DeLux
Player that used to be Lux
- Joined
- Jun 3, 2010
- Messages
- 9,302
As a penultimate contribution to the IC Boards, I thought I'd sort of talk about Desync Theory for a little while. Most of my time from here on out is going to be spent on maining Lucas and working on general Brawl mechanics in leading the Smash Lab.
I'll preface this post with the following info: I am pretty bad. I don't place well. But I do know the technical in's and out's of the character better than probably anyone else, save for maybe one or two people off the top of my head. I hope that doesn't come off as arrogant. It's just true at this point. I hope by leaving this all here, someone will pick up where I leave off and eventually know more than even I do. However, if my understanding of the mechanics are correct, the shenanigans I've come out with recently are probably the most efficient that we can get. I hope I'm wrong, because it's not QUITE enough to push us over the top. But all of it makes us pretty damn close in my opinion.
To begin, I started off playing asking myself, "Why is desyncing the best way to go when playing Ice Climbers?" Based on my understanding of the character from a matchup to matchup basis, I observed the following generality:
The matchups we're at a disadvantage in are the one's where the primary strategy has been to remained synced.
Metaknight? Everyone says stay snyced. Toon Link? Projectiles mean to play synced. ROB? Laser means you should be synced. Snake? The explosives everywhere make it hard to track Nana so you should stay synced. Peach? Syned bair her out of every thing or she'll Dair separate thing. I can't really think of anymore "bad" matchups off the top of my head, but for the most part, the bad matchups are the one's where our desync game gets jammed.
Which lead me to the following conclusion: perhaps the matchup isn't best played synced. Maybe all matchups are best played desynced but some characters just make it harder to desync safely. So I put my testing into finding faster and safer ways to desync in order to combat that.
Looking at in the context of proactive vs. reactive, this is how it breaks down in my mind. There are two dominant lines of thought on how to play ICs. First, which I am biased against, is the pivot grab everything mentality. In my opinion, this method of playing is weak because it is dependent on your opponent approaching. And while there is little doubt pivot grab is a powerful tool, I like to use it in the context of flashing it just to scare my opponents to not try to approach super fast where they might be able to overwhelm my more proactive desync play. Once they see I'm going to pivot grab an approach, they are more likely to try to pick and choose spots to approach rather than going hard aggro. This allows to safely desync, which is the second style of play. Desyncs lead to true grad setups and frame traps. So instead of waiting for the grab to come to you, you can go out and get it. If that doesn't scream proactive, I don't know what will.
So now that you have committed to a desync style of play, I want to make a distinction between standard desyncs and what I call setups. For my vernacular, desyncs are the actions that desync the ice climbers. To be honest, most desyncs are garbage and/or redundant to each other. What we want to move from is using a desync to using a setup, where the player is in absolute command of the frame trap. Typically the setups that are being have the following scenario: Popo is free to act while Nana is doing a special
There are a few reason this is better than when Popo is performing a special while Nana is free to act. For one, Popo is at a 5 frame advantage over Nana just based on the nature of inputs, so that's the first reason why this is superior. Minus a few exceptions, Nana is limited on the direction she'd be able to pivot grab. Popo is able to pull Nana along with him in terms of movement, while Nana is tethered to Popo in the case where he's stationary. If Nana dash grabs out of tether range, she might randomly upthrow, which never happens when you control Popo. Most of this is pretty intuitive, but I just wanted to throw it out there to cover the bases.
I'm going to expand on the setup idea though. For the most part, in order to safely set a frame trap, an IC needs the following things: 30 frames and one DSC length in spacing. Ice Block comes out on frame 17 as a hit box. You can't shield cancel run until frame 22. I arbitrarily added in 8 more frames because that's the length of shield drop.
To make this clear. IF YOU HAVE A HALF A SECOND OF FREE TIME AND SLIGHTLY OVER A DASH LENGTH IN FREE SPACE, YOU CAN BE SETTING UP. Outside of Falco, I can't really think of a single character that would be able to truly end that capability to desync for a majority of the match. Combining a setup with DSG theory as Sopo is broken against pretty much every character that doesn't have DDD grab range or ROB ftilt range based on my experience. Albeit, my main problem is sometimes I take "can" as "should" and figure since I'm able to do it, I do it for whatever reason. Playing from setups gives ICs a huge ground zoning advantage. You can (and should imo) use frame traps to zone for space in order to force your opponent in the air.
So now that we've hit our criteria, let's look at some setups that I favor and why based on the zoning and frame exchange:
Lux Desync 1: This is easily the most versatile and most powerful desync tool we have. The spacing trade on it is one dash length out of whatever buffer window it's performed out of. It can be done out of: Dash, Run, Dash Attack, Shield Drop, Landing Lag, Auto Cancelled Aerials, Non Auto Cancelled Aerial, Attacks, off the top of my head. For example, you can reverse aerial rush at an opponent and on landing elect for one of two options. First, you can LD1 setup straight out of landing. Or you could simply land and put a shield up, and then LD1 out of shield. You can retreat from an opponent, Shield cancel your run, and then reverse out of shield drop with a setup. Being able to use this in both retreating AND approaching situations is such a huge option limiter on the opponent.
Cost: Start up of one Dash, Spacing of one Dash
Lux Desync 2: This is the compliment for LD1 almost entirely for the fact that it's done out of the things LD1 cannot be performed out of. Those would be: Spot Dodge, Roll, Blizzard, Ice Block, and Crouch. I use this if I'm trying to cross from playing super synced to going to a setup. Out of crouch is probably one of the most efficient approaching setup we have
Cost: Start up of one turn around and one dash, Spacing of one dash
2 Dash Dance Desync: I use this if I'm trying to be more aggro. I tend to use LD1 as my bread and butter, so I like to use this one as a mixup because it has slightly more aggressive spacing. If done in the non shielddrop version, you gain a little less than a dash length from a standstill. If LD1 is done from a standstill, unless your turn around shield drop or do it out of landing lag, it's fairly neutral to defensive and retreating space costing.
Cost: 2-7 Frames Dash one way + 22 Frames Dash the other, spacing cost of some distance reverse direction + (Dash length - the reverse dash distance)
3 Dash Dance Desync: This is probably the most commonly used setup, and I typically hate how it's used. However, I've come to appreciate one major benefit it has. Most setups leave nana behind Popo. A 3 dash dance desync will keep Nana fairly close to Popo after the setup is out. It's slightly slower and can trip, but it slightly more aggro than even a 2 Dash Dance based on spacing trade.
Cost: 2-7 Frame Dash Dance + 2-7 Frame Dash Dance + 22 Frame Dash, spacing cost of usually one dash length depending on the inputs
Hylian Desync: Neutral spaced desync out of reverse initial dash. I'll use it as a mixup occasionally where I'd do an LD1 out of dash shield cancel, but there's the reality of 1 frame window of inputs that bothers me. Generally if I want to use this to net a grab, it's in a situation where if I mess up the input I'm going to get punished heavily. It's sort of the trade off of the inputs. However, if done correctly, this is THE FASTEST aggro desync we have to cover ground if dashing both times in the same direction. It's also a quick reversal, but again, that whole safety thing makes me tentative to use it.
Cost: 22 Frames Reverse Initial Dash, Spacing is *usually two dash lengths or neutral depending how used.
I * there because people (and by people, I mean Cheese) have documented being able to walk instead of performing a second dash. However, this is beyond my realm of tech skill so I just use it with dash because I throw what I know.
SND > Shield Drop: I'll use this if I have absolutely no space behind me and no space in front of me. And generally as only a mixup because it defies the criteria of having a dash length in spacing. But it's a very spacing neutral method of doing things.
Cost: 6+ frames of SND, 22 frames of Nana Dash, spacing is almost 0
Misc other desyncs I'll use:
Split Stick: If done out of cooldown of moves, it's dashless AND has nana setup to smash. It's fairly hard to setup because it takes planning ahead because the inputs require attack being held down, but using this as a mixup creates some interesting pacing gimmicks, especially with squall/belay cancelling.
Cost: Holding down attack, Fsmash Frames for Nana if not cancelled, Spacing 0
Squall and Belay Desyncs: Setting up out of recovery for obvious reasons. Costs are sort of N/A
Other Thoughts and Theory Craft:
Hitlag: Setups out of hitlag. If I get motivated, I'd look straight into this area. I have a theory on the hit lag desync mechanics. Nana does less damage than Popo. At a certain point, less damage means that there's less hitlag. Based on the principles around the Hylian desync, that means that if someone is frame perfect out of move cooldown on hitlag, they ought to be able to perform a setup out of hitlag, even if both climbers hit. I don't know exactly what the potential is in this area, but I'm sure there's quite a bit and it's the next frontier.
You can't Hylian Desync out of Roll, Spotdodge, Crouch, Ice Block, or Blizzard. Those 5 options are what I call the LD2 exceptions. I think there's something in the programming that circumvents the Nana flipping mechanics of being synced. So if you were to buffer a reverse initial dash out of spot dodge, instead of nana simply flipping by sync mechanic on frame 3, she's actually do a true 1 frame turnaround and dash mirroring Popo's movement. Since this is the case, you are able to LD1 out of reverse initial dash instead.
In theory, hitlag might also work in the same principle if the equation doesn't result in an inequality in the hitlags between Popo and Nana. So at certain percents, hitlag is probably an LD2 exception. At other percents, it acts as its own desyncing agent. I'm purely theory crafting on this one since I haven't really had the motivation to test it via the hit lag equation based on every staleness of move and every move's damage. Too much testing and not enough time.
Things I don't know: I don't understand why the dtilt desync works. It must have something to do with crouching and it being an LD2 exception, but for the life of me I don't understand it.
I might add more if I think of anything. But that's all I have for now.
Additional Note: Not all setups are frame traps using specials. There are other variations, but IB and blizzard are the most linear in terms of flow chart. There are some cool more complex things you can do with smashes in terms of frame trapping as well.
I'll preface this post with the following info: I am pretty bad. I don't place well. But I do know the technical in's and out's of the character better than probably anyone else, save for maybe one or two people off the top of my head. I hope that doesn't come off as arrogant. It's just true at this point. I hope by leaving this all here, someone will pick up where I leave off and eventually know more than even I do. However, if my understanding of the mechanics are correct, the shenanigans I've come out with recently are probably the most efficient that we can get. I hope I'm wrong, because it's not QUITE enough to push us over the top. But all of it makes us pretty damn close in my opinion.
To begin, I started off playing asking myself, "Why is desyncing the best way to go when playing Ice Climbers?" Based on my understanding of the character from a matchup to matchup basis, I observed the following generality:
The matchups we're at a disadvantage in are the one's where the primary strategy has been to remained synced.
Metaknight? Everyone says stay snyced. Toon Link? Projectiles mean to play synced. ROB? Laser means you should be synced. Snake? The explosives everywhere make it hard to track Nana so you should stay synced. Peach? Syned bair her out of every thing or she'll Dair separate thing. I can't really think of anymore "bad" matchups off the top of my head, but for the most part, the bad matchups are the one's where our desync game gets jammed.
Which lead me to the following conclusion: perhaps the matchup isn't best played synced. Maybe all matchups are best played desynced but some characters just make it harder to desync safely. So I put my testing into finding faster and safer ways to desync in order to combat that.
Looking at in the context of proactive vs. reactive, this is how it breaks down in my mind. There are two dominant lines of thought on how to play ICs. First, which I am biased against, is the pivot grab everything mentality. In my opinion, this method of playing is weak because it is dependent on your opponent approaching. And while there is little doubt pivot grab is a powerful tool, I like to use it in the context of flashing it just to scare my opponents to not try to approach super fast where they might be able to overwhelm my more proactive desync play. Once they see I'm going to pivot grab an approach, they are more likely to try to pick and choose spots to approach rather than going hard aggro. This allows to safely desync, which is the second style of play. Desyncs lead to true grad setups and frame traps. So instead of waiting for the grab to come to you, you can go out and get it. If that doesn't scream proactive, I don't know what will.
So now that you have committed to a desync style of play, I want to make a distinction between standard desyncs and what I call setups. For my vernacular, desyncs are the actions that desync the ice climbers. To be honest, most desyncs are garbage and/or redundant to each other. What we want to move from is using a desync to using a setup, where the player is in absolute command of the frame trap. Typically the setups that are being have the following scenario: Popo is free to act while Nana is doing a special
There are a few reason this is better than when Popo is performing a special while Nana is free to act. For one, Popo is at a 5 frame advantage over Nana just based on the nature of inputs, so that's the first reason why this is superior. Minus a few exceptions, Nana is limited on the direction she'd be able to pivot grab. Popo is able to pull Nana along with him in terms of movement, while Nana is tethered to Popo in the case where he's stationary. If Nana dash grabs out of tether range, she might randomly upthrow, which never happens when you control Popo. Most of this is pretty intuitive, but I just wanted to throw it out there to cover the bases.
I'm going to expand on the setup idea though. For the most part, in order to safely set a frame trap, an IC needs the following things: 30 frames and one DSC length in spacing. Ice Block comes out on frame 17 as a hit box. You can't shield cancel run until frame 22. I arbitrarily added in 8 more frames because that's the length of shield drop.
To make this clear. IF YOU HAVE A HALF A SECOND OF FREE TIME AND SLIGHTLY OVER A DASH LENGTH IN FREE SPACE, YOU CAN BE SETTING UP. Outside of Falco, I can't really think of a single character that would be able to truly end that capability to desync for a majority of the match. Combining a setup with DSG theory as Sopo is broken against pretty much every character that doesn't have DDD grab range or ROB ftilt range based on my experience. Albeit, my main problem is sometimes I take "can" as "should" and figure since I'm able to do it, I do it for whatever reason. Playing from setups gives ICs a huge ground zoning advantage. You can (and should imo) use frame traps to zone for space in order to force your opponent in the air.
So now that we've hit our criteria, let's look at some setups that I favor and why based on the zoning and frame exchange:
Lux Desync 1: This is easily the most versatile and most powerful desync tool we have. The spacing trade on it is one dash length out of whatever buffer window it's performed out of. It can be done out of: Dash, Run, Dash Attack, Shield Drop, Landing Lag, Auto Cancelled Aerials, Non Auto Cancelled Aerial, Attacks, off the top of my head. For example, you can reverse aerial rush at an opponent and on landing elect for one of two options. First, you can LD1 setup straight out of landing. Or you could simply land and put a shield up, and then LD1 out of shield. You can retreat from an opponent, Shield cancel your run, and then reverse out of shield drop with a setup. Being able to use this in both retreating AND approaching situations is such a huge option limiter on the opponent.
Cost: Start up of one Dash, Spacing of one Dash
Lux Desync 2: This is the compliment for LD1 almost entirely for the fact that it's done out of the things LD1 cannot be performed out of. Those would be: Spot Dodge, Roll, Blizzard, Ice Block, and Crouch. I use this if I'm trying to cross from playing super synced to going to a setup. Out of crouch is probably one of the most efficient approaching setup we have
Cost: Start up of one turn around and one dash, Spacing of one dash
2 Dash Dance Desync: I use this if I'm trying to be more aggro. I tend to use LD1 as my bread and butter, so I like to use this one as a mixup because it has slightly more aggressive spacing. If done in the non shielddrop version, you gain a little less than a dash length from a standstill. If LD1 is done from a standstill, unless your turn around shield drop or do it out of landing lag, it's fairly neutral to defensive and retreating space costing.
Cost: 2-7 Frames Dash one way + 22 Frames Dash the other, spacing cost of some distance reverse direction + (Dash length - the reverse dash distance)
3 Dash Dance Desync: This is probably the most commonly used setup, and I typically hate how it's used. However, I've come to appreciate one major benefit it has. Most setups leave nana behind Popo. A 3 dash dance desync will keep Nana fairly close to Popo after the setup is out. It's slightly slower and can trip, but it slightly more aggro than even a 2 Dash Dance based on spacing trade.
Cost: 2-7 Frame Dash Dance + 2-7 Frame Dash Dance + 22 Frame Dash, spacing cost of usually one dash length depending on the inputs
Hylian Desync: Neutral spaced desync out of reverse initial dash. I'll use it as a mixup occasionally where I'd do an LD1 out of dash shield cancel, but there's the reality of 1 frame window of inputs that bothers me. Generally if I want to use this to net a grab, it's in a situation where if I mess up the input I'm going to get punished heavily. It's sort of the trade off of the inputs. However, if done correctly, this is THE FASTEST aggro desync we have to cover ground if dashing both times in the same direction. It's also a quick reversal, but again, that whole safety thing makes me tentative to use it.
Cost: 22 Frames Reverse Initial Dash, Spacing is *usually two dash lengths or neutral depending how used.
I * there because people (and by people, I mean Cheese) have documented being able to walk instead of performing a second dash. However, this is beyond my realm of tech skill so I just use it with dash because I throw what I know.
SND > Shield Drop: I'll use this if I have absolutely no space behind me and no space in front of me. And generally as only a mixup because it defies the criteria of having a dash length in spacing. But it's a very spacing neutral method of doing things.
Cost: 6+ frames of SND, 22 frames of Nana Dash, spacing is almost 0
Misc other desyncs I'll use:
Split Stick: If done out of cooldown of moves, it's dashless AND has nana setup to smash. It's fairly hard to setup because it takes planning ahead because the inputs require attack being held down, but using this as a mixup creates some interesting pacing gimmicks, especially with squall/belay cancelling.
Cost: Holding down attack, Fsmash Frames for Nana if not cancelled, Spacing 0
Squall and Belay Desyncs: Setting up out of recovery for obvious reasons. Costs are sort of N/A
Other Thoughts and Theory Craft:
Hitlag: Setups out of hitlag. If I get motivated, I'd look straight into this area. I have a theory on the hit lag desync mechanics. Nana does less damage than Popo. At a certain point, less damage means that there's less hitlag. Based on the principles around the Hylian desync, that means that if someone is frame perfect out of move cooldown on hitlag, they ought to be able to perform a setup out of hitlag, even if both climbers hit. I don't know exactly what the potential is in this area, but I'm sure there's quite a bit and it's the next frontier.
You can't Hylian Desync out of Roll, Spotdodge, Crouch, Ice Block, or Blizzard. Those 5 options are what I call the LD2 exceptions. I think there's something in the programming that circumvents the Nana flipping mechanics of being synced. So if you were to buffer a reverse initial dash out of spot dodge, instead of nana simply flipping by sync mechanic on frame 3, she's actually do a true 1 frame turnaround and dash mirroring Popo's movement. Since this is the case, you are able to LD1 out of reverse initial dash instead.
In theory, hitlag might also work in the same principle if the equation doesn't result in an inequality in the hitlags between Popo and Nana. So at certain percents, hitlag is probably an LD2 exception. At other percents, it acts as its own desyncing agent. I'm purely theory crafting on this one since I haven't really had the motivation to test it via the hit lag equation based on every staleness of move and every move's damage. Too much testing and not enough time.
Things I don't know: I don't understand why the dtilt desync works. It must have something to do with crouching and it being an LD2 exception, but for the life of me I don't understand it.
I might add more if I think of anything. But that's all I have for now.
Additional Note: Not all setups are frame traps using specials. There are other variations, but IB and blizzard are the most linear in terms of flow chart. There are some cool more complex things you can do with smashes in terms of frame trapping as well.
![[FBC] ESAM](/data/avatars/l/92/92614.jpg?1365783210){kind=avatar}
