Official Ask Anyone Frame Things Thread

[Hunybear]

How would determine the speed in which Links Bombs would have to impact a wall/ground in order to explode?

 

[ThunderCat]

God Kadano, I really love your work <3

 

[khaleesi]

this may seem like a weird question but how do you do a regular turnaround grab? im very inconsistent with it for some reason.. like for example, when you need to turnaround when chaingrabbing a fox at low percents as marth and he di's behind you. do you need to hold the control stick in the opposite direction before the grab input or at the exact same time and does it matter how much force you put into the stick (like walking and dashing)? maybe it is because the timing for the jump cancel grab, which I do all the time instead of a regular grab, is different in some way. also, please do not confuse this with the pivot grab, which i can surprisingly do pretty consistently. thank you!

 

[Sycorax]

this may seem like a weird question but how do you do a regular turnaround grab? im very inconsistent with it for some reason.. like for example, when you need to turnaround when chaingrabbing a fox at low percents as marth and he di's behind you. do you need to hold the control stick in the opposite direction before the grab input or at the exact same time and does it matter how much force you put into the stick (like walking and dashing)? maybe it is because the timing for the jump cancel grab, which I do all the time instead of a regular grab, is different in some way. also, please do not confuse this with the pivot grab, which i can surprisingly do pretty consistently. thank you!

You can grab at any point in your Turn animation and you will grab in direction you are turning. All that is necessary to start the Turn animation is to move the stick into either of the blue zones on this input map:

There is no special input required. What may be causing your problem is that you are turning then JC grabbing. If you jump during a Tilt Turn, your character will return to the original direction they were facing before the turn. You will then cancel the jump with a grab facing the original, undesired direction.
Either learn to smash turn all the time, or just don't JC the grab which isn't really necessary until like 22%-ish.

 

[khaleesi]

sorry i'm still a bit confused.. so in order to do a turnaround grab i need to smash my control stick in the opposite direction instead of lightly pressing it, while also inputting the grab before the turn so i dont jump cancel out of the turn?

 

[Sycorax]

If you want to JC grab, you have to smash the control stick instead of tilting it lightly. Then input JC grab after starting dash.

If you want to just turn around grab in place, tilt or smash the control stick back and then press Z.

 

[Comet7]

How does knockback stacking work?

 

[schmooblidon]

How does knockback stacking work?

http://www.ssbwiki.com/knockback#Knockback_stacking

The wiki says they need to be seperated by less than 10 frames (after first hit's hitlag) for the knockback to be replaced, and 10 more frames will cause knockback stacking.
In experiments though I get less than 11 frames, although maybe it is about how you describe frames, the idea is the same though.
(Shine > dair)
0 - Last frame of shine hitlag
1 - Shine frame 2, victim sent flying
11 - Last frame to connect a dair to replace knockback
12 - If connecting dair now or onwards, knockback stacking will occur

I also tested 2 similar angles with a large gap but they seemed to merge, so I dunno how accurate that part is. I'll have to figure it all out eventually if no-one gives an answer

 

[Sycorax]

In my testing from a while ago, I found that what the wiki described about angles is accurate. The way they describe it though is confusing. For each knockback vector, horizontal and vertical, they are added if they are in opposing directions, and the bigger one is used if they are in the same direction.

This link describes it well too.

http://m.ign.com/wikis/smash-bros-project-m/Knockback_stacking

 

[schmooblidon]

In my testing from a while ago, I found that what the wiki described about angles is accurate. The way they describe it though is confusing. For each knockback vector, horizontal and vertical, they are added if they are in opposing directions, and the bigger one is used if they are in the same direction.

This link describes it well too.

http://m.ign.com/wikis/smash-bros-project-m/Knockback_stacking

Does same direction mean exactly the same, or wtihin what range? Does it take an angle from the hitbox properties, or does it take the KBVel X/Y (meaning DI is affected).?

So I've been trying to incorporate throw release points into the calculator, but I need some things cleared up. How do I know if a throw is weight dependent, and if it is how do I calculate the frame or sub-frame that the victim is released, so I can find the interpolated ThrowN offset?

I've also been getting very small inaccurate knockback velocity calculations for throws. Is there some other kind of multiplier or variable that is applied on top of the regular calculations for throws?

Magus420

 

[Kadano]

Does same direction mean exactly the same, or wtihin what range? Does it take an angle from the hitbox properties, or does it take the KBVel X/Y (meaning DI is affected).?

So I've been trying to incorporate throw release points into the calculator, but I need some things cleared up. How do I know if a throw is weight dependent, and if it is how do I calculate the frame or sub-frame that the victim is released, so I can find the interpolated ThrowN offset?

I've also been getting very small inaccurate knockback velocity calculations for throws. Is there some other kind of multiplier or variable that is applied on top of the regular calculations for throws?

Magus420

http://smashboards.com/threads/detailed-throws-techs-and-getups-frame-data.206469/

 

[Sycorax]

Does same direction mean exactly the same, or wtihin what range? Does it take an angle from the hitbox properties, or does it take the KBVel X/Y (meaning DI is affected).?

So when you're moving through the air in hitstun, your character has an X Knockback Velocity and a Y Knockback Velocity. These values diminish to zero over time by a factor proportional to angle the move sent you at (it's like .0515*sin or cos(theta) or something). I forget the specifics but it's easily deduced by investigating knockback velocity after being hit by Luigi's dsmash with and without DI. I'm sure it's somewhere in GentleFox's math for your calculator too.

When you're hit again while you have knockback velocity the game does not simply add the component vectors like what would happen in the real world. There are 4 vectors to consider, the X and Y residual knockback velocity vectors and the new X and Y velocity vectors of the new move you're being hit with. For each pair of component vectors X and Y, the game will add the two vectors if they have opposite sign (i.e. they are vectors pointing in opposite directions), and if they have the same sign, the larger of the two vectors will be used.

Let me give you an example. Pichu is flying through the air with knockback velocity (5,15). He is hit with a move that would otherwise result in a knockback of (-11,4). The two X components will add because they have opposite sign resulting in a new X knockback velocity of -6. The two Y components have the same sign so the larger of the two is used resulting in a Y knockback velocity of 15. The final resulting velocity is (-6,15).

It should also be noted that the "first 10 frames" do not count hitlag.

 

[Sycorax]

Sorry for the double post, but I forgot to reply to this part.

if it is how do I calculate the frame or sub-frame that the victim is released

Magus's thread is great, but it has at least two inaccuracies. The first is for calculating the release frame of the throw when hitlag is involved. I talk about the right way to do it here. If you don't want to do the math by hand, I made a spreadsheet that does everything.
But wait, there's more. Such a simple calculation is too pedestrian for the magnanimous depth of Melee. Those calculations will give you the right answer most of the time, but there are exceptions. The only one I have found is Captain Falcon's dthrow on Peach/Sheik/Zelda which will release them on frame 19 instead of the expected frame 18.
Edit: And I found another exception. Popo's dthrow on Peach will release on the correct frame but he will be actionable one frame later than expected, 46 instead of 45. My guess is that this stuff happens because of rounding :/

 

[Magus420]

Are you using 100 as the weight for throw KB? Throw KB is weight independent unless it's a hitbox after release that does the KB like Jiggs f-throw (though the hitbox can miss), Bowser klaw f-throw, and Falco d-throw on NTSC Fox.

 

[schmooblidon]

Kadano Sycorax Magus420

You babes! Looks like I have everything I need, I'll look through it in the morning

 

[reverie2]

Had a couple questions...

What frames of shine can I turn around in? Looked http://smashboards.com/threads/the-fox-shine-guide-by-njzfinest.67333/ and the Fox frame data thread but can't find it.

If I fullhop off stage, I can easily do a quick shine turnaround doublejump bair (using the left stick for jump).

However, if I fullhop off stage and do a shine turnaround, and pause for a long time, if I input up on control stick and bair with cstick, the timing is way off and it never comes out. I have to wait way longer between jump and bair. What's the explanation for this?

 

[Kadano]

Had a couple questions...

What frames of shine can I turn around in? Looked http://smashboards.com/threads/the-fox-shine-guide-by-njzfinest.67333/ and the Fox frame data thread but can't find it.

If I fullhop off stage, I can easily do a quick shine turnaround doublejump bair (using the left stick for jump).

However, if I fullhop off stage and do a shine turnaround, and pause for a long time, if I input up on control stick and bair with cstick, the timing is way off and it never comes out. I have to wait way longer between jump and bair. What's the explanation for this?

The wiki should always be your first address. It's ideal to have all accurate knowledge there.
http://www.ssbwiki.com/Fox_(SSBM)/Down_special

 

[reverie2]

The wiki should always be your first address. It's ideal to have all accurate knowledge there.
http://www.ssbwiki.com/Fox_(SSBM)/Down_special

Thanks.

Do you know if tilt turns vs smash turns apply with turning around in shine? As smash turns only take a frame opposed to 5-9.

It says there's 3 frames where jump cancel is disabled when turning, but that still doesn't seem to explain this situation:

"However, if I fullhop off stage and do a shine turnaround, and pause for a long time, if I input up on control stick and bair with cstick, the timing is way off and it never comes out. I have to wait way longer between jump and bair. What's the explanation for this?"

I did the turnaround and paused for a good few seconds, but the timing for doublejump bair is way different than if I do a super fast fullhop->shine turnaround-> doublejump bair

 

[Kadano]

Thanks.

Do you know if tilt turns vs smash turns apply with turning around in shine? As smash turns only take a frame opposed to 5-9.

It says there's 3 frames where jump cancel is disabled when turning, but that still doesn't seem to explain this situation:

"However, if I fullhop off stage and do a shine turnaround, and pause for a long time, if I input up on control stick and bair with cstick, the timing is way off and it never comes out. I have to wait way longer between jump and bair. What's the explanation for this?"

I did the turnaround and paused for a good few seconds, but the timing for doublejump bair is way different than if I do a super fast fullhop->shine turnaround-> doublejump bair

There is no tilt / smash input distinction for shine turns. I replicated the same situation you are talking about, and regardless of shine stall time, you can always input jump one frame after last shine reflector frame, and bair one frame after jump. So the fault is within your input timing.

 

[reverie2]

There is no tilt / smash input distinction for shine turns. I replicated the same situation you are talking about, and regardless of shine stall time, you can always input jump one frame after last shine reflector frame, and bair one frame after jump. So the fault is within your input timing.

Ah yep. just found out i was letting go of B early in the scenario where I paused a bit after the turnaround

 

[schmooblidon]

I've almost completed the release point calculations, but I seem to be getting incorrect release frames.

Take Fox's Upthrow on Falcon

Upthrow release: 8
Falcon weight: 1.04

8 x 1.04 = 8.32

But in game it is clearly 8.65, and the calculations to find the interpolated release point between frame 8 and 9, lines up with 8.65

I feel like I'm missing something obvious.

Edit: Is it 8 x (1.04 ^2) ? Because that makes 8.6528

Also just noticed the real release point seems to have different interpolations on the x and y

// Magus' release points
"f8":{"throwN":[-0.0352,22.2114]},
"f9":{"throwN":[-0.0670,22.3484]},

difference : -0.0318, +0.137
diff x0.65 : -0.02067, +0.08905
diff x0.6528 : -0.02075904, +0.0894336

real difference: -0.02287, +0.09301
multiplier: 0.7192, 0.679 <--- wtf

Only explanation I can think of is that the imaginary interpolation line is curved

edit2: Feel like I've been trying to calculate the animation frame they are released, and the formula I've been using is for game time frames only, and that the example above using 1.04^2 was just a fluke.

edit3: Ohhh I totally get it. 1/weight, to find how many animation frames pass each game time frame, then just see how far they are above the release frame they are once reached.

Formula (probs reinventing the wheel)

//animation frame per game frame
animPerGame = 1/weight;
animReleaseFrame = release + animPerGame - (release % animPerGame);

//use mod 1, so interpolations above 1.00 will be calculated properly
interpolation = (animReleaseFrame - release) % 1;

then use ceiling(animReleaseFrame) and floor(animReleaseFrame) to get the high and low frame to get throwN and transN

If I'm correct this means I would only need throwN and transN of the release frame plus the 2 frames above. As pichu having the lowest weight has the highest animPerGame of 1.808

 

[Deleted member 212841]

gcyftftdtfyxycyvj

 

[RedmanSSBM]

Kadano schmooblidon

I'm really confused as to why Ganon's down-b sent Falco in the reverse direction in this specific scenario. I know it tends to do that when the opponent is in the air, or when the down-b starts and the opponent is at the back-end of the down-b, but I have never seen it send someone in the reverse direction when they were on the ground when the down-b was already out for a little bit.

I'm guessing this has something to do with ECG, but this has really got me stumped at this point.

 

[Sycorax]

Kadano schmooblidon

I'm really confused as to why Ganon's down-b sent Falco in the reverse direction in this specific scenario. I know it tends to do that when the opponent is in the air, or when the down-b starts and the opponent is at the back-end of the down-b, but I have never seen it send someone in the reverse direction when they were on the ground when the down-b was already out for a little bit.

I'm guessing this has something to do with ECG, but this has really got me stumped at this point.

Reverse hits are determined by the relative positions of character's base positions, a.k.a. TopN bones. If the victim's TopN bone is to the right of the attacker's, the victim will be sent to the right. If it is on the left, they will be sent to the left. The TopN bone does not always follow the animation/hurtboxes of the move. A notable example of this is Marth's fsmash which moves his body forward a great deal, but his TopN bone remains behind his back foot. A similar things happens for Fox's and Falco's jabs though to not as great an extent. You'll notice Falco jabs right before he is hit. Further, Ganon's down-b be has his TopN bone situated rather far forward, just about right under his knee on the extended leg. He also moves relatively fast in his down-b. These three factors, 1) Falco's jab moving his hurtboxes forward but keeping his TopN bone back, 2) Ganon's down-b moving his TopN bone forward, and 3) Ganon moving somewhat fast, all combine to cause Ganon's TopN bone to pass Falco's as he is hitting Falco. This results in a reverse hit.

 

[RedmanSSBM]

Reverse hits are determined by the relative positions of character's base positions, a.k.a. TopN bones. If the victim's TopN bone is to the right of the attacker's, the victim will be sent to the right. If it is on the left, they will be sent to the left. The TopN bone does not always follow the animation/hurtboxes of the move. A notable example of this is Marth's fsmash which moves his body forward a great deal, but his TopN bone remains behind his back foot. A similar things happens for Fox's and Falco's jabs though to not as great an extent. You'll notice Falco jabs right before he is hit. Further, Ganon's down-b be has his TopN bone situated rather far forward, just about right under his knee on the extended leg. He also moves relatively fast in his down-b. These three factors, 1) Falco's jab moving his hurtboxes forward but keeping his TopN bone back, 2) Ganon's down-b moving his TopN bone forward, and 3) Ganon moving somewhat fast, all combine to cause Ganon's TopN bone to pass Falco's as he is hitting Falco. This results in a reverse hit.

Now that's a solid answer. Thanks a bunch

 

[reverie2]

Question about aerial interrupts...

This is when you aerial right before you land, the aerial doesn't come out, and you get the auto cancel lag, which is 4 frames, if I'm understanding correctly. I know samus uses this with upair when doublejumping off ledge, but why is it better than just doing nothing and landing (also 4 frames of landlag), or no impact land (1 frame of landlag)?

 

[Army805]

You can grab at any point in your Turn animation and you will grab in direction you are turning. All that is necessary to start the Turn animation is to move the stick into either of the blue zones on this input map:

There is no special input required. What may be causing your problem is that you are turning then JC grabbing. If you jump during a Tilt Turn, your character will return to the original direction they were facing before the turn. You will then cancel the jump with a grab facing the original, undesired direction.
Either learn to smash turn all the time, or just don't JC the grab which isn't really necessary until like 22%-ish.

What actions will cancel the turn animation? Also are pivots just actually just a smash turn where you don't hold the direction after the 1 frame of turn? The smashwiki on turns (only using smashwiki because there wasn't any readily available information, maybe I didn't look hard enough though) states that "A dash from a turn can be canceled into another turn from its first frame on, opposed to dashes done from other action states, which can only be canceled into turn starting from frame 4." Does this mean that if you dash out of neutral stand, you can't dash back until frame 4? Also it is not possible to tilt turn out of dash, only smash turn, correct?

 

[Kadano]

Does this mean that if you dash out of neutral stand, you can't dash back until frame 4? Also it is not possible to tilt turn out of dash, only smash turn, correct?

Yes. Yes.

 

[Sycorax]

What actions will cancel the turn animation?

Anything but neutral b.

Also are pivots just actually just a smash turn where you don't hold the direction after the 1 frame of turn?

Actually, you can be holding the direction after frame 1. You just need to move out of the Smash Turn zone.

Does this mean that if you dash out of neutral stand, you can't dash back until frame 4?

Kadano simply replied yes to this, but it may be worth pointing out that if you dash backwards out of Wait that you will first enter Turn. Thus, essentially you can cancel a dash backwards out of Wait into a dash forwards on the first frame of Dash.

 

[Sycorax]

Question about aerial interrupts...

This is when you aerial right before you land, the aerial doesn't come out, and you get the auto cancel lag, which is 4 frames, if I'm understanding correctly. I know samus uses this with upair when doublejumping off ledge, but why is it better than just doing nothing and landing (also 4 frames of landlag), or no impact land (1 frame of landlag)?

There's not much point to using an AI right before landing coming down. It's only useful when moving up through a platform or from the ledge.
AI is better than a NIL because it takes quite a while to set up a NIL. They generally can only be done from the apex of a character's jump once they start falling again. In the case of Samus, this won't happen until frame 28 or later of her double jump. If she AI's onto the stage from the ledge, she can land as soon as frame 14 of her double jump. NILing also requires dropping lower from the ledge than the AI.

 

[reverie2]

There's not much point to using an AI right before landing coming down. It's only useful when moving up through a platform or from the ledge.
AI is better than a NIL because it takes quite a while to set up a NIL. They generally can only be done from the apex of a character's jump once they start falling again. In the case of Samus, this won't happen until frame 28 or later of her double jump. If she AI's onto the stage from the ledge, she can land as soon as frame 14 of her double jump. NILing also requires dropping lower from the ledge than the AI.

In the case of Samus, if she performs the exact same dropdown doublejump timing, AI upair would make her touch the ground faster than doing nothing and getting the standard 4 frame impact land right?

 

[Sycorax]

In the case of Samus, if she performs the exact same dropdown doublejump timing, AI upair would make her touch the ground faster than doing nothing and getting the standard 4 frame impact land right?

Yeah. More than twice as quick.

 

[Hunybear]

Is there any resorce that documents which characters are most susceptible to shield stabs at their feet?

 

[Sycorax]

Is there any resorce that documents which characters are most susceptible to shield stabs at their feet?

It's called develop mode.

 

[reverie2]

Does anyone know what frame the sound comes out for tech vs no tech? I believe when you don't tech, you hear a bang sound when you hit the ground, and when you do tech it makes a grunt kind of sound (if i remember correctly, it applies to both techinplace and techroll)... Does it differ by character?

 

[reverie2]

Is falco's laser crouch cancellable? (do you get the hitlag reduction if you're hit when crouching)

 

[Sycorax]

Is falco's laser crouch cancellable? (do you get the hitlag reduction if you're hit when crouching)

Not full cc where you cancel hitstun, but hitlag and hitstun are reduced by 1/3.

 

[reverie2]

Not full cc where you cancel hitstun, but hitlag and hitstun are reduced by 1/3.

Hold on... I can't seem to find the source, but I remember getting a full cc reduces hitlag and knockback by 1/3. Doesn't falco's laser have set knockback based on weight, which means the hitstun won't change because hitstun is knockback*0.4? Are there instances (from other moves) where hitstun is completely cancelled from full cc?

 

[X WaNtEd X]

Could someone give me the frame perfect inputs for perfect wavedash? Just put in x for the jumpsquat or something. I keep messing it up in frame counter and just want to see where I'm going wrong.

 

[tauKhan]

Hold on... I can't seem to find the source, but I remember getting a full cc reduces hitlag and knockback by 1/3.

Crouching reduces hitlag and knockback by 1/3.

Doesn't falco's laser have set knockback based on weight, which means the hitstun won't change because hitstun is knockback*0.4?

Crouching reduces kb (and therefore hitstun) of any hit regardless of it being set kb or growing kb.

Are there instances (from other moves) where hitstun is completely cancelled from full cc?

If the move launches the victim into air (i.e it doesn't spike) and it does enough kb, but not too much so that it knocks down, the victim can hard land via ASDI down (which happens naturally when you're holding down for crouch), and all hitstun gets cancelled. So often when you successfully crouch cancel, you'll only have your regular landing lag instead of a considerable period of hitstun, and that is why cc + asdi down is so powerful.

Lasers happen to be so weak that hard landing doesn't occur, which is what SplitsOnTrees meant.

Could someone give me the frame perfect inputs for perfect wavedash? Just put in x for the jumpsquat or something. I keep messing it up in frame counter and just want to see where I'm going wrong.

If the jumpsquat of the char is X and you input jump frame 0, then you should input airdodge on frame X.