Haste VS ATK VS ACC Your view ?

[etchdog2000]

This is pure bullshit if I've ever seen it. Did you just find a pull to pull it out?

Accuracy isn't linear. Never has been, never will be. At lower levels, accuracy play a much more important role then it does on higher levels. You can't get 100% accuracy, so if your accuracy is 99%, no matter how much mroe you add, you won't ever reach 100%, ever. This is proof that accuracy isn't linear.

In the same way, +10 accuracy might make a difference, but +10 more won't matter as much. And thus, your observation is meaningless. Your tests are wrong, as accuracy can never, ever, be proven in percentages.

Oh is it now? do you know how easy that is to test. I have very very little ACC in my gear setup(as most DRK gearing for an attk setup). I feel every little bit of acc that I lose. I simply parsed with a sniper ring and then with out. Every fucking time its 2-3%. next I took off Ohat and put on a celeta. 4-6% change in ACC off the original setup. And hell thats with sushi, and 4 merits in scythe!

Its not a hard thing to test.....

 

[etchdog2000]

I am done on this thread. If you really want to know you can go to any job forum on the net and get opinions of people who have tried it. They talk about these types of setups on KI and lolAlla with a decent amount of frequency.

5% haste will only get people who attack every 3 seconds an extra hit per minute. .05 * 60= 3

 

[Uzor]

Oh is it now? do you know how easy that is to test. I have very very little ACC in my gear setup(as most DRK gearing for an attk setup). I feel every little bit of acc that I lose. I simply parsed with a sniper ring and then with out. Every fucking time its 2-3%. next I took off Ohat and put on a celeta. 4-6% change in ACC off the original setup. And hell thats with sushi, and 4 merits in scythe!

Its not a hard thing to test.....

You can't parse accuracy like that... Seriously, how stupid are you?

Yes, you can parse and see that adding 10 accuracy gives YOU a certian % of increase for that fight, but that's not proof that 10 acc gives that much accuracy to everyone. It only gives it with your stats, skill levels etc. Accuracy is *NOT* linear. What you're saying is that adding 20 accuracy will add 4-6% accuracy even if your accuracy is 95%, which would bring your accuracy to... 99-101%? How do you hit 101% of the time?

I am done on this thread. If you really want to know you can go to any job forum on the net and get opinions of people who have tried it. They talk about these types of setups on KI and lolAlla with a decent amount of frequency.

5% haste will only get people who attack every 3 seconds an extra hit per minute. .05 * 60= 3

Yeah. Let me waltz over to KI where they state that Sniper Rings suck for DRK because it gives -dark resist... That will apparently make dark knights gimp... because, you know, they have "dark" in their name. While we're on it, let's mention the largest noob community of them all, Allakhazam, where 90% of the population still believe that Millionaire's Desk have a hidden Treasure Hunter effect.

Yeah. Great idea. Let's listen to what they have to say.

As for your so-called proof, wtf is " 5% haste will only get people who attack every 3 seconds an extra hit per minute. .05 * 60= 3" supposed to mean? That's the worst mathematical proof I have ever seen. You're stull assumign that we only want to hit more often, and you completely disregard the fact that haste doesn't work like that. What you're saying is basically "Your delay will be the same, but at the 1 min mark you'll suddenly double attack if your delay is 180".

 

[etchdog2000]

Ok.

Delay 10 +10% haste gives you a delay of 9

10*.10 = 9

9 second delay * 6 hits= 54 seconds

Yep you guessed it. 54 seconds you have the same amount of swings as somone with 10 delay 0% haste in a 1 minute fight. and 6 seconds isnt enough to do jack shit except wait for your next hit that wont exist.

Simple math, simple equations, simple facts.

And seriously how old are you? Have you ever had a big boy discussion before? Or is your form of discussion calling people dumb asses for things you obviously dont understand.

I am sorry you dont get it....i really am.

 

[Klank]

The point is with more haste your hits come out faster with haste, thus killing the monsters faster even though you're getting in an extra swing. 10% haste means you do swing 10% faster. I think it would be beneficial to use a little haste on DRK because it does boost attack speed a little, and 20 attack isn't really that noticable. At least it's not in my experience with WAR.

 

[etchdog2000]

And btw, you are a complete idiot if you dont think I know that after you get a certain amount of acc it tapers off. Yeh pretty much everyone knows that, so lets assume that every knows that.

Dont tell me you think you really are telling me anything that any person with a high school education couldnt figure out.

You focus on all the wrong points I bring to the table, call me a dumb ass, when you cant even figure out a simple equation, that any tom dick or harry can figure out. You act like little arrogant kid.

 

[Uzor]

And btw, you are a complete idiot if you dont think I know that after you get a certain amount of acc it tapers off. Yeh pretty much everyone knows that, so lets assume that every knows that.

Dont tell me you think you really are telling me anything that any person with a high school education couldnt figure out.

You focus on all the wrong points I bring to the table, call me a dumb ass, when you cant even figure out a simple equation, that any tom dick or harry can figure out. You act like little arrogant kid.

I didn't call you a dumbass once in this thread. I will, however, say that you're wrong about the way you think haste and accuracy works.

And you skip my points completely. You completely ignored my mathematical proof about how even 5% haste will get you 1 more hit in a 50 sec fight.

 

[Marajade]

PSST. haste is about TP gain.

Oh noez, you loose 10 points of damage per swing. All you get to do is make a 1000 point WS 20% faster.

Put haste and acc on together where you can. A La Swift Belt, Homam hands/feet, and yes, switch your adaman Celeata for a Turban in TP.
Then switch to all your super STR/Attack gear, and kill somthing.
Repeat.

 

[Bajaratt]

Ok.

Delay 10 +10% haste gives you a delay of 9

10*.10 = 9

9 second delay * 6 hits= 54 seconds

Yep you guessed it. 54 seconds you have the same amount of swings as somone with 10 delay 0% haste in a 1 minute fight. and 6 seconds isnt enough to do jack shit except wait for your next hit that wont exist.

Simple math, simple equations, simple facts.

And seriously how old are you? Have you ever had a big boy discussion before? Or is your form of discussion calling people dumb asses for things you obviously dont understand.

I am sorry you dont get it....i really am.

What you're ignoring is that killing 10% faster actually matters. Let's say it takes 6 hits to kill a mob, with 10% haste, you get those swings 10% faster, which means you can kill more mobs in the same amount of time, which means more exp per hour which is therefore BETTER.

Using your numbers just for simplicity, 60 second fights, so 60 fights per hour. Let's forget all about exp chains and say 100 exp per kill, so 6,000 exp per hour. With 10% haste, you get those 6 hits in 54 seconds, resulting in 54 second kills rather than the 60 second kills without haste. Every kill nets you 6 more seconds to fight another mob, which you kill in 54 seconds as well, and get another 6 seconds to fight another mob.

Now, instead of 6,000 an hour, you're doing around 6,600 an hour. The difference is even greater when you throw exp chains into the mix, as you can hit higher chains more often by killing faster. Haste just increased your exp/hour, and you can get more exp when you go beyond 10% haste, just that's a relatively easy amount to get.

 

[Lavi]

OK, Scythe delays vary from 480 to 528 so I'll do calculations for both of them.

60 Delay = 1 second
480 Delay = 8 Seconds
528 Delay = 8.8 Seconds

In a 60 second fight, a 480 delay weapon will hit 7.5 times (60 seconds/8 seconds per swing) or 7 times since you cannot hit for half a hit and the half swing is wasted. For a 528 delay weapon, you'll have 6.82 swings (60 seconds/8.8 seconds per swing) or 6 full swings and .82 wasted.

Now, what amount of haste can improve those hits for me? Well, since the delays are different, we'll get slightly different numbers for both. Given the obvious nature, we'll need more haste for the higher delay weapon to get a full swing, but since we only need .18 swing for the high delay weapon and .5 for the lower delay, it's not so clear cut. Now, I want a full extra swing so I am only going to add enough haste in order to make full swings with no wasted delay (in other words, swings will be in integers of 8 and 7 respectively, adding more will not be a bad idea due to engage times, but skipping out on some would be detrimental).

8 Swings in 60 seconds:
7.5 Seconds per swing or 450 Delay
7 Swings in 60 seconds:
8.57 Seconds per swing or 514 Delay

In order to get each weapon an extra full swing in 60 seconds, I need to reduce a 480 delay weapon to 450 delay and a 528 delay weapon to 514 delay.

480 Delay Weapon:
[ (480 - 450)/480 ] * 100 = 6.25% Haste
528 Delay Weapon:
[ (528 - 514)/528 ] * 100 = 2.65% Haste

There you go. You now have a full swing in 60 seconds like you wanted. Now, to be honest, I never crunched your original numbers and I am wishing I had because they are wrong as far as I can tell. You only need a 6-7% haste max to get the full extra swing in per 60 seconds. I have a good idea how you messed up though. You deduced that you needed X percent haste to reduce your delay enough to get a full swing.

So your math was probably something like this:

6 Swings * 528 Delay = 3168
3168/7 Swings = 453
[ (528 - 454)/528 ] * 100 = 14.00% Haste

The only flaw is you forgot the partial swings in 60 second fights. Those are wasted swings that occur most likely because someone else was able to kill it before you landed your full swing.

 

[Rappa]

There's too much wrong with this post for it to be valid. For starters, it is assuming that the scythe-wielder has no haste spell, no march, etc, which for all intents and purposes isn't going to happen. It doesn't take into effect double attack, time used to use job abilities and weapon skills, etc. There are also a lot of other factors to consider when you are looking into haste gear, such as tp return (will this piece of gear require me to hit an extra time for 100% tp?) and other stat tradeoffs (losing atk for haste, for example). Haste is not the end-all, be-all stat to pile onto every job.

Honestly, your best bet is to go out and experiment on your own. Find out what each piece of haste gear you put on does for you in terms of speed, tp return, etc. Find the correct balance of atk, haste, and acc that works for you. Otherwise you're going to spend countless months acquiring 1337 gear only to find out it isn't as 1337 as you thought it would be. I would suggest sticking to the easiest items to acquire in terms of haste (w.turban), atk (black cuisses), and acc (thick/adaman, etcetc) and then just experimenting with new gear as you acquire it. These "fantasy builds" often end up highly lacking, and piling on haste at the expense of acc or atk will just result in weaker attacks missing more often and giving back crap tp. Screw math, just play around with what you've got.

 

[Lavi]

For more on the subject, let's do the math for the benefits to damage of haste. Let's just look at a 480 Delay weapon since looking at a 528 would be skewed even more so in my favor. Now we've gained 1 extra hit on top of the 7 previous hits that were already there. Let's say, you hit 80% accuracy with haste gear and with accuracy/attack gear, you are hitting 85%. Seems like a fair and realistic amount to me, but if you want to crunch different numbers, go ahead.

Odds with accuracy gear:
7 hits: .85 ^ 7 = .32 or 32%
6 hits, 1 miss: 39.6%
5 hits, 2 misses: 21.0%
4 hits, 3 misses: 6.2%
Everything else: ~1.2%

Odds with haste gear:
8 hits: 16.8%
7 hits, 1 miss: 33.6%
6 hits, 2 misses: 29.3%
5 hits, 3 misses: 14.7%
4 hits, 4 misses: 4.6%
Everything else: ~1%

Binomial Mass Distribution Function (used to calculate odds above):
{ n!/[ x! * (n - X)]} * [ p^x *(1 - p)^(n-x)]

Even at a 5% loss in accuracy, you can still make up for that loss by attacking quicker. To demonstrate, hitting 5 or more times with accuracy gear, I get a total of 92.6% probability, but with haste gear, I have a 94.4% probability. Even at a loss of accuracy, with an extra hit, the odds work in my favor. In order for haste gear to be at a severe disadvantage, you need a large loss in accuracy, probably around 10%. If you're eatting Sushi, it's probably not an issue especially while in a merit party fighting VTs.

There's too much wrong with this post for it to be valid. For starters, it is assuming that the scythe-wielder has no haste spell, no march, etc, which for all intents and purposes isn't going to happen. It doesn't take into effect double attack, time used to use job abilities and weapon skills, etc. There are also a lot of other factors to consider when you are looking into haste gear, such as tp return (will this piece of gear require me to hit an extra time for 100% tp?) and other stat tradeoffs (losing atk for haste, for example). Haste is not the end-all, be-all stat to pile onto every job.

Honestly, your best bet is to go out and experiment on your own. Find out what each piece of haste gear you put on does for you in terms of speed, tp return, etc. Find the correct balance of atk, haste, and acc that works for you. Otherwise you're going to spend countless months acquiring 1337 gear only to find out it isn't as 1337 as you thought it would be. I would suggest sticking to the easiest items to acquire in terms of haste (w.turban), atk (black cuisses), and acc (thick/adaman, etcetc) and then just experimenting with new gear as you acquire it. These "fantasy builds" often end up highly lacking, and piling on haste at the expense of acc or atk will just result in weaker attacks missing more often and giving back crap tp. Screw math, just play around with what you've got.

This is the best solution; however, ignoring haste all together because you think you need huge amounts to make an effect when you only need minor amounts is ignorant at best. While I do agree, parsing will get you the best results, sometimes a little math will send you in the right direction especially when the only way to parse specific results is to buy multimillion gil items to test with when it's very possible you may waste the gil on an item you can't get rid of and no longer want (EX items, items that rarely sell).

 

[metachronism]

WTF are people trying to even argue for in this thread?

uzor: do you have the acc/eva formula? unless you do, no one gives a shit about your "it isn't linear" spiel. When you're hitting mid-80%s and over, 1 acc rating does translate to roughly 0.5% more raw accuracy on T-VTs. The actual formula is probably logarithmic over the whole range, but it's impossible to test that while keeping the pc/mob level constant, unless you create some level 75 character with 0 weapon skill just for testing purposes.

On T-VT mobs (merit mobs), the acc rating/raw acc relationship becomes asymptotically linear (roughly 2:1 ratio), and then, reaches a hard cap at 95%.

You have to keep in mind that haste builds are mainly for people who are NEARING THE 95% CAP. If you are nowhere near that, you are usually better off boosting acc. nin/war/mnk/lolthf can get away with less acc and more haste because they can still generate a steady TP inflow from just the sheer number of attacks. That is not the case with the "ace" jobs.

For your reference, to hit 95%, the target acc ratings are:

Uleguerand (war/drk demons): 420
Uleguerand (blm/smn demons): 390
Lufaise (rams/abraxas): 415
Decorative Weapons: 380
Weak shit (puks and stuff): 360-380ish

And for those who don't know, the acc rating is calculated by:

200 + (skill - 200) * 0.9 + dex/2 + acc modifiers from eq

Sushi will add 15-18% (different types give different amounts) to this value.

The reason one would take homam over dusk would be to eat meat instead of sushi, but unfortunately, with your build, you won't hit the 95% cap without sushi on the stronger stuff (in fact, I'll tell you right now it's nearly impossible to hit 420 without sushi unless you make compromises). So, I would suggest that you go for dusk rather than homam for hands/feet to get more atk where you can, and assume you eat sushi.

 

[Lavi]

WTF are people trying to even argue for in this thread?

uzor: do you have the acc/eva formula? unless you do, no one gives a shit about your "it isn't linear" spiel. When you're hitting mid-80%s and over, 1 acc rating does translate to roughly 0.5% more raw accuracy on T-VTs. The actual formula is probably logarithmic over the whole range, but it's impossible to test that while keeping the pc/mob level constant, unless you create some level 75 character with 0 weapon skill just for testing purposes.

On T-VT mobs (merit mobs), the acc rating/raw acc relationship becomes asymptotically linear (roughly 2:1 ratio), and then, reaches a hard cap at 95%.

You have to keep in mind that haste builds are mainly for people who are NEARING THE 95% CAP. If you are nowhere near that, you are usually better off boosting acc. nin/war/mnk/lolthf can get away with less acc and more haste because they can still generate a steady TP inflow from just the sheer number of attacks. That is not the case with the "ace" jobs.

For your reference, to hit 95%, the target acc ratings are:

Uleguerand (war/drk demons): 420
Uleguerand (blm/smn demons): 390
Lufaise (rams/abraxas): 415
Decorative Weapons: 380
Weak shit (puks and stuff): 360-380ish

And for those who don't know, the acc rating is calculated by:

200 + (skill - 200) * 0.9 + dex/2 + acc modifiers from eq

Sushi will add 15-18% (different types give different amounts) to this value.

The reason one would take homam over dusk would be to eat meat instead of sushi, but unfortunately, with your build, you won't hit the 95% cap without sushi on the stronger stuff (in fact, I'll tell you right now it's nearly impossible to hit 420 without sushi unless you make compromises). So, I would suggest that you go for dusk rather than homam for hands/feet to get more atk where you can, and assume you eat sushi.

Simply not true. While it would be nice to get to the 95% cap AND get a haste build, you don't need to. Mathematically, if you are hitting more frequently, which with lower delay weapons happens a lot easier than higher delay, you can make up for accuracy loss with more attack swings. Think of it like a machine gun and a pistol. While someone with a pistol needs to take care to hit a target, I can spam a million shots and hopefully one hits the target. Say I had 100% acc, I then instead remove a bunch of accuracy gear and replace it with 50% haste so I hit twice as often. Now, if I have 50% accuracy, I can still hit as many times as I did without the accuracy (law of averages). Now this is an extreme example, but the math that I did with the binomial function actually demonstrates as much with a little more realistic numbers.

In the previous example, I have a 32% chance of all 7 hits landing in a single fight while in accuracy gear. With haste and at the loss of accuracy, I now have a 33.6% chance to hit 7 times in a single fight AND a 16.8% chance to hit 8 times. Now tell me, which is the obvious superior? Granted, those numbers are fake, but they can be easily replicated with a little effort by any melee given a little effort. When I get 75, I will do just that (if I can get a decent haste build).

 

[metachronism]

Simply not true. While it would be nice to get to the 95% cap AND get a haste build, you don't need to. Mathematically, if you are hitting more frequently, which with lower delay weapons happens a lot easier than higher delay, you can make up for accuracy loss with more attack swings. Think of it like a machine gun and a pistol. While someone with a pistol needs to take care to hit a target, I can spam a million shots and hopefully one hits the target. Say I had 100% acc, I then instead remove a bunch of accuracy gear and replace it with 50% haste so I hit twice as often. Now, if I have 50% accuracy, I can still hit as many times as I did without the accuracy (law of averages). Now this is an extreme example, but the math that I did with the binomial function actually demonstrates as much with a little more realistic numbers.

In the previous example, I have a 32% chance of all 7 hits landing in a single fight while in accuracy gear. With haste and at the loss of accuracy, I now have a 33.6% chance to hit 7 times in a single fight AND a 16.8% chance to hit 8 times. Now tell me, which is the obvious superior? Granted, those numbers are fake, but they can be easily replicated with a little effort by any melee given a little effort. When I get 75, I will do just that (if I can get a decent haste build).

And you claim that haste and accuracy come at equal opportunity costs? Yes, you have a point IF haste and accuracy gear are interchangeable at equal costs. It is clearly not the case, both mechanically and financially.

Mathematical models are meaningful only when you have the premises correct. It makes no sense to interchange 14% haste for 5% accuracy, as they do not come at equal costs (which is what your model assumes). In terms of equipment, your model is the equivalent of changing your belt from a life belt to a speed belt, and squeezing out 8% more haste out of thin air. OF COURSE you'd do more damage with that kind of an exchange...

Yeah, if we had speed belt equivalents in every equipment slot that was cheap, your argument would be valid, but I'm afraid that's not the case.

 

[Mishaela]

200 + (skill - 200) * 0.9 + dex/2 + acc modifiers from eq

Just curious, as I'm bad at math so I might be totally wrong, but bolded a part to ask...what's the point of this? The two "200"s totally negate eachother nomatter what your skill is.

*Edit* Okay, yeah, figured it out, skill - 200 * 0.9 first. Nevermind >_<

 

[metachronism]

It's just a roundabout way of saying, skill < 200 gives you 1 acc per point, but skill beyond that gives only 0.9 acc per point.

 

[Rappa]

Just curious, as I'm bad at math so I might be totally wrong

Always remember: Please Excuse My Dear Aunt Sally!

 

[Uzor]

200 + (skill - 200) * 0.9 + dex/2 + acc modifiers from eq

Without proof, you can't say that this is the accuracy formula. So I ask you, where did you find this?

It's just a roundabout way of saying, skill < 200 gives you 1 acc per point, but skill beyond that gives only 0.9 acc per point.

Your formula is saying that all points, uncluding those before 200 skill, gives 0.9 points

Either way, it doesn't change the fact that haste builds are more usefull then either of you give credit for. While yes, I do agree that you shouldn't give up all accuracy and attack for haste gear as DRK, SAM or other hardhitting jobs, but as I already proved earlier (and Lavi did too), even 5% haste (which is very east to get) gives any job an extra attack in before the mob dies during a 50 sec fight assuming he uses no spells and weapon skills.

 

[Mai Ikari]

Without proof, you can't say that this is the accuracy formula. So I ask you, where did you find this?

Your formula is saying that all points, uncluding those before 200 skill, gives 0.9 points

It may not be saying much, but it's the generally accepted description of ACC. All skill until 200 = 1 ACC per. Over 200 = .9 ACC. ACC gear = 1 ACC. 1 DEX = .5 ACC.

And, as Rappa said, Please Excuse My Dear Aunt Sally!

200 + (remaining skill*.9) + (DEX/2) + ACC gear



As for the main topic... Haste is good, haste makes kills faster, even if you won't parse a single point of damage higher. Assuming you can hit in the first place.