...which seems to fit right in with what Alex was so patiently trying to explain to me. That makes a lot more sense now. Thanks for pointing that out!
This is something I have been working on correcting in my game. I have been burned so many times trying to move an S piece to the left side at level 19. If you take a look at Eli's maxout, specifically at 07:08 in the video, you'll see an S piece coming in right after a single line completion. I 'think' the stack (being 4 high on Eli's game) is too tall to attempt an S piece going to the left wall. He does it just fine, but I'm guessing he either has an extra tap at the end, or the flip that he does at the last moment is enough to get it into place. http://www.youtube.com/watch?v=sy-D8ByS3Kk My question to the experts is whether you have developed the extra tap technique, or would you be more likely to avoid the move to the left wall in that situation?
That's a tough call... I think you kind of have to go for it. Check out these diagrams -- if you take a conservative S placement, you're going to have to make a more difficult decision when L appears. It's manageable if you're very clever (or at least scared out of your wits enough to try something crazy), but if you do something standard and that O appears as it does in Eli's sequence, you're likely in some serious trouble.
Six frames can be a big difference on 19, as that amounts to three rows dropped. It is always best to get a wall charge whenever possible. I try to do it on every piece, whenever a wall charge is available, even if I do not need a wall charge. The wall charge means the next piece will make its first shift immediately, rather than one to six frames later, which translates into beginning your shifting two to three rows higher in the matrix. In other words, if you are moving an I all the way to the left side, and you get there with time to spare, even if it is only a couple of frames, you should let go of the left-button, and re-tap left. Now you have a wall charge. If not, you will find yourself somewhere between 10 and 16, rather than a full 16.
I may have mentioned this in another thread, but it is totally relevent for this thread, and what you noticed..... SNES allows for diagonals, and I believe the "odd" behavior of SNES DAS charge is due to "accidental" downs when we are pressing D-Pad left or right. 29-31 is intense, and there is a natural tendency to press the D-pad quite hard, and if you just *barely* nudge "down" at any point with your left or right during the first few frames after ARE, you will kill your 11 frames of pre-gravity, which would totally change the dynamics of how much "time" you have to get the piece all the way to the side. So I advise favoring the "up" side of the d-Pad. Just sort of move your thumb a tad upward and sort of think about avoiding the "down" section. I think Ben Mullen will back me up on this.
A couple of things to discuss on this. I think we've already discussed this before, but how did you come up with eleven frames? In my dissection of SNES Tetris, I found that the program added $0A to the counter. It's been ages and I don't have the hardware to double-check, but I am sure it was $0A. That's why I used $0A in my "SNES Level 29 Emulation" NES Game Genie code. And the code I referred to above seems spot on to me. Of course, how much difference could I tell between ten frames and eleven frames anyway? I could alter the game genie code to add 11 and try it out I guess The other thing I wanted to bring up is regarding: Ben Mullen and I have discussed this at some length, both of us having played a considerable amount of 29+ SNES, and it does seem that there is something weird going on. The biggie being that it sure as heck seems like clearing a line helps with DAS charge. I could swear that, all things being equal, clearing a line means your next piece is going to make all the way, while not clearing a line drops the probability considerably. Leastways, that how I used to think. Now that I've played SNES gravity rules on NES, I can say pretty confidently that it is an illusion. I've had very similar experiences with the pieces making it or not making it on NES. And I know nothing quirky is going on with NES. One frame gravity is just a bitch, that's all. Of course, my earlier theory on this had to do with "accidental downs"--that the player is accidentally hitting a diagonal--and that theory is still relevant. Levels 29-31 are intense, and I don't know about other people, but I tend to squeeze my controller too tight and push the buttons too hard, all which would invite an accidental "down". Clearing a line would give the player a brief respite, thus easing the stress factor, thus reducing the likelihood of an accidental down.
I got the number by framestepping through the game in an emulator. I think there are a couple possibilities here: 1) the piece is visible for an extra frame but only actively controllable for the next 10 frames before it starts to fall (1 visible + 10 active = 11 frames apparently in the playfield), or 2) the value written in the game is 0x0A = 10, but the code has an off-by-one bug or something else that results in it being controllable for an extra frame before falling (0 1 2 3 4 5 6 7 8 9 A = 11 values, so perhaps pre-gravity could be 11 frames). I think one of the potentially important issues is that neither 10 frames nor 11 frames is evenly divisible by the DAS period of 6 frames. I can think of some weird disadvantageous patterns coming from that. For example, if a piece enters with negative edge DAS (i.e., 10 frames on the counter, a full 6 frames until the next move) and assuming pre-gravity is 11 frames, then the piece will move once during pre-gravity and fall a frame short of moving again before beginning to fall (11 = 6+5). Every other charged DAS value (11~16) will move twice before the piece begins falling. And, of course, 1G is just a bitch. It is going to exacerbate DAS issues to levels otherwise unseen. You're rarely going to reach the wall or spend enough time there to build to full DAS naturally. Every cell the playfield grows taller is one less frame you get to charge DAS before the piece locks. Getting charge from wall tech on a 1 cell tall obstruction is frame perfect. You're going to be topping off DAS much less often, and the impact of having partially charged DAS is going to be much more severe.
