Hi everyone, I was wondering if anyone could answer a few questions I have about frame speeds in TGM series. After watching some of the insane grandmaster videos I decided to try and knockup my own tetris game that would operate at the same sort of speeds. Thanks to your wiki and some forums posts Ive built a fairly accurate mechanics and rotation system. However there is certain bits of info that I cant find and dont know if it is accurately known, basically Im unsure of how the game alters its speed after it hits 20G. How do the number of frames for ARE and lock delay decrease based on the games level? Also how does the amount of frames for DAS alter based on the game speed? And when using DAS how many blocks of movement can a piece move by in a single frame, and does this alter based on the games speed/level? According to the wiki the LockDelay is fixed on TGM1 at 30 frames (Which I thought sounded a lot, but after checking it emulated it does seem right), and gets shorter as the level increases in the sequels. After experimenting with some different values I would guess that it wouldnt drop below 5-10 frames (as it would be almost physically impossible set any lower). I couldnt find out what the starting values for ARE were, I reckoned about 15 or 30 frames to begin with then dropping down to as low as 5 frames (Would the game be possible with ARE set to 0 frames?). Anyway if anyone knows what the values are and how they change over different levels could you please help me?
The timings only change at level multiples of 100. And the game is certainly possible with ARE set to 0, though you'll need a longer lock time to make up for this. This removes the challenge considerably, and also ends up not pushing the player to be as fast. You'd also lose IRS. As for specific timing changes, sure I could list them... But why not try for something more original? Plug some values in and see what works.
yeah, tgm's a bit slow and restrictive by today's standards anyhow. (btw, i'd disagree with ct's assessment about how ARE pushed the player-- 0 ARE allows for faster play).
Also, TGM is not restrictive in general as TGM3 is fine. Faster than you anyhow. And I didn't say ARE pushes the player faster. I said short lock times push the player faster.
[20g + no ARE + Sega Rotations + infinity] = way faster possible records than [20g+ ARE + Sega Rotations + finite] that's all i'm saying. some may view one better than the other, but i prefer faster-- and by tgm, i thought he was asking about tgm1 specifics, not tgm3. and what's with the personal attack? faster than me? oh come on-- let's not go there. in addition, "And the game is certainly possible with ARE set to 0, though you'll need a longer lock time to make up for this. This removes the challenge considerably, and also ends up not pushing the player to be as fast" is very misleading. having less ARE does not mean you need longer lock times. you can still push the player with shorter lock times without ARE (moreso, i'd say).
It wasn't a personal attack. You're not fast enough to hit the upper limit of speed in TGM3... And nobody is yet so that's not an insult. By using the blanket statement "ARE is bad", and claiming this is true even for TGM3, you're really asking for more room than you need. That's all I'm trying to say. Your experience with tool assisted replays is skewing your perception of what is limiting.
but i didn't say "are is bad," i said are is slower. tgm3's fast b/c of it's lessened are times. near the end, there's nearly no are.
Yes, technically, you are correct. ARE limits performance. But if TGM3 is faster than you can handle, why complain? By the time you can handle it, TGM4 will be out. I guess I just interpreted your comment as more of a complaint than a mere statement, and there is simply no reason for anyone to be complaining at this point. Are people even breaking 4 pieces a second? Maybe they do in the more insane portions of the Shirase videos, but I don't think anyone has sustained that kind of speed for an extended period of time.
Yes. Nicholas sent me a link to a video of DTET play at what he described as 250+ ppm (left = cpu; right = keyboard).
technically YES, but in practice not really. Don't miss DAS. It is not because you can quickly stack 3-4 tetraminos on the center really quickly that you will be able to do the same for the side. Without ARE you still have to wait the DAS, your tetramino resting stupidly on the center. if you report in Ti this short time into ARE, the speed gameplay become the same but turned more esteticaly.
yeah, cpu's can go well over 1000ppm =]. fastest i've heard of is a player who pushed it to 230ppm max on a 12wide field, which is extremely impressive, and might translate to something like 250 in a 10wide field. the shirase videos might have a few moments where the player fit 4 (maybe more) pieces into one second, but definitely not for long periods. the last video's best consistent ppm was about 180ish for a little less than a minute. as for the das, it doesn't necessarily depend on are. in fact, the game i just mentioned with the 230ppm had zero are, but crazy das (my favorite das). one last thing to "But if TGM3 is faster than you can handle, why complain?" here's why: i've TASed tetris dx's 40 line mode and went 200+ ppm. now since i can't go 200+ ppm, does that mean i can't complain about tetris dx going fast enough? even though it's quite possible!
Though not humanly possible. TGM is a very hardcore game, yet it is finely tuned to be played by humans. It is both fast, and meant to be played fast. None of this "techinally possible to play fast" nonsense that hasn't translated into actual fast plays like Tetris DX. I'm sorry but that human isn't breaking 4 or even 3 pieces per second. The fastest reported Shirase time is about 3.26 Pieces Per Second (source: Mihara's blog). The fastest reported DTET time is about 2.94 Pieces Per Second (source: DTET website). I just did the calculation. The player averaged 3.08 Pieces Per Second over the whole game, or 184.8 ppm. This estimation is guaranteed to be slower than the actual performance. This makes it mathematically impossible for the best consistent speed to be as slow as 180 ppm.
"180ish" meaning 188 i think exactly. i timed the short period between like 900 and 1000. right before the green blocks. and how exactly did you calculate it? those garbage blocks can really upset an estimate. the best way to do it is to count the pieces exactly. as for dtet, the best record for 100 line clear is 1:15.15. the site says 261 blocks, which is 3.47pps. edit: i just want to go ahead and say it-- the above mentioned time record, advanced mode, plays without ARE. ;]
184 was the average, so 188 can't be the peak. (Well techincally yes, but just looking at the video you can tell the speed wasn't invariant enough for that.) I got my time much the same way I calculated that 3.26 in that one thread, except I was able to be more precise because we have a video. DTET: I was referring to the 300 lines in 4:15. IMO relying on a single minute is terrible. You're playing for that lucky game where pieces "just work" together instead of showing your true abilities. 750/255 = 2.94
i was comparing it to a section time (which is less confident than a 100 line run). it's over a minute of playing, and i'd say that's pretty damn good, and any other player wouldn't get that lucky in a million games. it just doesn't happen that way. i don't think it's fair to compare it otherwise, as tgm3 enjoys a much larger competition than dtet.
