Wednesday, December 31, 2008

Rapid Improvement: Part 2

In part 2 of this series, I want to focus on the next thing that causes beginners to lose games: Basic execution of special moves. There are two components to this.

1) reliable: Can you consistently do the move in training mode?
2) atomic: Do you think of it as one single action instead of a sequence of steps?

I'm sure everyone can identify with the first part. "My opponent is jumping in, but I don't want to risk doing a dragon punch because I'm afraid I'll screw it up". You must have a strong command of the mechanics of doing the moves that you want to do. You simply cannot compete otherwise. I've seen beginners have a difficult time with what seem to be straightforward moves like Guile's flash kick:

* Insufficient charge
* Press button too early and get a hopping knee
* Press button too late and get a jumping kick
* Move joystick too slowly to up position
* Move joystick to forward position instead of up

Regardless of why they fail at execution, the impact of the failure is clear. Assume that you deal 15 damage on successful execution of a flash kick (to simplify I won't even go into the positional advantage of setting up a crossup afterwards). But if you fail, you eat a jump roundhouse, low roundhouse combo for 20 damage.

If you fail your flash kick half the time, then on average you are losing 5 damage per flashkick. Clearly blocking is better than taking 5 damage. The conclusion from this is: It's not worth trying for special moves unless you know you can do them.

Atomic is the term I use when a move or sequence of moves is so ingrained that you cease to think of it as a series of joystick movements and button presses, and instead think of it as one "thing" to do.

It takes a lot of time and effort to think about what your next move has to be in a sequence. Here's an example from my personal play in another game, Virtua Fighter. In that game, the character Wolf has a combo whose stick motions are back, torward, punch + kick, torward, kick, up torward, punch + kick, towards, kick.

At first when I was learning this, I would get about one third of the way through the sequence on muscle memory, and then I'd have to think "what's next?", and remember the next series of button presses. By the time by brain finished remembering what was next in the sequence, I'd have very little time to actually execute.

If you can precache the entire sequence of desired moves into muscle memory, then you don't need to waste time thinking about how to do them during combat. This frees up a ton of time and attention on other things.

Level 2: The one hundred drill
In this drill, you go into training mode and do the move in question unil you have one hundred successes on each side, each day, for one week. So for example

flash kick, right hand side: 20 failures, 100 successes
flash kick, left hand side: 50 failures, 100 successes

Do this every day for one week, rain or shine, even if you are sick or tired. This will drill into into your muscle memory. You may be wondering why moves like flash kick, which are symmetrical, need to be drilled on both sides? The reason is that in play, you will be starting from a different blocking position on each side. So in play, the moves are different because they start from defensive crouch (which is different on each side).

Write down how many failures and how many successes you had each day, on what side. Post it on a public blog to track your improvement.

In the second week, do the same, but first to fifty successes.

After the second week drill is completed, do runs to twenty once a week to keep yourself in shape.

Tuesday, December 30, 2008

Rapid Improvement, Part 1

Recently I've been teaching a friend of mine the basics of the game. I've taught a fair number of people how to play, and I've come to realize that there are no good guides (that I know of) on how to improve, and improve rapidly.

Watching two beginners play street fighter is a lot like watching two beginners play chess. P1 makes a huge, game losing blunder. P2 then fails to capitalize on it, and makes his own game losing blunder. P1 incorrectly capitalizes on that, and makes his own blunder. It's hard not to cringe while watching.

But we've all been there. We've all started as newbies. When I was in college I had a systematic system to teach rank beginners how to play and how to improve. So I'd like to start a series on how to improve in your street fighter play rapidly.

Level 1: The blocking drill

Most rank beginners that I observe lose their games simply by failing to block. A typical losing beginner sequence will be something like the following (vs Ken)

1) Get hit by low roundhouse (should have blocked high)
2) Opponent jumps in with jump roundhouse and hits (should have blocked high)
3) Opponent follows up with a low roundhouse (should have blocked low)

I'll see this 3 hit sequence happen again and again in beginner play. The fundamental problem isn't uncommon at high level play either. I've won many games that I should have lost when my opponent simply failed to block a lethal jump in combo.

In college, I devised a drill that helped new players learn to block correctly. We both pick Ryu. One player has the role of attacker, and the other (the one doing the drill) has the role of defender.

The attacker attacks the defender with the following moves:
* any crouching kick
* any standing kick
* any jumping kick

The defender may only block.

The defender wins the drill if, by the time the timer has reached zero, he has taken no life.

Some tips for the attacker:
Jump roundhouse, low roundhouse
Walk up low roundhouse
Walk up, jump straight up roundhouse (on the way up)
Walk up, jump straight up roundhouse (on the way down)
Jump roundhouse, jump roundhouse again
Jump in whiff, low roundhouse (this one seems especially hard for beginners)
Stand roundhouse
Stand roundhouse, low roundhouse
Stand roundhouse, jump roundhouse

10 or more hits: Basic blocking is a fundamental problem. Keep practicing.
4-6 : You're getting better, but you will still lose your games because of blocking.
1-3 hits: Better. Keep doing this from time to time and strive to reach zero.
0 hits: Very good

Thursday, November 27, 2008

I'm number one

I know these numbers are completely meaningless, but I can't help myself. It's probably the first and only time I'll ever be ranked first on a 360 game :)

Wednesday, November 26, 2008

I played a marathon all night session last night with the local seattle players to celebrate the arrival of ST:HDR, slept till 2, then promptly woke up and played again. It's fun to be excited about a video game again, I haven't felt this way since college!

I'm going to run down my impressions of the game. I'm sure that anything I post about balance will be wrong in the long run, but it's fun to speculate anyways!

Best new stage: Fei Long. His stage went from the most annoying stage music ever to one of the best. I love it!

Unranked matches: When I play unranked, I play a few games with my opponent and then a bunch of other players end up in the lobby. When I lose, I have to wait several games before it's my turn again. Why is it set up this way? I'm not particularly interested in watching two players I don't know play it out until it's my turn. Why can't I just keep playing against my opponent, GGPO style?

I found a wierd bug in the unranked matches. At one point stage music completely turned off, only to turn on again intermittently. Also, I was playing Blanka, and electrified Fei Long. He stayed with a crackling aura the entire game, it even lasted through the next round!

Ranked matches: I really wish your rank was tied to the character you select. Since the character selection is a double blind process, and you only play one game anyways, there's no meta game for the character selection process the way there is in standard tournaments. If I'm playing Chun Li, I'd like to have my ranking kept seperate from my ranking as M. Bison. This may be something that was out of Capcom's control, but it's still annoying.

I was really afraid that the gameplay would be dreamcast like. I don't know why, but I've always had a really hard time pulling off combos on the DC. I can do them fine on arcade, but there's some subtle timing difference or something that makes the DC combos much harder for me. I was afraid the same would carry over on the xbox. Nope, not at all! I don't know if it's because the "randomness" was removed or what, but I can do my favorite Bison combos! Yay!

Chun Li: I picked Chun Li with a sense of dread, thinking she'd been hit hard with the nerf bat. Reduced super damage, no big upkicks combo afterwards, and nerfed neckbreaker. She's be a shadow of her former self.
What I found is that she's a shadow of her former self if you are Nuki, NKI, Akishima, or some other excellent Chun Li player. If you're a good, but not great Chun Li like me, then she's got this incredible new move called lightning legs.

In classic ST, I've never been able to pull off combos like Standing Fierce -> Lightning Legs. I even have a really hard time with low forward -> legs. So now that the number of mashes have been reduced (an excellent idea), I can pull them off consistently. I can now poke with walking up low foward, lightning legs. It's a huge, huge difference for me. So top players that could already do all of this are probably sad at the state of the new Chun. But for me, I feel like I've been given a boost and can now do the things I could always see them do, but I could never do. I love her.

M. Bison:
M. Bison is my nomination for most well designed all around character. I cannot stop gushing about how great his changes are. One tiny, tiny change was all that was needed, and it was the perfect tiny change. He has kept his powerful offense, and has been given one "get out of jail" card to improve his atrocious defense. The new devil's reverse, with invulnerability, gives him an option out of corner fireball and SPD traps that he never had before. The reason I think this change is so great is that it's not a guaranteed way out, it's just an option. If I'm in the corner and my opponent does a move that would trap the old Bison (last minute FB, or low jab-> SPD), I can now devils reverse to fly out safely. But if my opponent knows I will do this, he can simply delay his attack and nail me with almost any move. So Bison is still at a great disadvantage in the corner, but just not a game losing nothing-you-can-do disadvantage. Perfectly done.

Blanka is my prediction for new monster, and I'm going to try to learn him. All of his old Komoda blanka nastiness, which was already deadly, and an overall improved set of moves. I think he will end up being very, very good.

Cammy: I can't seem to get the right "flow" of Cammy, but Jason Cole was destroying everyone last night with her without even breaking a sweat. She seems to have all the offensive power of Bison, Blanka, and Balrog, but without even having to think about charging moves. I guess we'll have to see what happens with her.

Balrog: Felt pretty much the same, honestly.

Monday, October 6, 2008

Tournament reports - PNW majors

So I've realized that my old tournament report strategy is fail. What I've been doing is after a tournament, instead of posting immediately, I want to wait for youtube videos to come out so that I can analyze the matches and write about what I did wrong/right.
However, what happens is that by the time the video comes out it's old news (to me), and I don't have the motivation to post anymore. So I'll write up my backlog of reports now.

Pacific Northwest Majors (6/28/08)
Round one, I played against a Dhalsim player and almost lost. Holy shit! That was kind of a kick in the ass for me, and I think I "woke up" and played better.

In the winners finals, I played LTB's O. Ken on my very own ST cab. He just flat out raped me (I went with Boxer). However, I did notice something -- the way that he was raping me. What was happening was that he was throwing out fireballs from long range, letting me jump over them, and then walking up and throwing me. I know that's a basic as hell strategy but it's also very effective. I use it myself all the time when I'm playing against Boxer.
So I thought a lot about my loss, and came up with an idea of how to beat him if I was lucky enough to make it to winners finals.

Then I was in the losers finals where I played Alex Kelly. His Guile has gotten a lot better. He's pretty consistent about landing his crossup short combo, and that can be pretty scary. I decided it would be too risk to play Boxer or Dictator against him, so my choices were Claw or Chun Li. I feel that although Claw can dominate more, it's also an easier match to lose than Chun. In other words, I see Chun Li vs Guile as a 60-40 match with low variance, while Claw is like a 70-30 match but with high variance. I decided I preferred the more consistent match, so I went with Chun.

I ended up losing the first match, which was another kick in the ass. I think I was playing too conservatively, and letting the match become 50-50. I realized that I had to take more risks, but that those risks would be in my favor. I changed my gameplay and was able to win against Alex.

Then I had the finals vs LTB. I started with Claw, which was a bad idea, and got raped. I then went with Boxer where I had a new idea. My idea was to jump over his slow fireballs, but use jumping strong, which would hit his head if he started to walk up to throw me. If that didn't work I would immediately crouch and be ready to hit low fierce, because I noticed LTB liked to do an immediate jump RH to set up his throws. I figured even if that traded, it would put me in a great position of knocking him down. Plus more importantly, I wouldn't lose ground. My ultimate goal was to get him to be in the corner, where he likes to do jump up and jump back roundhouses. I figured that if I can get him to play that game, I've won -- it's just too hard for Ken to guess right every time.

I was able to do that and you can see the video here.

I'll follow up with a post about Evo, and then one about this weekend's Random Select Fall showdown.

Monday, August 4, 2008

Old A3 tournament report from Japan

While I was looking up on google groups for my old postings about Kurahashi, I stumbled on this:

Japan A3 Tournament Report

I thought it might be good to post for some laughs :)

Landing combos more consistently

One of the nice things about GGPO is that you can watch what other players do, and pick up tidbits that you haven't incorporated into your game yet. I was watching Hentaikamen, who is very strong (at least by my standards :) ), and who pretty much took me apart with his Dictator.

One thing I saw was that he was starting a lot of his jumpins with jumping jab. I thought this was a little bit wierd, as usually my Dictator jumps in with

a) Roundhouse, because it does more damage.
b) Strong, because I can hit once on the way up and once on the way down, for more chances to hit if my opponent jumps back at me.
c) Forward, because it has hitting frames all the way to the ground.

So it seemed wierd to me that this great player was using what appeared to be a suboptimal jumpin. Why was he using jab?

I experiemented with it, and I found that jump jab lands you much closer to your opponent than the other options above. This has a real effect on gameplay, because it's very easy to be out of range for your favorite combo with Dictator. Specifically, one thing I absolutely hate is when I jump in from the front, land my jump in, then follow with a standing short, low forward, scissor kick combo -- except the low forward whiffs out of range. That's really annoying and can often lose you the round in terms of opportunity cost.

When I tried with jumping jab, I never whiffed my low forward. So by using jumping jab, you're giving up damage for consistency, something that is totally worth it.

It might seem really obvious to some, but I'd never really thought of mixing up my jumpin attacks (Why use anything other than the "best" choice?).

It's the same "aha" moment I had when I first saw Kurahashi playing boxer in Japan

And saw him playing against Dhalsim too.

Kurahashi would just jumping strong (what I call his "curly punch") to jump in with instead of roundhouse, for the exact same reason. Jump strong lands you closer to the opponent, letting you do low jab x2, low strong, into a series of options (at least that was what Kurahashi mainly used). Back then, I would only jump in with Roundhouse, because it did the most damage. It took my seeing a superior player using something else for me to question my basic assumptions.

Looks like almost ten years later, I still haven't learned my lesson.

Saturday, August 2, 2008


Last week I discovered GGPO. I'm amazed at what I've been missing not playing on this. There's some really top caliber players there, including Graham and Alex Wolfe, Alex Valle, and some very good Japanese players.

There's also been nightly tournaments at 9pm PST, which are great practice for Evo. I find one problem on GGPO is that I'm always tempted to play my "fun" characters instead of my tournament characters, but the tournaments are a great way to force myself to play my real characters.

It's also very nice because you can watch other players. I highly recommend it for everyone, whether a new player or an experienced veteran!

Sunday, July 27, 2008

Leaning back with Boxer

Here's a little tip I learned from Graham Wolfe.

Boxer leans back really far when he's in the standing block position. His sprite is actually a few pixels further away than when he is crouching. So if you have your super charged, and are waiting for your opponent to throw a fireball, you can use this to your advantage to get a few extra frames.

What you do is sit in the crouching back position, as you normally would (to charge down and back). When you see your opponent throw a fireball, move the stick to the standing back position, and THEN do your super motion. You'll have a few extra frames to do the super before the fireball hits you, and a few extra frames is a big deal!

Friday, June 27, 2008

Pacific Northwest Summer Majors

Tomorrow is the pacific northwest majors! It will be a lot of fun! It looks like last winter's tournament attracted enough attention that we will have some Canadian players coming down.

Since the event is in Seattle, I'll be bringing my cabinet there. The thing is *very* heavy, but I have two friends coming to help move it. Sacha and I barely got it up the stairs when I first had it delivered over a year ago, but we were only two. Hopefully having three people will make a difference.

I'm still not sure who to play. It's been about a year since I started playing Dictator. He's by far my favorite character, but I still don't feel really "tournament worthy" with him. I still feel that when push comes to shove, I'm better off playing Claw or Boxer, my old staples. Case in point, at Evo I didn't play a single round of Dictator. Now if the tournament were on my cabinet, I might seriously consider playing him. I've found that I land my combos much more reliably on the cabinet. I'm pretty sure that there is a speed difference between Dreamcast and Arcade, and since most tournaments are on Dreamcast these days, but I only have the cabinet at home to practice, my timing is off. Combos matter much more with Dictator than they do with Boxer or Claw, so I really feel that my expected win percentage goes down a lot on Dreamcast.

I'm also interested in following the news of the ST:HDR beta on 360. I don't plan on getting the beta, but I'll be interested in seeing it at Zach's tournament tonight. From what I've read on SRK, it sounds like Ken is pretty ridiculous. That's funny, because Ken seemed really scary when I read about his list of changes :). Hopefully he will be toned down in the final build -- then again everything is changing so much that it's hard to find a point of reference. Who knows what "too good" means when everyone has been buffed up to the "too good" level anyways?

Wednesday, March 12, 2008

Random Select Winter Showdown

I've been meaning to write up a post about the Random Select Winter Showdown (Pacific Northwest Regionals), after Pablo wrote a post about the experience. I wanted to wait until videos were up, so that I could look at them, and analyze how I could do better.

First of all, I want to congratulate Pablo and Alex for doing an amazing job. Pablo went undefeated through the winners bracket until the finals. Alex beat RayBladeX, an excellent Boxer, and previous top 8er at Evo, not just once, but twice -- in the winners bracket and in the losers bracket. In fact, I think I owe the tournament to Alex. I played Ray Boxer v Boxer in the 5 on 5 exhibition match afterwards, and he trounced me. Alex's Guile has improved dramatically in the last 6 months. Congratulations.

Here is my winners finals match with Pablo (I am playing Dictator)

Here is the grand finals, with Alex (I am playing Dictator/Chun Li)

Reviewing these matches, one thing I notice immediately about my play is that I'm missing a lot of combo opportunities. There are several points where Alex whiffs a flash kick, where I should be able to do standing short, low forward, scissor kick. Even just low forward into scissor kick would be simple and damaging. Yet I find myself not doing this basic combo properly -- a combo I can do in my sleep in casual matches.

One way I can improve then, is to simply get better at execution. It's easy to fall into the trap of only practicing high end, difficult combos. But failing bread and butter combos like the ones I mention above can cost the match. I think my problem is that since I "know" that I can do them already, I don't bother practicing. But then when it's tournament time, I miss the opportunities I am given.

When I played in Japan, I remmeber playing against Kurahashi and getting *one* chance like that a game, if I was lucky. High end play requires maximizing these chances every time, and I need to do better there.