanced mob/npc. For example, for me to gain 5 pistol speed, I only need 10 minutes (to go from Marksman 0/0/0/0 to 0/0/0/1) or 10 weeks (to go from pistoleer 0/0/0/0 to 0/0/0/1).
If after ten weeks of pistoleer, I only get the same increased (5%) in DPS as a newbie marksman, I will not be very satisfied. (yeah, I know, the pistoleer gain would is really 6%, not 5%, but the principle is the same).
Also, Noules formula means the best gain is 74% increase in DPS (master pistoleer without master marksman). There's no way a master pistoleer can take on the advanced creature with only a 74% increased in DPS.
We're on the same page in that we all want balance (for both PvP and PvE). We all recognizes the 2 shortcoming of the current system (namely the singularity of speed as speed skill approaches 100 and the low effective cap for pistoleer).
I like the system the way it currently is in that newbie damage potential is low and their gain is slow whereas master damage potential is high and their gain is proportionally higher.
The problem I have with noules formula is it make master pistoleer nearly worthless. Why would I spend 3 months to gain master pistoleer for an increased of only 44% increasedDPS (44 pistol speed) when I can spend 3 minutes with a smuggler who can sliced my weapon for a 44% increased in DPS (44% increased in weapon damage)?
The current system let you go from a 25% increased DPS (marksman 0/4/0/0) to 400% increased DPS (master pistoleer). Now, this is something I'm willing to spend 3 months working on.
The problem I have with the current system is one can also go from 400% increased to 200,000% DPS, which I think is uber-crazy.
You're right that with Noules's equation it means that players that currently have high weapon speed skills (more so for other classes than pistoleers because we hit the cap more easily) would see a sudden decrease in their damage output (DPS).
Noules and I tossed that back and forth in this thread or maybe a different one, and both liked the idea of adding an additiona skill that you'd gain as you went up the trees that was a damage skill - giving an increase to the other half of the DPS equation.
So basically what's being suggested is that first the speed equation gets fixed so that concepts like "Pistoleers are fast, but do less damage, while riflemen shoot more slowly, but for higher damage" and "Stopping Shot does a lot of damage, but it's slow, while Fan shot does less damage, but is a better / high level skill because it is faster and will do more damage over time" actually work and don't fall apart with high speed skills. Then secondly you add damage skill bonuses (just like the ones that the melee classes already get - this isn't a new concept for the game) to restore the pve balance.
YetiIronFist/Noules: I'll answer your question with another question 
Is speed broken in general or only at high end? eg is speed broken at 0, 25, 50, 75?
Now something to consider: Let's use noules formula. A master pistoleer DPS is 1.74 that ofa newbie with zero experience. Now, DPS average only. To consider probability of true damage, one must consider standard deviation also. Let's make fantasy and say the newbie can use a FWG5 50 - 200 at speed 2.0. DPS for newbie is 125; master pistoleer, 218. The standard deviation is around 63%. So what this mean is roughly 5% of the time, the newbie will kill the master pistoleer. Sorry, but this possibility is not very satisfying to me. No newbie should be able to take down a master pistoleer 5% of the time. Now consider the same newbie against a master marksman (163 DPS). Standard deviation remain the same. The newbie will kill the master marksman 1/3 of the time. Definitely not good.
My answer: speed the way it currently work right now work fine until the skill goes past 75 (read BH dabbler, rifleman). Since it work great up until some limit, just cap it at that artificial limit.
For example, Pistol line I BH = 10 / Pistoleer first speed modifier (lvl I too) = 6. If someone get them both, you only get a +10 modifier.
This way, even being master pistoleer with all BH modifier, you can't reach 100. I think this solves the problem profession wise actually.
And the advantage would be to REALLY think twice where to go to be "the best at pistol".
Like it is now, it is MUCH too easy for a BH to gain new speed bonuses...if their own are better than the profession, why should they bother "wasting" some more points, or to be master ?
Now, what about errasing all duplicated modifiers when you reach Master ?
The only way to be over 200 would be to be master BH and Master Pistoleer.
Good luck
It's true that equation would make things alot harder, but I dont think that's necessarily a bad thing. Actual good specials could be added then, because currently, things have been nerfedleft and rightbecause players in general have been too powerful compared to the mobs since release. It would definitely require some more balancing work than just changing how the equation works, but I'm pretty convinced the work is small compared to the problems that'll come from that equation later on in the game if it's not changed.
A simple cap would solve the balance problem, but it'd make the other (which I think is bigger) problem even worse. What happens when more items, more skills, more guns, more things in general get added to the game? Everyone is already at the cap, with zero expansions, loot additions, skill mod items, anything. And that's a cap that CANT be changed for the sake of game balance, because otherwise you'd reach that infinity point quite fast again, and there's nothing to do about it with how it's calculated right now.
It just doesnt work in a mmorpg that intends to have a lifespan of over half a year, you'll need to start adding new neat stuff to the game eventually to keep it interesting for the players. I see the cap as a huge limit for that, it's totally eliminating any additions related to speed in any way, which is a big thing overall and I think should remain as something that makes a difference.
YetiIronfist wrote:
Aye - good points Noules.
I was thinking something along the lines of what you were saying with giving more speed at higher skills so that you got a similar relative effect (say... receiving a 5 or 10 percent relative increase to damage).
Another solution to bring things back to where they were before, but without destroying the balance between fast / slow classes, fast / slow eapons, fast / slow specials (having everything fire once per second regardless) would be to add a damage skill as well. So you would receive the speed bonuses you do now, which would give you less of a benefit at higher skill levels with the fixed equation, but also receive bonuses to damage with some of the higher skills. Something like this may be needed to bring players back into balance without modifying pve parameters.
I absolutely agree about the damage skills. Melee classes get accuracy, damage and speed (all of which improve effective DPS) while ranged classes only get accuracy and speed. This makes balancing the two rather difficult as melee classes will improve faster (they get the third stat, and those all have multiplicative effects on the DPS), meaning melee has to either start out worse or end up better in terms of DPS, or both.