Forums › English Language Forums › General › The Arsenal

Buscar

For all you DPS number-crunchers out there...

5 respuestas [Último envío]
Jue, 08/16/2012 - 12:27
Imagen de Raisinfist
Raisinfist

Hi all,
Not sure where to post this, so please tell me where to move if necessary, but...

So I know many people place a high priority on DPS when choosing a new weapon (which seems obvious) but I think it's possible to be misled by the DPS figure for many weapons.

For example, I find that, during normal PvE play, I rarely use the final strike of the 3-strike combo in Brandish upgrades (such as Shockburst Brandish ) and I often use only the first combo strike for Sealed Sword and Troika lines, except when I have plenty of time on my hands in front of an immobilised opponent. This is because letting the combo run to completion leaves me vulnerable to attacks before I can shield up in normal battle situations versus lichen colonies, for example. So it strikes me that, in a normal battle situation, although Shockburst will do less damage per hit than Sealed Sword, I will get twice as many hits in the same amount of time, since I will be forced to shield to preempt enemy attacks.

Now suppose I take an imaginary Troika line weapon and an imaginary Brandish line sword with identical DPS figures. The DPS is normally calculated by the damage done per combo divided by the time taken. But in normal combat, I will only get a chance to use (slightly less than*) half the total available damage available for the Troika combo, whereas for the Brandish I will get to make use of (slightly less than*) two thirds of the combo damage pool. So for me, the Brandish sword is the obvious option (although you can argue benefits of say knockback with Sealed Sword and similar).

Now for the radical bit - I propose an alternative damage measure method to account for this. Of course this could not replace DPS, since DPS is still valid for situations vs. an immobilised/incapacitated enemy like an attack to the rear of a Lumber-type monster. You could do a calculation based on DPS of all strikes bar the final in the combo, which would be more helpful for PvE. However I have recently been toying with Cutter-line swords, and I notice you can't get four out of five combo strikes on most enemies without incurring damage in return. So I suggest that rather than making the length of the combo the base unit for DPS, we instead calculate the damage that can be done in the first second of the combo, allowing for a shield block before a monster has time to react.

I know I've used sword and shield examples here, but a similar thing applies to guns - I often run to the penultimate shot on my guns to avoid being vulnerable in the reload animation, since you can't shield whilst reloading (a bizarre notion, but necessarily implemented by OOO to avoid OP'ing guns). This makes my Damage Per Combo reduced by a half-combo for Alchemer line guns, but only by one sixth for Antigua lines.

So, thoughts? ideas? questions?
Cheers,

R.

P.s. Maybe you could call this alternative indicator DinFS (Damage in First Second, since DFS has too many sofa-related connotations [;)

(*Of course, the final combo hit of any sword does more damage than other combo hits - another fact that skews DPS figures.)

Jue, 08/16/2012 - 12:56
#1
Imagen de Bopp
Bopp
yet another idea

Nobody really believes that DPS-from-comboing is the only way to measure damage output, or even the only way to measure DPS. More kinds of measurement would be great.

I don't speak for Lancer Knightz, but I think that they started with DPS-from-comboing for swords and guns just because it seemed like a very clear-cut measure, that didn't require any skill to achieve reproducible results.

That guild has started on a new way of measuring sword damage output: almost-combos per minute. (An almost-combo is a combo that omits the final swing.) That's not identical to what you're suggesting, but it's similar in its motivation. Getting reproducible results is difficult, though. The first few times I tried such a test, I found my numbers steadily climbing, as I got better and better at it.

In your idea to measure damage in the first second, how are you going to handle the fact that maybe 1.5 strokes fit into the first second of time? Are you going to count only complete strokes? That will make your results highly sensitive to the unit of time that you've chosen.

Vie, 08/17/2012 - 07:13
#2
Imagen de Raisinfist
Raisinfist
1 second?

Bear in mind, this is entirely Cann-Ed-Fodder's idea. I'm just maintaining the thread.

@Bopp
From what I understand from his explanation of this, the strikes that can be put into 1 second are the strikes that can be put in before most monsters attack. Since you cannot stop halfway through strikes, I assume that this would be 'number of strikes that can start within 1 second'. Also, the KB from the (2nd) strike would give you a little time in which to finish the swing and shield.
An 'almost-combo' wouldn't work for fast swords such as the Spur, Cutter or Flourish-esque lines, hence the suggestion of hits within one second - 4 hits for a Cutter go on for more than one second - so the almost-combo may not exactly fit the bill.

May I suggest, by the way, for the almost-combos, hitting the attack button until the animation for the second swing begins? This should only result in two swings, but I haven't checked that.

Vie, 08/17/2012 - 07:27
#3
Imagen de Bopp
Bopp
yep

Bear in mind, this is entirely Cann-Ed-Fodder's idea. I'm just maintaining the thread.

So far, I think it's a fine idea. I want you to jump in, get some data, discover the best way to run the tests, and so forth. Get cracking! :)

May I suggest, by the way, for the almost-combos, hitting the attack button until the animation for the second swing begins? This should only result in two swings, but I haven't checked that.

You mean, hit Attack for the first swing, and then hit Attack a second time, but holding it down only until the animation just barely begins? I don't think about it that way, but it's probably close to what I do. Probably it's what Lancer Knightz does.

Another issue is that you want to shield-cancel to cut animations as short as possible. Doing this all at high speed for a solid minute is tiring. If you mess up then you have to start over. And then you need to repeat the whole test a couple of times, to make sure you're getting a stable answer. Collecting high-quality data is not a walk in the park.

Mar, 08/21/2012 - 10:42
#4
Imagen de Havenihaveaproblem
Havenihaveaproblem
Would there be anything wrong

Would there be anything wrong with running the test at 20-30 seconds? I tried to reproduce some of the Lanzer Knights data myself once, and I noticed that my combos per 20 seconds were proportional to the combos per minute number.

You also have to determine how to incorporate UVs. Partial combos, for example, might be more relevant to a player who has no AS UVs as opposed to one with AS max! But any data would be useful as long as the UVs are consistent.

For me, if there are a group of enemies, you are probably better off using a charge attack than darting in for one or two swings anyways. But I'm not the authority on swordplay.

Mar, 08/21/2012 - 12:32
#5
Imagen de Raisinfist
Raisinfist
@HavenIhaveaproblem

1. YES, there IS a problem with running at damage per 20-30s:
- 1. The DPS we are going to calculate will, most times, omit the final swing (which does extra damage). In 20-30s, there are a lot of last swings. Damage per 30s / 30 =/= damage per 1 second.

2. We aim to do any data with NO UVs or buffs. Evidently somebody with a lot of ASI could pull off a full combo in less time, thus making incomplete combo data almost worthless. However, many of the mainstream weapons already have average full combo DPS data, so they could use that (ie, not the D. P. 1s.) for their choices.

Thanks for replying!

Sitio elaborado con Drupal, un sistema de gestión de contenido de código abierto