Sunday, December 30, 2007

Tournament website

I've been spending the last few weekends working on a new project, my Super Turbo Rankings Website. I've finally launched an alpha release. I haven't worked on any styling, right now it's just functionality over form.

I've also taken this opportunity to explore coding in ruby on rails. I could have done this in Perl, which is a language I've developed in professionally, but I thought I'd take this chance to try something new and learn a new language and framework. I was extremely impressed by Ruby, and by the Rails framework. I think it actually took me less time in RoR, even taking in the learning curve, than it would have taken me in Perl!

I want to describe the application a little bit. There are two main pages, the players page and the tournaments page. The tournaments page shows you a list of tournaments, and you can click on a tournament to view details about that tournament. Click on any player name you see to be taken to that player's information page. Right now I only support double elimination tournaments, and I plan on having support for single elimination, round robin, and just a random collection of matches.

On the players page, you can see a list of players ordered by handle. It also shows you the players' TrueSkill ranking. If you click on that player's profile, you can see more information about how their ranking came to be, including a history of all their matches, and how their ranking evolved over time. There's also a graph showing how their ranking has evolved over time.

Your ranking is defined by two numbers, your mu and your sigma. These two numbers define a probability curve. Your ranking is somewhere under that curve, with the most likely ranking being the mu value. The sigma is a representation of your uncertainty. If you have a small simga, your curve is tall and narrow (like the red curve below). If your sigma is high, you have a short and fat curve (like the greed curve below).

Your total level is mu - 3*(sigma). This is a conservative estimate of the actual level , meaning your actual level is probably much higher. This prevents people from winning one tournament, and then being at the top of the rankings forever. Such a player would have a high mu but a low sigma. We don't want that player to be more highly ranked than someone that does well consistently.

I plan to add a plot showing the distribution curve of their ranking.

I also show how a player's ranking evolves over time.

Here's a sample graph:

As expected, the red curve (mu), is always higher than the green curve (level). Also, as expected, as the player plays more games, the level approaches the mu more closely. This is because the player's sigma value goes down over time.

I think this should be a fun way to compare the skill of two players who have never played each other. It's also good to have a detailed record of tournaments. I don't think we have the full brackets for older tournaments (like old Evos), but we can enter the data we do have, and see how the numbers shake out. It is my hope that in the future, this will become the comprehensive site for ST tournament records.

Monday, November 12, 2007

Rebalanced mode and fake fireballs

I'd like to comment a bit on the decision to give ryu a fake fireball in ST rebalanced mode.

Before I start, I want to say that I disagree with the premise of how capcom/backbone is choosing to address the rebalancing. As they mentioned in the first blog post

My goal is to buff up the worst characters so they reach the middle (or upper middle at best). Next, buff the middle characters slightly, but not so much that they become top tier. And finally, leave the top tier characters intact.

I don't disagree with the idea that power levels need to change. What I do disagree with is keeping the top tier intact, and "buffing up" so that lower and mid tier characters are higher in power.

As I mentioned before, I think that the way to do it is to keep the "core" tier intact, and bring the power levels closer together by slightly reducing the power of the top tier characters, so that they are still very good, but missing the really broken stuff, and then buffing up the low tier slightly.

I think the core tier characters are:

Ryu and Guile especially stand out to me as examples of excellent design. They are both extremely strong in the best hands, but no one would call them the best characters. They are the epitome of skill rewarding play. Players that win tournaments with ryu or guile are just plain good.

I think it's a much better design decision to keep these core characters they way they are, and bring the rest around them. You actually end up changing fewer characters this way: In both systems, you have to help the lowest tier. However, in the "keep the core" system(that I am a proponent of), you only have to bring the top tier down, which is four characters (Claw, Dhalsim, Boxer, and Sagat). On the other hand, in the "buff people up" system, you have to bring the entire middle tier up. Since the middle tier is much larger than the top tier, you touch more characters.

The other reason I think that the "buff up" system is wrong is highlighted in this quote from David Sirlin:

Also, as I said when I rebalanced Puzzle Fighter, we already know what the game felt like with the previous top tier characters, and it was fun, so it’s better to balance the game around that power level than a new, lower power level.

As I mentioned on, I don't think that the game at the top tier is nearly as fun as the game at the mid-tier!

I completely disagree with Toodle's position of "never nerf". Why? As a vega player, I can tell you that I've beaten very very good players with repeated wall dives. I don't think this is a fun part of the game. In fact, I've played the Japanese system of soft banning Vega and O Sagat, and you know what? The game was much more fun. And it was more fun because I didn't have to play or play against these idiotic, repeatable traps.

Chun li vs Guile is a fun, intense matchup. Every pixel of spacing matters. There's countless decisions to make. Do I block his jab sonic boom? Do I jump straight up over it? Should I try to trade with a low forward, or will he anticipate that with a backfist?
I haven't seen that kind of depth in most fights that involve O Sagat and Vega. Vega *can* be played in a powerful, complex, and dominating way, but the easiest way to just win is to do wall dives. Similarly with O Sagat's tiger shots.

I think it's high time to remove these kinds of moves. I want more fights like Chun li vs Guile, Zangief vs Hawk, Ryu vs Ken, and Dictator vs DeeJay. Those are fun, stimulating, and exciting fights. There's books to write on those fights.

I want to give Dhalsim, Claw, and O Sagat a chance to have those fights too. Getting rid of those extremely specific and abuseable moves is a great way to keep the inherent power and complexity of the character without destroying the entire game ecosystem they reside in.

The mid tier's fights are more exciting, more deep, and reward decision making better than the fights at the top tier. Changing the mid-tier is dangerous, because you risk losing those great mid-tier fights that I think really define the game.

So that said, I disagree that Ryu should be given any changes, because I think Ryu and his core compatriots are perfect as is.

That said, that is a disagreement with the premise of the change, which is the "buff up" system. I disagree with the "buff up" system, and I think the right call is the "keep the core" system, for the reasons above.

However, I'd still like to discuss the change in Ryu given the "buff up" system. So even though I disagree with the "buff up" system, I want to put aside that fact for now, and discuss what Ryu's change will do in the context of that system.

In this context, I think Ryu's fake fireball will make him stronger. This is hard to evaluate without context, because I don't know how other characters have been changed. Since I don't have that information, I have to think of how the normal ST characters would be affected. This would make Ryu more powerful at mid range, and especially more powerful in the corner. It would be even more difficult for characters like Honda, Chun Li, Dicatator, and Ken to get out of his corner traps.

There is one thing that I think is an absolute must. Ryu's fake fireball should not give him super meter. One doesn't have to look any further than SF3 to see why. SF3 is a game dominated by running away and charging meter. Players are rewarded by not interacting with their opponents, and instead doing safe moves repeatedly that charge up their meter. Once they have it, they approach their opponent.

Ryu already has one of the safest meter building moves in the game, with his jumping back hurricane kick. This is a move that at least forces Ryu to give up a resource, his position relative to the corner. But if the fake fireball built meter, Ryu would have a way to charge meter that would (presumably) be impossible to punish at mid range (if it was possible to punish at mid range, you might as well do a normal fireball, and the move would not be very useful). This would completely change Ryu's dynamic. Ryu players would have a huge incentive to do fake fireballs to charge meter, at which point they have one of the best supercombos in the game at their disposal. It would greatly sadden me to see one of the pinnacles of character design be overshadowed by a meter building strategy.

No meter for fake fireballs.

Saturday, September 8, 2007

Option selecting Psycho Crusher/Scissor Kicks

I've been playing a little bit more of dictator at my weekly gatherings, and I've realized that sometimes I just don't have enough charge for a scissor kicks combo. I realized this when I crossed over a zangief player with jumping forward, and then did three standing shorts. Usually this combos into scissor kicks but sometimes the kicks weren't coming out.

This might be because of the variables charge times in super turbo, and because dictator's scissor kicks is the longest charge time move in the game. The speed at which I press the standing shorts is probably affecting my charge time. If I did a pyscho crusher instead, I would probably have gotten that out every time, but I was greedy and going for the longer charge time pyscho crusher.

Should I stop going for crossover scissor kick combos? No! I had the idea that maybe I could go for an option select.

The idea would be:
Jump with crossup forward
Keep holding lever back
Stand short x 3
Move lever towards opponent
Press roundhouse kick and fierce punch one right after the other

If you have a long enough charge for the scissor kick, the roundhouse kick input should give you a scissor kick. However, if you don't have a long enough charge, then the roundhouse kick should come out for a very short time (1 frame or so), and then get kara cancelled into the fierce pyscho crusher!

So you get to have your cake and eat it too. You can try for the scissor kicks, but if it doesn't come out.

This can also be useful in Dictator's other standard combo
Crossover forward
Stand short x 2
Low foward
Fierce Psycho Crusher

Replace the last "Fierce Psycho Crusher" with a roundhouse instantly followed by a fierce, and you should be able to try for the more damaging combo safely.

I've tried this on my machine, and it seems to work. However, to really prove that this works, perhaps some kind reader would try it out on an emulator? I really need to learn how to set up inputs on emulators. Proving this kind of stuff would be great.

Wednesday, August 29, 2007

Evo 2007 main tournament report

I was quite happy that I had qualified, but now that I was in the main tournament, it was only going to get harder. Of the Seattle ST players, only myself and Nate Montes had qualified. Pablo had almost made it, losing a tough fight to Tony Cannon in the losers finals.

My first fight was against a player I didn't know. He was a Chun player, and I went with Claw. I won that match, and I was starting to feel good about my chances. I had four wins in a row so far!

My next match was against another player I didn't know. He played Chun li. I decided that my Chun Li was pretty good, and if I didn't know this person, he wouldn't be great. Well, in my first 20 seconds of play it was apparent that this player had amazing reflexes. He was very consistent with his reversals, hitting me every time I went for the neck breaker. The point at which I really realized that I was outclassed in the reflexes and timing department was near the end of the round. I had a super charged, and threw a jab fireball. He jumped over it, and I used the stored super trick to hit him as anti air. I saw the blue flash of my super, and then saw ANOTHER blue flash. He landed and counter supered me for the win. Now that I think about it, he must have stored his super and jumped, pressing kick when he landed to hit me. I don't know if I mistimed my super and did it too late, or if that's a legitimate counter, but either way, I was very impressed by his timing.

I was a bit shaken by this loss. I could still salvage the match though. Should I go with Claw or Boxer? I went with Boxer. We played a very close set, and at the end of the third round we both had zero life. His low forward hit me and won the game and match.

Afterwards, I learned that this unknown player is Buktooth, one of the top CvS2 players. I should have stuck to my guns and not changed tracks simply because I didn't know my opponent. Regardless, I was very impressed by Buktooth's innate talent.

Now that I was in the losers bracket, things would get rough. In the losers bracket you advance twice as slowly as in the winners bracket, so that means you have to work twice as hard to go the same distance. It sucks! Plus, there were many good players in the losers bracket. Many many extremely good players were there, because the brackets were set up in such a way that they had to fight each other in the winners bracket early on. I wasn't very happy being there so early, and I resolved to play without mercy and without underestimating any opponent (something I should have done in the winners bracket).

My first opponent in the losers bracket was a player that I didn't know (I've just learned that his nickname on SRK is "shirts"). He picked old Dhalsim, and I was able to win with my standard anti dhalsim strategy. I used a Claw trick that I had been practicing for a while against his Dhalsim, and I was happy to see it pay off in this situation.

After winning against Shirts, my name was announced again. My opponent was OhNuki. My heart sunk. Ohnuki had won SBO (the largest Japanese tournament) a mere month before this event, winning with a perfect against YuuVega, one of Japan's deadliest players. Additionally, I had played against Ohnuki in a tournament when I lived in Japan. This was on a different game, Street Fighter Alpha 3 instead of Super Turbo, but Ohnuki devastated me. Gian had called him one of Japan's five "God Players", along with Tokido and Umehara. I was done for, and I knew it. I picked Claw for my first match. I know that Ohnuki is a great player, and Claw should stay on the ground in this match. My twisted logic was that since Ohnuki was so good, I shouldn't do the BEST moves against him, because he would clearly be too good for that. Instead I should do BAD moves and try to suprise him with wall jumps. This is an idiotic strategy. It's called outhinking yourself or pysching yourself out. It's what bad players do against good players to make themselves play even worse. I lost all of my rounds to Ohnuki, and pretty badly.

So that was it for me. I wasn't super happy with my tied for 25-32nd place. I felt like I could have beaten Buktooth if I had played smarter in the metagame, even though he has more raw talent than me, I think I could have won with Claw if I had picked him.

That evening, I thought a lot about my mindset for Evo, how I felt before and how I felt afterwards. I had been trying to objectively look at my performance, taking my ego out of it. This is something that is hard for me to do, but something that is essential to being a good player. I was especially motivated by an interview I read with Muteki guile, and one of the things that he said was that he looked at every loss and tried to learn why it happened, and to learn from it. So objectively, I would rate myself as "good but not great". I did "well" in casuals before the tournament, I won my qualifying pool, but I lost decisively to one of Japan's top players, and I also lost to a very talented american player. I want to think about my mindset, and how it affects my play.

In order to do this, I need to look at how my mindset and self evaluation has changed over time. Before I moved to Japan, I considered myself "very good". I didn't think there was anyone that was better than me. Sure, there were people that beat me, or did better than me in tournaments, but in my "heart of hearts", I was certain that if I practiced enough, I could beat anyone. I felt that my ability to learn, and ability to impose my will on my opponent was as strong or stronger than anyone elses, and that although my execution wasn't naturally great, that I had the ability to excecute anything if I really put the time into it. Part of this self evaluation was reinforced by my tournament performance. I did fairly well in tournaments, and there was no player that I had never beaten. There were players that did consistently better than me, but I had been able to beat those players both in tournaments and in casual, and I didn't feel that I couldn't do it again.

After I moved to Japan, my mindset changed. The Japanese players were just so much better than me. When I played at MORE arcade, there were players I simply could not beat. I lost to MORE's "Sky High Claw" player 20 games to 0, in the Boxer vs Claw matchup, something I was certain I would win every time. In Japan, I almost never won tournaments (except one). After a while, I changed my mindset. I started playing for the enjoyment of the game, and I didn't have any illusions of being anything other than an "okay" player. I could win in arcades, but I was nowhere near the level of the top players there.

But now, my mindset has changed again, and twice during this tournament. After my loss to Ohnuki I really felt like I was never going to be good.. but watching Graham play in the finals really inspired me. He completely destroyed Gian and Ohnuki. Graham did everything that I should have done against Ohnuki. He stayed on the ground. He played smart. He didn't lose his cool. He didn't change his strategy because of fear of another player. Graham was amazing. Watching Graham has made me come to the conclusion that I can do that well too. I'm not the best ST player, but I am good, and I can use my head to be even better. I've gone from disappointed in my performance to energized and inspired that I can do better next year.

Finally, there is one additional way in which my mindset has changed. When I lived in the US before moving to Japan, I really cared about my own performance. So much so that if I identified another player as a possible threat to my own tournament results, I would not want that other player to improve. After I was eliminated from the main tournament, I saw that a former classmate of mine at MIT, David Sirlin, was in the top 8. David uses Claw. I thought about the two Claw tricks that I had practiced for the last weeks, the tricks that I had hoped would win me the tournament. My old self would have just kept them a secret, and used them next year. But this time I thought to myself that I should wish my friend luck, and give him every opportunity to win. He's a great player, and telling him this could certainly hurt me at a future date. But at this point, I wanted to wish my friend luck and success, and so I told him everything. I'm happy that I can be competitive at this game, and play to really win, but I'm also happy that I'm not so competitive that I won't share and discuss strategies with people that have the potential to do better than me. I'm glad that I've changed for the better in this respect.

Monday, August 27, 2007

Rebalancing ST

After hearing that ST might have a rebalanced mode, I started giving some thought to what should be changed, and why. My first suggestion is that we (as a community) should collect data. We should get data from all the Major ST tournaments, both in the US and in Japan, and get character selection popularity, as well as win percentage. Ideally I'd like to see a list that goes something like this (the numbers are made up)

Characters by popularity:
1. Boxer (1000 picks)
2. Ryu (800 picks).
.. etc..

The reason for this would be to see how often characters appear in competive play, and that way we could validate some of our assumptions of how popular certain characters are.

Anyways, I talked about this with some other players, and thought about the issues. I want to identify the cast of characters I think are at the right power level in terms of work/reward and effectiveness, and then the list of those characters that are above or below.

The right characters:

I think these characters are very well balanced, are strong, and have an appropriate effort/reward ratio.

The too powerful
Dhalsim: He's the best character in the game. There's two aspects of him that are very very good. The noogie trap, and the long range control. I talked to Seth and they said they were considering toning down the damage of the character, but to me, that doesn't address his problem. I would suggest giving him a tiny bit more recovery on the yoga fire. The problem is that he can shut down so many characters with yoga fires, forcing them to jump.

Boxer: Boxer has two dominating factors. His damage and his throw trap. I think reducing his damage is a great way to address both issues.

Claw: Claw's off the wall dive wakup trap is just too easy and too dumb. I would suggest making it so that you always block it the way it came from (if he came from the left wall, you block as if were a left attack). I think with this change he would still be a top tier character but not overwhelming.

O Sagat: His tigers are too fast and he does too much damage. He is probably the second easiest to dominate with character after Claw. Capcom went a bit too far with New Sagat, something in between is probably correct.

The too weak
Zangief: He's a tough one. I don't feel like I know Zangief well enough to be really informed, but it seems to me that toning down the "too powerful" tier is good enough. He seems to do well against the Balanced and Low tiers.

N Sagat: See Old Sagat

Cammy: Cammy is very dangerous in the right hands. I would suggest doing one or both of the following: Make her jab spinning knuckle invulnerable for a slightly longer period of time, or shorten the recovery of her cannon drill slightly.

Fei Long: I don't know what to do for fei long. Maybe he's already good enough if the top tier is lowered?

T Hawk: Give him O Hawk's normals, and make his condor dive faster on startup.

Chun li: I'm not sure if Chun li is balanced or not. I think she's on the border of Balanced and Good, which is probably fine. I would just make her neck breaker kick always blockable the same way. Make it always block backwards, so that it's something newer players have to learn. Otherwise she's probably okay as is.

Honda: I hate Honda. We need a way to help him be better vs projectile characters that doesn't make him dominate even more over non projectile characters.

Blanka: The thing about Blanka is that he's actually pretty close to being good. He was a monster in hyper fighting. This might be another case of leave him alone and just tone down some of his predators.

Dictator: With the top tier being toned down, he might be too good? Unsure.

Evo 2007 pool play report

Saturday at Evo was pool play. This is where the unwashed masses fight and jostle for the chance to make it out of the pools, and into the 64 man tournament that determines the ultimate winner of the event.

One of my biggest fears was not making it out of the pools. I originally thought there would be eight pools of eight, and that you would need to make top four of your pool to qualify. I didn't think that was so bad. Then I found out that the pools were much bigger! My pool had 17 players! That meant you had to be in better than the top 25% of the pool just to get a chance to play in the "real" 64 man tournament. Yikes. With strong players like Gian (Evo 2005 champion) and Ohnuki (SBO 2007 champion) unqualified, that meant it would be very easy to get knocked out in the pools!

With my pool size, I needed to win three straight games to qualify. In a double elimination tournament, wins in the winners bracket are twice as valuable as wins in the losers bracket. It's a long, tough slog to go through the losers, because even though half the field is being eliminated each round, you're getting half the winners backet injected into your field each time as well. So the more you win early on, the easier your road will be. I won my first two games against players I didn't know. I used Boxer and Claw -- I switched about evenly between the two so as to make my main pick unpredictable. If it came to a character switch, I would pick Boxer if my opponent was Claw, and Claw if my opponent was Dhalsim.

I've played a lot of Boxer vs Claw, so I knew the matchup well. You have to be very patient, and never ever lose your charge, or you will get wall dived to death. You can either fierce headbutt the wall dive (gets you safely out of range but does no damage), or you can kick rush him as he is going for the wall. My second opponent told me later he had an "uh-oh" moment when he saw me doing the kick rush :). If Boxer doesn't know about those two things it can be a very rough match.

My third match came up, and this was for a qualification. It was against Jet Phi, a tournament organizer and a player that had appeared at the first seattle tournament at Preppy's house that I went to several months ago. At the time, I had won a very close set to win the tourament. His honda had beaten my Boxer several times during the set. I checked the bracket -- if I lost to Phi, I would have to face and defeat Jef Perlman, a classmate of mine from MIT. I said hi to Jef.. I hadn't seen him in ten years! We joked that we could form "team MIT" with David Sirlin for the 3-on-3 team tournament. We never got a chance to ask Dave if he wanted to do that, but the team tournament never happened anyways so it didn't matter. I had played Jef a lot at the MIT arcade, but it was all on alpha 2, never on Super Turbo, so I didn't know how good he was. With that in mind, I started my match against Phi.

I knew that Phi was a honda player, so I went with Chun Li. Now my traditional characters for tournaments have always been Boxer and Claw, but during my time in Japan I studied under the apprenticeship of Akishima, the best Chun Li in Tokyo. I had been hoping to get good enough with her that I could play her in a tournament. I feel that Chun Li has a very big advantage over Honda. As an aside, I later saw Seth Killian pick Honda against a Chun Li! Strange, since I think Honda has such a disadvantage. I asked him about it, and it sounds like a lot of US players think Honda wins that fight! My strategy against honda with Chun is to throw jab fireballs, and use low roundhouse to hit him if he jumps. If he's at a range where I can't throw a fireball (too close), I jump back with forward kick, which hits his torpedo and his superman. Jump straight up fierce works well too. The main danger is ending up in the corner, where one jump in can mean death due to ochio throws. This is a match where Chun can easily perfect Honda. I stuck to my game plan, but Phi suprised me by using jump roundhouse to hit my low roundhouse when he was jumping over my jab fireballs. This was not good, because my whole plan centered on this. I went to my plan B of jumping back with forward, but although I could do about 50% of his life with this, I ended up getting killed by ochios. My game plan didn't work because without the low roundhouse to do a significant chunk of his life and knock him down, I wasn't making it difficult to get me in the corner.

After my loss, I thought a bit. Should I stick to Chun, or switch? I knew Phi could beat my Boxer, since he had done so before at the aforementioned tournament. I went with Boxer, because at least I knew my gameplan with him, and I didn't want to have to figure out a second plan with Chun Li on the fly.

My Boxer plan is to again play very defensively. Stand jab and low strong hit his torpedo, and I use jab rushes to do blocked damage. I stay charged and headbutt any jumpins. I'm even happy to trade with low fierce if I'm not charged. The main fear is that he will superman from close range, and cross me up, which hits the buffalo headbutt. If this happens you have to jump back with strong, but that won't work in the corner since you don't jump back far enough. So, as usual, you have to avoid the corner. I was sobered by my first loss, so I played very carefully and won with a perfect in one of the rounds.

I had qualified, and better yet, won my pool! Next was the main tournament, which would be 32 pool players like me, and the top eight from each of the four Evo Regional tournaments. Before I wrap up this post, I want to mention the amazing performance of Nate "XTG" Montes. He never heard his name called in his pool, and was given a second chance where he had to play against Japan's KKY in his first game. KKY is a dhalsim player, and Nate plays Zangief exclusively. This is pretty much an impossible matchup. Nate lost, and was at the very bottom of the losers bracket. Nate climbed his way up, winning an astounding SEVEN MATCHES IN A ROW. His last match was against Ino, a Guile player from Japan. There was a huge crowd around Nate's game, and time after time, he was down to almost no life when he would get his opponent in the corner and kill him with a spinning pile driver. Nate made excellent use of what I like to call "grappler fear". Basically every one of Zangief's moves is easily countered, but a good player will make his opponent so afraid of one move (like a jumping splash) that he will hit with another move (like a walk up piledriver). You could palpably feel the terror of his opponents as they sat in the corner, trying to counter every move precisely. Nate would eat several low forwards from Ino, and then jumping splash twice in a row. This is something Guile can easily defeat, especially a player of Ino's caliber, but Nate had him so terrorized that it worked. The cheer from the crowd when Nate won was deafening. Congratulations Nate! You were really amazing.

Sunday, August 26, 2007

Sunday Evo 2007

I just had a pretty good afternoon of practice sessions with Graham Wolfe, Alex Valle, Jason Nelson, Jason Cole, and Brian whose last name I don't know. Brian and I found Graham right after his incredible evo performance to congratulate him. We had lunch and then we got the other players together to practice. Nelson was really into money matches. He was trying to make every game a money match! We played some money matches.. I beat Nelson's Guile and Dictator with Chun li, and then lost to his Old Boxer(!), and one other match I don't remember. We ended up going 2-2. I should have just quit when I was 2 games ahead.

We played from around 3pm to 7pm, so a good four hours. The most useful for me learning wise was when the others had left to get stuff and I played Valle's Ken 1 on 1 for 20 or so games. Playing again reminded me why I think he is a level above almost everyone in natural talent. He was doing things like supercomboing my Boxer's rushes. Maybe I was being predictable, but he was doing it a lot, and a lot more than happened in Japan.

Anyways, I played a lot of Claw and Boxer against him, and I was having a pretty hard time overall, but especially against his old Ken. I felt like I had no way around the jab fireball trap. Graham and the others came back in, and I watched how Graham won. Then both Valle and Graham helped me with what I was doing wrong.. namely I wasn't punishing jab DPs with dash uppers, and I wasn't timing my jab headbutts over fireballs right. I was doing them too early and always getting swept or thrown.

With Claw, playing against Alex forced me to be really patient. After some tips from Valle, I was able to jump over his new Ken's jab fireballs, but I still had a hard time with Old Ken. The general pattern was that with both Boxer and Claw, I was being too aggressive and impatient getting over fireballs, instead of being patient and blocking. This is something that it helped to have these players point out explicitly, because in my mind, I wasn't being agressive ENOUGH! I thought I was being really defensive, but in actuality I was being overly agressive. Now that I write this, I realize that this is something I see lot of other players do too. They keep attacking when they shouldn't and get beaten bad because of it.

Thanks a lot to those players for taking the time to play. I felt like I really improved my game. At one point, someone knocked on the door and there was a crowd of like 15 or so players that wanted to play too. I'm sorry guys for turning you down. I don't mean to exclude anyone, but the whole reason we had gone to play in a private room was to avoid the ridiculous lines on console play. Also, when you want to sit and practice matchups, you only want to be playing with a handful of players, so the rotations are quick and you can get your game in right away if you lose. We had that many players on Thursday night, with the Japanese guys, and it was pretty long.

I want to write about my own performance later, but I want to write about Graham now before I forget about it. Graham played in a really inspiring way. Last night, after I was eliminated, I was feeling pretty angry and frustrated. I lost really badly to Ohnuki, and I really thought that I just could not beat him, ever. But watching Graham play completely changed my mind and my outlook. His match against Gian was incredible. I'll put a link to it here when it goes up on youtube. Gian is someone I played for YEARS and was never able to beat him with Boxer. Graham completely destroyed him. Even so, when I saw his draw against Ohnuki I knew he was going to lose. But Graham played a perfect Claw. He completely destroyed Ohnuki too. He stayed on the ground, and kept him in the corner, playing patiently, and didn't try to go for wall dives. I'm now officially a Graham Wolfe barn.

What really impressed me about Graham's play was that he thought about the match analyzed it, and then used his understanding to win. This process was very visible in the way he played. I'm reminded of the Muteki guile interview that inspired me before Evo.

Tuesday, August 14, 2007

SF2 code v1.0

I posted this sometime back in early '97 on I've seen it get picked up a little bit in signatures on SRK, so I thought I'd repost it here.

SF2 code v1.0

(The following section adapted from Robert A. Hayden's Geek code v2.1)

We Street Fighter players are many. Some of us are newbies, some of us are masters, some are scrubs. This is my first attempt to provide a code for fellow SF2 players to
recognize each other simply, and without undue flamage.

How to tell the world you are a SF2 player, you ask? Use the universal SF2 code. By joining the SF2 players organization, you have license to use this special code that will allow you to let other un-closeted SF2 players know who you are in a simple, codified statement.

The single best way to announce your SF2 player status is to add your SF2 code to your signature file and announce it far and wide. But be careful, you may give other SF2 players the courage to come out of the closet. You might want to hang on to your copy of the code in order to help them along.

The SF2 code consists of several categories. Each category is labeled with a letter and some qualifiers. Go through each category and determine which set of qualifiers best describes you in that category. By stringing all of these 'codes' together, you are able to construct your overall SF2 code. It is this single line of code that will inform other SF2 players the world over of what a great SF2 player you actually are.

Some of the qualifiers will very probably not match with you exactly. Simply choose that qualifier that MOST CLOSELY matches you. Also, some activities described in a specific qualifier you may not engage in, while you do engage in others. Each description of each qualifier describes the wide range of activities that apply, so as long as you match with one, you can probably use that qualifier.

Also, pay particular attention to case-sensitivity, there can be a big difference between a 't' and a 'T'.


SF2 players can seldom be quantified. To facilitate the fact that within any one category the SF2 player may not be able determine a specific rating, variables have been designed to allow this range to be included.

For variable, said trait is not very rigid, may change with time or with individual interaction. For example, SF2 players who use ken's short short throw all the time but think all other throws are ultra cheap would have a rating of t+@--

For indicating "cross-overs" or ranges. SF2 players who go from M+ to M--- depending on the situation (i.e. mostly "M+") could use M+(---). This could be a polite player who sometimes gets really obnoxious when he wins a lot.

For 'wannabe' ratings. Indicating that while the SF2 player is currently at one rating, they are striving to reach another. For example, T- -> +

@ is different from () in that () has finite limits within the category, while @ ranges all over.

SF2 players vary in their uses and views on throws.

My name is Rasheed Rankins, or Lord Baal. My goal in life is to throw you. I'll throw instead of combo you when you are dizzy, just to make a point.

I have no compulsions about using repeated ticks. They win me the game, don't they?

I'll throw you if you're turtling, but I don't go out of my way to throw.

I'll sac throw, or walk under throw, but that's it. Walk up throws and/or ticks are cheap!

I'll slam your face on the machine if you throw me. The last Zangief player I fought now is now missing a testicle.

SF2 players range from Ken Kombo Kings to struggling beginners.

I can do DeeJay's 15 hit combos in my sleep. I think only wimps use one digit combos.

I can do Guile's Fierce/Fierce/Sonic Kick and Ken's TOD. If you are dizzy against me, you lose.

I can do Fierce/Fierce/Fireball most of the time.

Wow! I got a two-in-one off the other day!

I heard that jumping Roundhouse/crouching Roundhouse was a combo, but I'm not sure...

I've spent time developing my one combo, which I can do consistently, but I can't do others of the same caliber.

Timing is an important aspect of SF2, not easily learned.

I never fail well timed attacks. I supercombo people on reaction to jabs.

You'd better not fireball against my charged Balrog. Not even Vega can jump in on my Ryu.

If you jump in on my Ryu, you'll eat a DP 90% of the time. I get most walk under throws.

I find air throws difficult to do. I have a hard time safely jumping straight up over fireballs.

I can't seem to get Vega to throw after a wall dive, even when my opponent is just standing there.

Oh! It's do the joystick motion, *then* push the button!!

Are you a lean mean killing machine? Or are you a softhearted sissy?

I ask my opponent to tie third round and then throw a fireball one second before time runs out. I try to get 5 year old girls to play me so I can stay on the machine longer.

I crush newbies like the girly men they are! I like to use big combos and ticks when they're trying to figure out blocking!

I use any method within the game to win. I'll pretend to be weaker against weaker players to 'milk' them for the most amount of money.

I kid around sometimes, but I don't give second rounds, not even to newbies.

I'm a nice guy. I'll give second rounds, and I'll play goofy characters even against good players just for the fun of it.

I don't care whether I win or lose. The fun is in playing the game! I make it a point not to stay on the machine for more than 2 turns.

I've devoted my life to playing a goofy character and playing him as ruthlessly and crushingly as possible! I travel the world to beat on overconfident Ryu players! Arggghhh!! *foam* *froth*!!!!!!!! HAHAHAHAHAHAHA!

How well have you mastered the fireball and it's various traps?

One I get someone in the corner with my Ryu, they never get a hit on me again.

I'm considered a "master" at local arcades. I fake fireballs, and DP them when they jump in.

I can throw fireballs, and DP, but people jump over my fireballs and hit me occasionally.

Are they telepathic or what? Every time I throw a fireball, I get kicked in the face!

Oh, you can vary the speed of the fireball with the button pressed!? Ahhhhhh....

Some SF2 players are adept with T-Hawk and Zangief. Others are still learning...

I can tick with Zangief's Supercombo. Consistently.

I can SPD 100% of the time. I've spent quite a bit of time with grapplers, and I can crush you like an empty tin can if you blink.

I'm a fairly average grappler. I can SPD most of the time, I guess.

I spin the joystick around like crazy and mash on the buttons, shaking the machine and jumping up and down. It works more often than you think!

Forget it. After 5 years and 100K, I've given up on trying to get an SPD.

I'm on a holy crusade to brand as cheap anyone who used an SPD. It's the devil's work.

Okay, fess up. This is the true test of your passion....

I spend over $50 a week on street fighter. I've bought my own machine too.

I'll easily spend over $5 in one sitting. Once I've found good competition, the money just keeps on flowin'

I'm restrained enough to put in one quarter, play one game, and leave afterwards.

I spend less than a dollar a week on street fighter.

I still have tokens left over from when I last played... in CE.

I'm so damn good that I'm still playing the same game from last week!

I don't keep track of how much money I spend. That way, I can play guilt free!

Just how good are you?

People leave when they see me put my quarter up. I don't lose games.

I've had winning streaks of over 10 against serious competition. Most of my opponents pay more than I do.

I win some and lose some. I don't rule the machine, but I've beaten even the masters occasionally.

I hardly ever win, but I'm getting better!

Win? What's that?

Everyone around me sucks so much that I think I must be skilled for beating them. I tell everyone I meet how great I am, never realizing that I am in fact a scrub.

I usually lose, but then again, my competitors are so damn good that I can justify my losses. I'm really quite good, really I am.

Some SF2 players play Ryu only, and some play the spectrum. How varied are you?

I play every single character.

I play 4 or 5 characters regularly.

I have a main character and a backup character if I lose.

I play one character. I'm pretty much screwed if I can�t beat my opponent with him.

I've never even tried any other character but my own.

I'm still deciding who to play.

Polite? Rude? Tell us about your 'arcadiquette'.

If I beat my opponent, I offer him a quarter and a smile.

I always compliment my opponent on a game well played. I never put anyone down.

I tend to ignore my opponent as a person, prefering to concentrate on the screen.

I laugh at my opponent when he screws up, and brag about how great I am to the crowd.

I'm hated at arcades all over the world. I publicly humiliate my opponents and put my elbows on their playing space.

The Net:
How much are you involved on The first value is for reading, the second for posting. So someone with n++:-- has been a total lurker on agsf2, but has read every single article ever posted.

I have read every article on agsf2 since it's creation:I'm known as a net.personality
on agsf2.

I read most articles on agsf2: I post quite a bit, and the veterans recognize me.

n :
I read agsf2 when I'm bored: I've posted a few articles.

I rarely read agsf2: I posted once or twice, but I've never really had anything to say.

I just stumbled here by accident. Who are you wierdos anyway?: I'm a total lurker. Never posted.

Just how big a part of your life do you devote to this game?

SF2 is my life. I stay up at night thinking of combos and strategies. I'm going to marry Chun Li.

I play every day. I read SF2 fiction, and tell SF2 jokes. My friends think I'm wierd.

I play occasionally, but that's about it.

I really don't have time to make SF2 a priority.

SF2 fiction? You guys need to get a life!

I don't really care about the game, but I still want to marry Chun Li.

With that posted, here's my SF2 code:

SF2 Code v1.0: t+ c+ T+ r+(-) f g+ m+ s+ v+ M+(-) n+:++ o+

Sunday, August 12, 2007

Speed equivalences

Tonight, Umbrellastyle and I tried to put to rest something that has been bothering us for a long time. We attempt to answer the following questions

1) What is the difference between Free Select and Arcade fixed speed?

2) What is the difference between USA and Japanese arcade boards of super turbo?

3) What speed on CCC2 corresponds to the arcade versions of Super Turbo, both American and Japanese?

In order to answer these questions, we needed some equipment. We have
* A PS2 with CCC2. The monitor used was NOT HD, to avoid any HD lag
* A japanese arcade board of Super Turbo, on an actual arcade machine (not super gun)
* A usa arcade board of Super Turbo, on an actual arcade machine (not super gun)

We decided on the following methodology. Using a stopwatch, we would time the round timer three times. During our testing, we found that the speed of the stage affects the speed of the timer, so we timed the most common stages (Ken, Guile, and Fei Long).

Now before I get to the data and our conclusions, let me explain the various settings on the various machines.

The US Arcade version of super turbo has five settings for game speed. They are:
Turbo 0, Turbo 1, Turbo 2, Turbo 3, and Free Select.

The Japan Arcade version of super turbo has also five settings for game s peed. They are:
Turbo 1, Turbo 2, Turbo 3, Turbo 4, and Free Select.

Finally, CCC2 also has five settings. They are:
0, 1, 2, 3, and Free Select

Now part of the confusion here is that when it comes to the turbo settings, the american numbers start at 0, while the Japanese numbers start at 1. However, the Free Select options are 1, 2, and 3, regardless of game. Because this confusion, I am going to use the following terminology for the remainder of this document:




With me so far? :)

Now, the assumption that everyone has been making so far is that CCC2 is perfect emulation of the USA arcade. A fairly reasonable assumption. But I haven't found any actual data backing this assumption up. Part of our goal was to determine whether or not it was true, and to settle the question once and for all!

On fei long's stage, we got the following data (each of these numbers is the average of three runs)



CCC2FREE SELECT 2: 56.47666667

A couple of strange things jump out here. Let's go over them one at a time

First of all let's take a look at the CCC2 Free select speed 3. It clocks in at 49.62 seconds, where USA free select 3 clocks in over a second slower, at 50.76.
Now some of you may be asking "so what? It's only a second?". Fair enough. But note that the difference between USA Free Select 2 and Free Select 3 is a second. So CCC2 Free select 3 is faster than ANY arcade machine. It's faster than ANY speed the Japanese arcade or USA arcade machines can even produce! If anything, it would be equivalent to "free select 4" if such a thing was possible!

The other wierd thing we noticed was that on Fei long's stage of CCC2, Free select 2 is actually SLOWER than Free select 1! This sounds unbelievable, but here's the results of three separate runs:

Fei Long's stage:

CCC2FREE SELECT 3: 49.73, 49.63, 49.5

CCC2FREE SELECT 2: 56.47, 56.54, 56.42 <- SLOWER THAN FREE SELECT 1!! WTF??

CCC2FREE SELECT 1: 52.67, 52.79, 52.85

Okay, so maybe Fei Long's state is wierd. Let's take a look at Ken's stage.

Ken's stage:

CCC2FREE SELECT 3: 50.04, 50.12, 49.93

CCC2FREE SELECT 2: 49.91,49.85,49.94 <-- FASTER THAN FREE SELECT 3!!

CCC2FREE SELECT 1: 53.51, 53.48, 53.26

So CCC2 Free Select is just messed up. Sometimes Free Select 2 is FASTER than Free Select 3, sometimes SLOWER than Free Select 1! We did not see this behavior on either of the arcade versions.

USA Free Select options are almost identical to the Factory set options. The equivalences are:

JPN Free select options are almost identical to the Factory set options, BUT ONE SPEED OFF. The equivalences are:


I hope this clarifies much long standing confusion. Now what conclusions can we draw?

First of all, CCC2 is really not a very good version. Unfortunately, it's too late to change the logistics, so we should make do with what we have. Free select 3 is DEFINITELY too fast. It's wrong enough that there's a significant difference. So what speed should we use?

Well, aside from the craziness regarding Fei and Ken's stages, it turns out that CCC2 Free select 2 is equivalent to USA FREE SELECT 3, and CCC2 Free Select 1 is equivalent to USA FREE SELECT 2.

Our final, and most important conclusions are:



If you are interested in viewing our data, you can see it here.

Dictator's Jab->Super

There's a Dictator combo I've always seen Japanese players do, and I've never been able to figure out how they do it. It's simply standing jab into super. Now that I've got a PS2 and CCC2, I've messed around in training mode, and I was able to do it. Here's how you do it

Charge back for 2 seconds.
Joystick towards
Joystick back
press jab
Joystick towards
press kick

It works! This is also how you do Boxer's stand jab into super. Now it's just a matter of practice, but the main combo (for both characters) is

low jab x 2
stand jab

The timing that seems to work for me (at least for Boxer) is to just do the regular super motion, and press jab twice, the first jab being a tad early. I'm not very consistent with it but it's getting better.

Monday, July 23, 2007

Five easy pieces

Now that I have my own machine, and that I get the chance to play uninterrupted, I've been able to give more thought to some of those little mysteries I've always noticed but have never been able to really explain. Here are three realizations I've come up with that could be useful in competitive play.

Why does a normal throw come out when I'm trying to SPD?
I noticed this one when I first started playing T. Hawk back in college. Every so often I'd get that stupid fierce grab instead of a SPD when I was sure I was doing it right. I'm sure this happens to everyone. What's going on?
My theory is that you are ending your SPD in the forward or back direction, and that you are simply in range of your normal throw. Since the command for a throw is joystick towards (or back) plus punch, that means that if you do your 360 motion on the stick in such a manner that in ends in towards (or back), you're inputting the command for both a 360 and a normal throw. The easiest way to avoid this confusion is to simply make sure your joystick motion ends in defensive crouch. Note that this doesn't help Zangief, who has a down+punch grab. Another reason to play Hawk!

Why does a jump come out when I'm trying to SPD?
As in the situation described above, the problem is that the joystick is ending in the wrong position. If the stick is in any one of the toward jump, neutral jump, or back jump positions, you have a good chance of jumping if you press the button too early or too late. Just avoid this problem by ending in defensive crouch.

Why can I walk up SPD with a button press but not a button release?
This one was driving me nuts. I simply can't do a walk up, spin the joystick, and release punch to get an SPD. I get a jump, every single time. But when I walk up, spin the joystick, and press punch, an SPD comes out. Why? My theory is that when you press the button, a punch starts coming out for a few frames before it is interrupted into an SPD. This grounds your character for those crucial frames! When you do a button release, there is nothing to keep you on the ground as soon as the joystick hits up. So if I'm right, this means that the walk up SPD command is more like:

towards, down towards, down, back, down back, back, press punch, back up, up, up forward, release punch.

Why do some guiles charge faster than others?
This I just figured out today. Maybe it's old news to some, but it was news to me! It will definitely help in competitive matchups. I found out while experimenting with Dee Jay that you can do a sonic boom with the following motion:

back, forward, back, punch

instead of the usual

back, forward, punch

This doesn't seem like a lot, but it makes a difference. In the latter example, you go back to charging as soon as your sonic boom is released. But in the former example, you are charging before the sonic boom is released. This gives you those few extra frames needed to get your next boom out before the other guy.

Why does Dee Jay's low fierce interrupt when close but not when far?
The situation I'm thinking of here is one that has been driving me crazy. Dee Jay can cancel his crouching fierce into a sobat kick, sometimes. At least, sometimes for me. When the computer whiffs a fierce DP, I can do it every time. But when computer Dictator whiffs a roundhouse slide, it never comes out! Also, I can't seem to ever combo crossup forward, low jab x 2, low fierce, sobat kick. I can do crossup forward, low jab, low fierce, sobat about half the time. What gives?
I think what's going on is that the fierce has close hitting frames that exist at the beginning of the animation, while it's still interruptible. After a few frames, the hit box grows, but by that point the move is no longer interruptible. So that's why it combos
I'm less sure about this one, more investigation may be needed. I'll continue to work on it and see if I can solidify it.

I really wanted to find another tidbit so I could call this post "Six Easy Pieces" :)

Sunday, July 22, 2007

Machine Gun Upper

For some reason, I keep coming back to Dee Jay. There's something cathartic about playing him against the inhumanly difficult CPU. The CPU is so hard that it forces you to really get every one of your moves down. I've been practicing combos around the Machine Gun Upper for the last few days, and there's mainly two I've been working on.

Crossup forward, low jab, low short, MGU
Crossup forward, low jab, low jab, MGU

These two combos seem very similar, but there are differences. It seems as if the low short interrupts faster into the MGU than the jab. However, I seem to get thrown a lot by the computer a lot in between the jab and the short when I am trying the latter combo. I think this is because the short button press has to happen faster than the second jab would. So I've been switching to jab.

I can't seem to get the last hit of the MGU. I get 3 hits on it, but the fourth remains elusive. I'm also not sure how to control the speed of the MGU. Sometimes they hit slowly, and sometimes they all hit so fast you can't see it.

Friday, June 29, 2007

Thursday gathering

I've starting setting up regular super turbo gatherings at my house. We had one yesterday, with record attendance of 12ish people. That's actually a bigger regular crowd than many arcades.

We've been doing round-robin tournaments, and this time we did it Japanese style, where everyone picks a character and sticks with it for the tournament. I went with Dee Jay, since I wanted to practice. I still can't seem to regularly combo his sabot kick, and I can't even understand why sometimes it doesn't even combo with itself. Nevertheless, he's a fun character, a lot more fun to play than Guile.

I'm really happy with my japanese super turbo cabinet. It plays just like my days in Japan at MORE, game in Namiki, Tri-tower, and shibuya kaikan arcades, my old super turbo stomping grounds. I'm actually very seriously considering getting a second one, for head to head play.

I played a lot of Hawk in casuals, and I'm really happy to say that I was doing his safe jump/release piledriver about 3/5 tries. I won several matches with the "unbeatable" series, where you do that in the corner and just repeat it, and it was tremendously satisfying. I get a sense of real pleasure doing that unlike other corner traps.

As an aside, the Japanese have a pretty name for corner traps. They call it:

鳥かご - "torikago"

Which means bird in a cage. Cute, isn't it? :)

I have a confession to make. It's ugly, and I'm ashamed to say it, but I'd only be lying to myself if I didn't just accept it. I can't reliably dragon punch on the first player side. It's shameful. I've been hiding this weakness for years, using sad substitutes like throwing backwards to get myself on two player side, jump up air hurricane kicks (with ryu), and other random acts of garbage.

For those of you reading that have no idea what this means, let me try to explain.

As you can see in the
picture, there's two sides to street fighter. The one on the right is called player two, and the one on the left player one. The characters start out by facing each other, so the moves for player 1 are mirror images of the moves for player two. If the characters jump over each other so that end up switching sides on the screen, then you have to do all of your inputs in the other direction.

So anyways, there's a particular move called the dragon punch, whose inputs on the player one side are:
joystick right
joystick down
joystick down/right + punch

when you are on the player two side, it's:
joystick left
joystick down
joystick down/left + punch

What I'm saying is that I can do the second one, but not the first one. Well I can do it, but not consistently. Not every single time. And in this game, if you miss one when you need to do one, you can lost the game because of that mistake. So not being able to do it is a pretty big flaw in your game. I've been covering up this weakness by having my character jump around on the screen and try to switch sides with my opponent when I'm on the side that I can't do the move on. Smart opponents of course figure this out and block you.

Enough. Enough is enough. If I can learn button release spds, I can damn well learn how to freaking dragon punch on the one player side, can't I? How hard can it be?

Well, as I learned last night, it's still pretty hard. I'll get there.

Monday, June 25, 2007

Release SPDs

I was practicing yesterday on my ST machine after my gathering on Saturday, and I figured I'd try a little bit of DeeJay to see how he played. I looked for some videos on youtube, and found several ones that look pretty useful:

Jump forward, low jab x 2, upkicks (6 hits)
Jump forward, low fierce, RH Sabot (4 hits)
Crossover jump forward, low jab x 2, stand strong, RH Sabot (6 hits)

I experienced an "a-ha" moment after playing around with those. Up until now, when I was playing a character or trying a combo, I would try to just do the move without really thinking about exactly where the joystick position was at any given time. In other words, the game was moving so fast that I'd just kind of "try it", rather than explicitly knowing what I was doing at every frame.

Well, playing yesterday, there was a point where the game just started to slow down. It's no so much that the game slowed down per se, but more like all of a sudden it was slow enough that I could really play attention to what I was doing. The moment came when I was doing low jab x 2 into stand strong. I had been just kind of mashing the standing strong and sometimes getting a low strong, when all of a sudden the game slowed down, and I could very easily just move the joystick into the back position, and press the strong button. Voila, it worked!

Excited by this new way of seeing how the game works, I thought I should try my hand at a combo I've always seen, but thought nigh-impossible. Stand strong into upkicks. This kind of combo has always seemed really advanced, because you need to press strong when the joystick is in the neutral position, on the way from crouch to up. So the motion is

Joystick in down/back position for two seconds
Joystick in neutral position, press strong
Joystick in up position, press forward

I was able to do it on the first try! And I had no problem getting it four more times in a row. It just felt like the game was slower, and I had all the time in the world to make a careful move into neutral while hitting strong, and then up with roundhouse. It didn't feel rushed or complicated at all.

So with this victory under my belt, I thought I should try my holy grail -- the release SPD tick. This is a core part of the monster T Hawk strategy I am trying to develop. The idea is, you base your whole game on the safe jump / release spd tick. It starts with a knockdown (DP for example). Then

Safe jump jab on your opponent.
Joystick to defensive crouch, pressing and holding jab.
Immediately after the jab, press and hold strong and fierce.
After pressing fierce, do a circular motion on the stick to back, up, forward, and back to defensive crouch. It's important to go in that order, so that you maximize the number of joystick positions on the ground near the end of the input.
As your stick reaches defensive crouch again, release fierce.
Release strong.
Release jab.

The end result is a fairly standard jump jab, low jab, SPD. The magic here is that your SPD comes out safely. If your opponent did a reversal DP, you just sit there in defensive crouch. If your opponent didn't, you SPD him.

I've been trying to get this tick down for the last several days, but it's always felt hurried and rushed. When I play against the computer, I always find myself doing a low jab and then jumping back with fierce, or doing a standing fierce, or something else that makes it obvious that I'm not doing the inputs right.

If I'm really doing the inputs right, then even if I screw up, the worst thing that should happen is that I end up in defensive crouch. I went back to T Hawk, and played against the computer, and found that this is exactly what was now happening. Now, I'd do jump jab, low jab, and then either an SPD would come out, or I'd be blocking. No more flailing on the stick, no more randomly jumping back with punches coming out. I was carefully hitting each of the 4 corners on the stick too, no more wild spinning and hoping something comes out.

This is the first gestalt moment like that I've had for a long time in this game. I'm definitely looking forward to coming home and seeing if this feeling still persists!