I actually checked it, and it really does seem punishable, slowest moves I could punish it with were Kitanas b4 and d4 at 15 startup. Anything slower than that like Cages f2 didn't seem to work, but of course my testing methods weren't exactly super scientific (trying to control two characters at once, wahey). I did find it a lot easier to punish it with moves with fast startup (actually something like Strykers 4 will work pretty consistently) so it probably wasn't just me accidentally releasing block.Lomyn said:
Really hate sounding like such a whiner, you guys are doing a fantastic job.
I did it deliberately for no other reason than to spite you, and only you.
Try doing an "instant" crouch block. A good test is to try crouch blocking Nightwolf's shoulder ram on reaction (or even in anticipation). You'll get hit every time.
I don't see how it's all that beneficial, so I don't think I'll be taking the time to calculate it, but if you'd like to do it I will add the information to the collected frame data.
No one is being left out. He's just not listed yet. As stated in the OP, every character will be completed.
The jumping method may or may not work, but the method I use to test hit/block data is confirmed to be 100% accurate (or at least as accurate as our execution frame data). So I will be using the methods I've already developed for this. It's a slower process, but I'd rather be confident in my calculations.I'm almost certain Sprint is right about not being able to attack before being able to move. I tried hitting p2 character with stuff, then either holding up or mashing some fast attack on the other controller and couldn't see at least any huge difference in when you could either jump or attack and can't really see why there would be any. Pretty ghetto test but meh, at least the difference in timing between when you can block and when you can attack was obvious. I even used almost the same example with jump punches about blocking in Kitana thread, just with Raiden, heh. I guess what I'm trying to say is you can release the data you have collected on hit/block advantages without having to worry too much about it being wrong
I actually forgot you already had a method for testing this, oops. Anyway, if the jumpy way gives even somewhat correct results it can be useful for people who can record gameplay but don't have a friend/gimp/hooker for testing reversals and such against eachother. I guess it's also a bit easier to do, then again I have no idea how much effort the whole recording-then-viewing-frame-by-frame thing takes. Which I'd guess is still a lot, heh.DrDogg said:
Not sure if you're trying to troll or what but the OP has stated several times now that this thing is a work in progress.
I believe this is what DrDogg is going to use for figuring frame data, which seems pretty laborous but if we assume we know what the start up frames are right it works. Of course you kind of have to take into account that it's impossible to do attacks at first possible frame consistently and all that (which is why it'll take a lot of work). Which is kinda why I like bringing up the "do stuff and hold up" method because while it might not give perfect results at least for crouching attacks it gives kinda good enough info for knowing whats obviously safe and what isn't. Clearly exact data is better than good guesses but I'll take anything that's available at this point.DrDogg said:
Yeah I'm curious to see how fast her roll is up close and at max distance.
Vulcan Hades said:
You mean command normals (forward +
)?
Forget that question, I wasn't clear.You mean command normals (forward +
Because they don't have it written down. It isn't programmed by frames apparently and just advantage/disadvantage.
Didn't know that universal moves like uppercut have impact frames that can vary on a character basis as much as by over 100% (i7-i15).