Base Knockback / Scaling formula?

[Jdaster64]

So, I'm a bit confused about how knockback is calculated. I'm familiar with the Base Knockback and Knockback Scaling values in PSA and such (ignoring Weight Knockback for sake of simplicity), but I don't know how exactly they affect the launch speed.

My best guess at a bare-bones formula (no stale moves or external factors) is something like:
d*b + d*D*g/w

where d is the damage the move does, D is the total damage on the target after the move is executed, b is the base knockback value, g is the knockback growth value, and w is the weight of the target.

Am I on the right track, or is this all totally off? I could probably figure this out myself given enough time through editing stuff in PSA, but I wanted to see if I could get this verified here first.

Alternatively, I don't imagine there's an Action Replay code or such that shows the launch speed of the last move used by a given character?

 

[Toomai]

We know what the formula is in SSB64:

((((((damageAfterHit/10) + (damageAfterHit * damageDealt/20)) * weightFactor * 1.4) + 18) * knockbackScalingFactor) + baseKnockback) * ratio
(for fixed knockback: damageAfterHit = 10, damageDealt = weightBasedValue)

We also know the following:

• Base knockback is "the minimum amount of knockback that can be dealt".
• Knockback scaling is "as a percentage, how much knockback is affected by damage". So 100 is the default, 50 means it gets stronger at 0.5x the default rate, etc. I'm pretty sure this has to be input into the formula as a percentage.
• The knockback values reported at the end of a match in Brawl are divided by 0.03 - a launch speed of 3000 is a knockback of 90.
• The Project M team may or may not have the formulas for SSBM and SSBB. If so they refuse to tell us, if not then I don't know how the rumour started.

 

[Jdaster64]

I tested a move that I gave pure upward base knockback of 100 and no scaling on a 100-weight character and it showed 3221mph. That times 0.03 = 96.63; would you expect that is probably supposed to be 100? If so, the dividing factor's actually just above 0.031.

At any rate, thanks for telling me about the factor; I was wondering why when I kept putting powers of 10 in for all factors it kept giving me knockbacks like 322 and 3221. (Also, I'm using Brawl-, which supposedly has no stale moves reduction when a move is not fresh).

 

[Toomai]

Forgot to mention that (in Brawl at least, don't know about Melee) the gravity of a character has an effect on the knockback they take. Also be warned about truncation/rounding, and I would recommend making no assumptions.

The factor is exactly 0.03, it's a value in a known block of numbers that includes things like the Sakurai angle and the clang window.

 

[Jdaster64]

Ah, didn't think about gravity. I'll mess with that some until I get a nice set of values to test with, then.