• Welcome to Smashboards, the world's largest Super Smash Brothers community! Over 250,000 Smash Bros. fans from around the world have come to discuss these great games in over 19 million posts!

    You are currently viewing our boards as a visitor. Click here to sign up right now and start on your path in the Smash community!

Knockback Formula & Values

Status
Not open for further replies.

SCOTU

Smash Hero
Joined
Mar 16, 2007
Messages
6,636
Location
Northville, MI
ColinJF Model

This research was completed by Amazing Ampharos and I (Colin) over the last few days.

1.How Damage is Stored
2. A Note on Methods
3.The Damage Formula
4. A Note on Training Mode
5.The Knockback Formula
6. Further Research

Abstract

In this post, we present an accurate damage formula and an accurate knockback formula. In particular, we develop new results in the areas of how damage is stored, stale move decay, Lucario's aura power, pokemon fatigue, charging smashes, and more. We develop exact models for many of these things that predict the precise damage, including decimal places. We also present a correct formula for knockback.

1. How Damage is Stored

At least three decimal places of precision are required to explain the exact results we obtained using our damage formula. Whether damage is actually a decimal number is difficult to determine, but whatever it is, it is precise enough to store the equivalent of three decimal places. Contrary to some previous reports, the maximum damage is in fact 999%. Damage over 999% is discarded.The game truncates damage for the purposes of display. For example, 125.651% is displayed as 125%.

2. A Note on Methods

To determine the actual damage a move was dealing, we used it repeatedly and recorded the sequence of numbers after each use. Using a program I wrote, we were able to determine the actual damage the move was dealing, to full precision. At first, we had to work within training mode to avoid stale moves, but once we cracked stale moves precisely, we were able to account for these in the sequence and determine the exact damage of moves directly from sequences of numbers in versus mode.

The source for this program is provided here in case other people want to do further research involving decimal damage:

Code:
import java.math.*;

/**
 *
 * @author Colin
 */
public class BrawlDamage {

    private static final double[] stale = {
      1.05, 0.9, 0.81, 0.73, 0.66, 0.60, 0.55, 0.51, 0.48, 0.46
    };
    
    private static final int[] data = {
        /** Place sequence of numbers here. **/
    };
    
    /** Change to false for training mode. */
    private static final boolean versusMode = true;
    
    public static BigDecimal[] getInterval(int[] data, 
            BigDecimal lower, BigDecimal upper,
            int precision, boolean versus) {
        BigDecimal x = lower;
        BigDecimal y = upper;
        BigDecimal step = new BigDecimal(1);
        for (int i = 0; i < precision; ++i) {
            step = step.divide(BigDecimal.TEN);
        }
        for (; x.compareTo(y) <= 0; x = x.add(step)) {
            boolean works = true;
            BigDecimal sum = new BigDecimal(0);
            for (int i = 0; i < data.length; ++i) {
                double factor = 1.0;
                if (versus) {
                    factor = stale[(i < 9) ? i : 9];
                }
                BigDecimal addend = x.multiply(new BigDecimal(factor));
                sum = sum.add(addend);
                int j = sum.round(
                        new MathContext(0, RoundingMode.FLOOR)).intValue();
                if (j != data[i]) {
                    works = false;
                    if (j > data[i]) {
                        if (upper.compareTo(x) > 0) {
                            upper = x;
                        }
                    } else {
                        if (lower.compareTo(x) < 0) {
                            lower = x;
                        }
                    }
                    break;
                }
            }
            if (works) {
                lower = upper = x;
                System.out.println("x = " + x.toString() + " works.");
            }
        }
        return new BigDecimal[] { lower, upper };
    }
    
    public static void main(String[] args) {
        BigDecimal x = new BigDecimal(data[0]);
        BigDecimal y = new BigDecimal(data[0] + 1);
        for (int i = 2; i < 10; ++i) {
            BigDecimal[] interval = getInterval(data, x, y, i, versusMode);
            x = interval[0];
            y = interval[1];
            if (x.compareTo(y) == 0) {
                break;
            }
        }
        System.out.println("x lies in (" + x + ", " + y + ").");
    }

}
3. The Damage Formula

This formula gives you the exact damage that a move will do in any mode of the game. This formula works for both training mode and versus mode.

damage = base damage × stale multiplier × charge multiplier × aura multiplier × fatigue multiplier

We explain the value of each of these multipliers in the following sections.

3.1 Base Damage

Every move has some integer damage associated with it. This is its base damage. However, most of the time, the actual damage dealt by a move is not an integer, but this is because of multiplicative factors in the damage formula.

3.2 Stale Move Multiplier

The formula for the stale move multiplier is



This is explained below, including the definition of s.

In training mode, stale moves are not in play, so the value of this multiplier is exactly1. If the move is not affected by stale move decay, then the value of this multiplier is 1 even in versus mode. The following moves are not affected by stale move decay: Luigi's down taunt; DK's cargo throws; tether attacks; Zero Suit Samus's neutral air.

In versus mode, for moves affected by stale move decay, this factor is never 1. If a move is not in the stale move queue before being used, the value of this multiplier is exactly 1.05.

In versus mode, for moves affected by stale move decay, a move is placed in the stale move queue after it is used. The stale move queue contains nine entries, and each entry has a numerical weighting as depicted in the following diagram.



The left of the diagram represents the most recent move to be added to the queue, and the far right represents the oldest move to the added to the queue. Then we have

stale multiplier = 1 – s

where s is the sum of the values of the positions in the queue occupied by the move being used. For example, if the move being used occupies the .1 position, the .06 position, and the .03 position, then the stale multiplier is 1 - (0.1 + 0.06 + 0.03) = 0.81.

3.3 Charge Multiplier

This is the formula for the charge multiplier



For partially charged smash attacks, the value will be somewhere between 1 and 1.4, but we were unable to determine the exact distribution experimentally; it is too difficult to test. However, we speculate that it follows the same distribution as the aura damage multiplier. We speculate that if you charge the smash 1/n of the way, the multiplier is the same as the aura damage multiplier for (170-75)/n+75 %.

3.4 Aura Multiplier

This is the formula for the aura multiplier:



where



In coin mode, each coin counts as a stock for the purpose of the stock multiplier.

The damage multiplier is in the interval [0.7, 1.4]. It is 0.7 for all damages <= 20% and 1.4 for all damages >= 170%. It is exactly 1 only when the damage is 75%. A very good approximation for the damage multiplier can be found by reading the value from this chart:



For example, the damage multiplier for 95% is 1.09.

This chart was created by using quintic interpolation between some exact values that we worked out.

3.5 Fatigue Multiplier

Not everything is known about this multiplier. However, it is 1 if the character is not a pokemon trainer pokemon. Otherwise, it is in the interval [0.7, 1] where 1 is fully unfatigued and 0.7 is fully fatigued. There is in fact a spectrum of fatigue levels and the multiplier can take on various values in between 0.7 and 1. The multiplier cannot take on every value in this interval, however. When the pokemon shows no signs of fatigue, the multiplier is 1.

4. A Note on Training Mode

Our damage formula explains one aspect of training mode that was previously a mystery. In training mode, the stale move multiplier is 1, since stale move decay is not in play. As a result, most moves do their integer base damage always. In versus mode, however, the stale move multiplier is never 1 for moves that respect stale move decay. As a result, moves in versus mode do not usually do integer damage. This is the cause of the inconsistencies noticed between training and versus mode.

5. Knockback

The starting place for our research on Knockback was MrSilver's “Character weight list, fall speed list and random things” thread. MrSilver's work is a good starting place, but is incorrect in several ways.

Throughout this section, the words “launch speed” and “knockback” are used interchangeably because the quantity that corresponds to knockback on the results screen is the “max launch(er) speed”.

5.1 Constants Associated with Characters

Firstly, for the purpose of calculation of knockback, every character has two associated constants, namely weight1 (w1) and weight2 (w2). MrSilver determined both of these constants for every character, but mysteriously referred to weight2 as “fall speed”. In fact, weight2 has nothing to do with fall speed. Here is the list of weight1 and weight2 for each character. (These values are copied from MrSilver's document and included for ease of reference with the following formula.) Weight1 is the first value and weight2 is the second value

Bowser: 1.111, 0.91
Donkey Kong: 1.091, 0.20
Snake: 1.076, 0.16
King Dedede: 1.071, 0.29
Charizard: 1.061, 0.33
Ganondorf 1.056, 0.91
Samus: 1.051, -0.54
Yoshi: 1.045, -0.05
Wario: 1.045, 0.30
R.O.B.: 1.040, -0.32
Ike: 1.035, 0.28
Captain Falcon: 1.030, 0.92
Link: 1.030, 0.46
Wolf: 1.020, 1.84
Lucario: 1.010, -0.40
Ivysaur: 1.010, -0.16
Mario: 1, 0.00
Luigi: 0.995, -0.32
Sonic: 0.985, 0.33
Ness: 0.980, -0.12
Pit: 0.980, 0.00
Lucas: 0.980, 0.49
Diddy Kong: 0.975, 0.55
Toon Link: 0.970, -0.16
Ice Climbers: 0.970, 0.07
Peach: 0.960, -0.37
Marth: 0.944, -0.25
Sheik: 0.935, 2.21
Zelda: 0.935, -0.26
Falco: 0.919, 1.23
Olimar: 0.919, -0.47
Zero Suit Samus: 0.914, 0.44
Fox: 0.910, 3.41
Pikachu: 0.904, 0.40
Meta Knight: 0.904, 0.68
Kirby: 0.899, -0.44
Squirtle: 0.884, 1.05
Mr. Game & Watch: 0.884, 0.01
Jigglypuff: 0.849, -0.78

5.2 Constants Associated with Moves

Every move has a base knockback constant (b) and most moves have a knockback growth constant (g). For moves with fixed knockback, the base knockback constant is defined as the knockback taken by Mario. For moves with variable knockback, the knockback growth constant is defined as

g = (f - m)/e

where m is the knockback taken by Mario at 0% before the hit, e is any damage you like (say, 100), and f is the knockback taken by Mario at e% before the hit.

Then the base knockback constant is defined as

b = m – dg

where m is the knockback taken by Mario at 0% before the hit, d is the damage the move deals, and g is the knockback growth constant.

Moves with fixed knockback have have two weight variability constants (s1, s2), which describe how much the knockback of the move varies with weight1.

All sufficiently strong moves have a weight2 scale constant (c) which describes how important weight2 is to the move. For most moves, c = 100. For some moves, such as Ness's up throw, c is over 160.

5.3 Knockback Equation

We determined that the knockback equation has two cases depending on the value of a constant k.

To calculate the knockback, first work out k.



where b is the base knockback constant, d is the damage the move would deal, x is the damage of the target before the hit, g is the knockback growth constant, w1 is the weight1 of the target, and s1, s2 are the weight variability constants.

Then we have the equation for knockback:



where w2 is the weight2 of the target, and c is the weight2 scale constant.

5.4 Additional Multipliers for Knockback

We found a few things that cause simple multiplications of the final knockback value. A move being “super effective” is a multiplier of 1.1, and a move being “not very effective” is a multiplier of 0.9. Obviously, this only affects the Pokemon Trainer. In fast mode special brawl, all knockback is multiplied by 1.5, and in slow mode special brawl, all knockback is multiplied by 0.5. Heavy mode special brawl does not affect knockback. As was previously known, attacking a target who is charging a smash other than Ness's up and down smashes is a multiplier of 1.2.

6. Further Research

It is not known how stale move decay affects knockback precisely. Another area for further research is how charging smash attacks affects knockback.


Colin and Amazing Ampharos
 

MuBa

Smash Lord
Joined
Nov 4, 2005
Messages
1,958
Location
Dragon Kick you into the Milky Way!
You know leafgreen has been doing a lot of research on knockback and gravity. I will have to tell SamuraiPanda to give him an invite too. He will prove to be a GREAT asset on this research.
 

SCOTU

Smash Hero
Joined
Mar 16, 2007
Messages
6,636
Location
Northville, MI
Can I get someone impartial, who has good analytical skills, amazing numerical analysis skills, good chart making skills, and a history/ reputation for number crunching/comparisons please compare these two models and draw up some charts that shows which is more accurate, and in which cases it is moreso?
 

Pierce7d

Wise Hermit
Joined
Dec 20, 2006
Messages
6,289
Location
Teaneck, North Bergen County, NJ, USA
3DS FC
1993-9028-0439
I know I haven't said much, which is due to the fact that I've been trying to study all of this data so that I don't just blabber on uselessly, and can actually contribute, make sense and help.

I did want to mention though, that in ColinJF's Model, section 3.3: Charge Multiplier, I was under the impression that there are only 3 levels of charge. These levels are, uncharged, partially charged, and fully charged. I will test this, and get back to you.

Also, I know these charts need to be tested. I'll try to get it done then. If not tonight, then tomorrow hopefully.
 

ColinJF

Smash Ace
Joined
Dec 21, 2007
Messages
712
This might be a good place to mention a few updates to the original model, since we first posted it.

Firstly, I made a mistake using the term "knockback" to refer to "launch speed". They are not the same thing, and I suggest we use the correct terms from now on. Knockback is, however, related to launch speed, but involves other factors. For example, for vertical knockback, launch speed versus average fall speed is linear for a fixed amount of knockback.

Secondly, the mysterious constant "c" seems to be a function of g and the launch angle of the move, although this doesn't affect the actual calculation of c.

Thirdly, w2 is nothing other than a - 1 where a is vertical acceleration.

I've been using this script to work out g, b, c for arbitrary moves, but it's easy to do it by hand as well: http://colin.shoddybattle.com:81/gbc.php

Also, I've practically applied our work on launch speed & my more recent work on knockback numerous times recently, and so far everything has been consistent with the model.
 

SCOTU

Smash Hero
Joined
Mar 16, 2007
Messages
6,636
Location
Northville, MI
Also, Ankoku and I have discussed calling the "weights", "launch resistance" since that's what its more akin to.
 

ColinJF

Smash Ace
Joined
Dec 21, 2007
Messages
712
That wouldn't be a bad idea, but we'd have to careful to avoid saying "resistance to knockback", since several things help you resist knockback (e.g. fall speed) that aren't weight. There shouldn't be a problem with "launch resistance" though, aside from making it initially catch on, since "weight" is very widespread.
 

infomon

Smash Scientist
Joined
Mar 11, 2008
Messages
5,559
Location
Toronto, Canada
Hmmmm, so the weight1 and weight2 constants are used for knockback calculations, but aren't related to the character's fall speed(s). That's very interesting. I know from the "momentum-cancelling" / using-aerials-for-survival stuff, that fastfalling an aerial out of hitstun helps counteract your vertical knockback. So at some point after you're hit, it would seem that your fall speed(s) become relevant to survival.

I guess that's one question we need to answer: is it the character's fastfall speed that helps reduce vertical knockback? Empirically, this seems to be the case; fastfalling knockback with Link and Fox is much more visually obvious than with floaty characters. What can we say about how much the fastfall helps? This is probably a simple motion equation, but I'm not sure of the mechanics.

This will make it relevant to collect fall speeds and fastfall speeds for every character. In addition, I wonder if there are characters with a "fall special" state whose speed does not match their regular aerial fall/fastfall speeds? Or is aerial mobility always the same before and after a recovery special?

Edit: Also, I suspect than an ordering of characters by fall speed will be different than an ordering of characters by fastfall speed. It would really be interesting to see such a chart; but I don't know how to measure fall/fastfall speeds exactly.
 

ColinJF

Smash Ace
Joined
Dec 21, 2007
Messages
712
As I said above, we made a mistake not using "launch speed" in the original document. Launch speed is a function of weight and fall speed acceleration. But knockback itself depends on other things. Vertical knockback is related to launch speed by average fall speed over the total distance, and horizontal knockback is related to launch speed by average horizontal air speed over the total distance.

w2 is fall speed acceleration by the way.

Anyway, I haven't written down everything I've figured out about knockback yet, but there really aren't any mysteries as far as what quantities are involved.
 

infomon

Smash Scientist
Joined
Mar 11, 2008
Messages
5,559
Location
Toronto, Canada
Ummmm.... ok, I'll have to read/think it over a bunch before I understand how/when fastfall speed factors into all that. Thanks for your reply, I might have more questions later.
 

ColinJF

Smash Ace
Joined
Dec 21, 2007
Messages
712
I can already tell you how fast fall speed is involved (in principle).

There are four "regions" of "falling" when you are knocked up: when you are accelerating to your top speed, travelling at your constant speed until your aerial finished, accelerating to your top fast fall speed (not such how long this takes though), travelling at your top fast fall speed. When you can calculate the average fall speed like this for a bunch of characters, you discover that average fall speed versus launch speed is linear for fixed s (knockback).

If there's no fast falling involved, average fall speed f[average] is given by

f[average] = s/(s/f - (a * (f/(2a))^2)/(2 * f) + f/(2a))

where s is the distance, f is the top fall speed, and a = w[2] + 1.

This formula becomes a bit tricker when you involve the aerial to fast fall, but it's the same principle.

I still don't have a formula for knockback in terms of fall speed, fall speed acceleration, and launch speed unfortunately. So, some analytical work remains, but I have a pretty good idea how vertical knockback works in principle. Basically I think when you are knocked up there are two forces that help to get you down: a sort of "air resistance" that opposes all motion by decreasing your speed by a bit each frame, and your fall speed (which changes with time).

Some of this is just speculation though... I'll write a more thorough exposition when I'm more sure. Until then, sorry for being a bit vague. :p
 

infomon

Smash Scientist
Joined
Mar 11, 2008
Messages
5,559
Location
Toronto, Canada
Oh, interesting.

One small note, you can fastfall as soon as the aerial begins, the fastfalling needn't wait until the aerial ends.

I wonder if the earliest time at which you can use an airdodge is the time at which you reach top knockback speed. Probably not tho, that'd be way too early......

Edit: oh no worries about vagueness, I'm not really following along with the math yet anyway. I should get to that at some point ;) Mad respect for the work you've done on this.

Edit 2: Oh, another thing. Using a Jump or a momentum-changing Special during knockback gives an extra boost away from the stage. The magnitude of this boost is proportional in some way to the "amount" of knockback left on your character (presumably your speed, I'm not sure what attribute exactly is what matters), as well as something dependent on the move.

It's as if the game tries to give you an extra boost away from the stage, by the same amount of distance that you tried to gain by using the jump/special. However in all my tests, it doesn't really seem correct or consistent in this determination. That is, I suspect each momentum-changing special has some scaling factor associated with it explicitly for telling Brawl how much of an extra boost it should get (proportional to ?knockback speed?), if the move is used during knockback. The existence of exceptions ("braking" a-la bucket braking, also DK's up-B and Yoshi's side-B) also gives evidence to this theory.

Do you have any understanding of these boosting effects? (I'm not sure how relevant it will be to competitive Brawl, but w/e.)
 

leafgreen386

Dirty camper
Joined
Mar 20, 2006
Messages
3,577
Location
Playing melee and smash ultimate
I really don't think that the time you can airdodge is when you reach top speed. Because afaict, you're at your top speed the frame you're launched. You probably lose a set amount of velocity each frame (that's counted toward knockback), although I couldn't tell you what it's relative to, yet.

As for various momentum changing aerials, have you actually confirmed that they do give a boost to your momentum? Momentum canceling all seems like a glitch to me... I don't see why they would purposely put something like this in. Fox's shine is one of the most obvious momentum altering moves, but when I look at vids of fox's shine interacting with the kb, it just looks like he's following his regular trajectory just without the extra pull from gravity. Moves like peach's bomber cause you to move backwards at the start of it, so that doesn't work for testing, either. What moves specifically give a boost forward that are quick enough that you can clearly see the boost given to knockback?
 

infomon

Smash Scientist
Joined
Mar 11, 2008
Messages
5,559
Location
Toronto, Canada
http://ca.youtube.com/watch?v=njcYW0rFmUg
Look at 1:48+. Fox's Shine gives an extra boost upwards.
Fox's up-B as shown is also pretty interesting.

It's more obvious with some moves than others, but let me know if that's not sufficient to convince you. Maybe I don't understand what you're saying about gravity and w/e...
 

leafgreen386

Dirty camper
Joined
Mar 20, 2006
Messages
3,577
Location
Playing melee and smash ultimate
It's very possible that the "boost" seen there is due solely to gravity being canceled/weakened by the shine. He already had some time to accelerate, and when that fall speed suddenly became zero, it looked as if he was "boosted" upward. Horizontal and vertical momentum and knockback momentum are handled separately (or they should be, anyway, from what I've seen so far), which is why I think this is the cause. I'll go test this on zero gravity with a smaller than normal hitstun constant, so it should be especially obvious.
 

3GOD

Smash Ace
Joined
Jul 23, 2006
Messages
745
Location
Athens, GA
@ColinJF - I thought you (and perhaps some others as well) might like to look at this. Perhaps you guys can figure out the correlation (if any) between the stale multiplier and reduction in hitlag, blockstun, etc.

I tested Wario's Fsmash in Training Mode as well as several different levels of decay in VS mode. If any further explanation of the data is needed, just let me know. What (if anything) you do with it is entirely up to you. Also, I'd be glad to test some other moves if it would help you determine the trend.

Training Mode:
Hitlag: 12
Blockstun: 6
Blockhitlag: 12
Base Damage: 19%

VS Fully Fresh:
Hitlag: 12
Blockstun: 6
Blockhitlag: 12

VS Used 1 Time:
Hitlag: 11
Blockstun: 5
Blockhitlag: 11

VS Used 2 Times:
Hitlag: 10
Blockstun: 5
Blockhitlag: 10

VS Used 3 Times:
Hitlag: 10
Blockstun: 4
Blockhitlag: 10

VS Used 4 Times:
Hitlag: 9
Blockstun: 4
Blockhitlag: 9

VS Used 5 Times:
Hitlag: 9
Blockstun: 3
Blockhitlag: 9

VS Used 6 Times:
Hitlag: 9
Blockstun: 3
Blockhitlag: 8

VS Used 7 Times:
Hitlag: 8
Blockstun: 3
Blockhitlag: 8

VS Used 8 Times:
Hitlag: 8
Blockstun: 3
Blockhitlag: 8

VS Used 9 Times:
Hitlag: 8
Blockstun: 3
Blockhitlag: 8
 

infomon

Smash Scientist
Joined
Mar 11, 2008
Messages
5,559
Location
Toronto, Canada
leafgreen: ooooohhhhhh you're saying that using Fox's shine nullifies your fall speed (which I'm guessing has both a horizontal and vertical component), which normally means it stops your downward velocity, but when you're in knockback you're still travelling up from knockback but your fallspeed was helping counteract that, so instead of Fox stopping in place, he travels upwards even faster when his fall speed is set to zero. That makes perfect sense!! You've convinced me, that makes so much sense it's almost certainly what's going on.

So there must be something unique about the way G&W's bucket / Yoshi's side-B / DK's up-B work, since they totally ignore this and reset your *actual speed*. And similarly, many momentum-changing specials also do that at some point, ex. Sonic's side-B and down-B override knockback when they're released after charging (producing a braking effect).
 

ColinJF

Smash Ace
Joined
Dec 21, 2007
Messages
712
If the "base blockstun" is slightly higher than 6 then its values follow exactly what you would expect by multiplying the base blockstun by the stale moves multiplier.

It looks like hitlag is only decreased because of the damage of the move decreasing.

Which leaves me with two questions: how is blockstun calculated that it is slightly higher than 6? and what is going on with this data point?:

VS Used 6 Times:
Hitlag: 9
Blockstun: 3
Blockhitlag: 8
It's the only one where hitlag and blockhitlag aren't the same, and I can't imagine any good reason why they would suddenly become different... whatever operation is done to transform the value for stale moves, I'd think it would be the same for both...

***

Also, fall speed doesn't have a horizontal component, but air control is analogous to fall speed for resisting knockback.
 

3GOD

Smash Ace
Joined
Jul 23, 2006
Messages
745
Location
Athens, GA
Which leaves me with two questions: how is blockstun calculated that it is slightly higher than 6?
Maybe blockstun is some percentage (like 32% or so) of the damage given? Just a wild guess really, but we could check few cases.

and what is going on with this data point?:

It's the only one where hitlag and blockhitlag aren't the same, and I can't imagine any good reason why they would suddenly become different... whatever operation is done to transform the value for stale moves, I'd think it would be the same for both...
Your guess is as good as mine there...I tested it a couple of times to make sure I wasn't seeing it incorrectly. I'll recheck again this afternoon just to make sure...

If we assume the numbers for Blockhitlag and Hitlag are the same, one possibility for this difference is that one might simply be truncated and the other rounded. However, one would expect the difference to appear more often than just this one instance.
 

leafgreen386

Dirty camper
Joined
Mar 20, 2006
Messages
3,577
Location
Playing melee and smash ultimate
Like I said over aim, this is great. Although also like I said over aim, it still needs to be verified whether the constant in determining the speed at which you slow down from kb is the number of frames kb lasts for (if the number of frames of kb is constant, then all kb should last for 35 frames; the rate of slowing would then be the launch speed divided by 35) or if it's just the number itself (probably 115). The latter is much more likely, but I never was able to verify it.
 
Status
Not open for further replies.
Top Bottom