I came across this on nico video. Some people here probably would have seen this. Posting for those who haven't. This site requires registration to watch videos. http://www.nicovideo.jp/watch/sm7678617
This is ZAB's most recent SHIRASE (CLASSIC) time attack record, from a few months back - S13 green line in 4:09:26. For those of you without a nicovideo account, you can use this link: http://d.hatena.ne.jp/video/niconico/sm7678617 It's a shame that it's a green line as he fails the BIG roll at the end, but I know that ZAB also has performed an orange line S13 in 4:11:46. Lee_n has the higher quality .mp4 hosted on his site, but I'm not sure if it would be appropriate to share the link publicly without both ZAB's and lee_n's permission.
I had my money on that as well. And his username just looks like a bunch of random letters. I almost felt disappointed that it wasn't spam.
I did some calculations by counting the line clears and the stack left after each section. Also accounted for the rising garbage by counting garbage line clears as 0.1 of a normal line clear. It's entirely possible that there might be some mistakes in reading the values, as I did this only in a couple of hours and no one else double checked the values. However, the number of tetrominoes dealt (second last column) in each section did turn out to be an integer in each case. Obviously this is only a verification and not a proof, but anyway ... The columns that are not highlighted with gray color mean that they have been entered manually. The gray colored columns are the ones that are based on a formula.
wksrdg, thank you so much for taking the time to make such an awesome spreadsheet! I had wanted to do the exact same thing myself, but could never find the drive to even start such an arduous task! If you don't mind, would it be alright if someone were to put the data in a google doc so that those of us without Microsoft Excel can view it too? Ignoring the first 3 sections where there's substantial spawn delay, ZAB sustains an average of 217.8 tpm for well over 2 and a half minutes from 300-1200. That's quite some endurance! Also, despite playing at what seems to be a blazingly fast speed, he didn't earn even a single ST Gold medal. Evidently, none of his section times were as fast as his previous bests set on that machine. It's possible to estimate what sort of speeds he must have been playing at when he set those records by assuming that the number of tetrominoes dealt in a section is 55-65, and using jin8's unofficial ranking page. I estimate that on a good day, ZAB is sustaining 256-303 tpm for entire sections. That's erm... pretty fast!!! Mihys (the author of DTET) did a similar study a few years back, using what was at the time the fastest known Shirase video. You can see the study on Mihys' website here: http://dtet.web.infoseek.co.jp/dtet/index.html (look on the left hand side, a little way down).
Yeah I don't mind this at all. I am just unsure what exactly are google docs Either way, I will try to re-check the data (for a possible mistake) later, when I find some free time. Well good section time depends heavily on what kind of stack you have when you enter the section. A simple two or three extra tetrises can boost up the level a lot and hence decrease the tetrominoes dealt. The level rewards in this game seem to be quite high. Also, it's worth noting that about 0.6 or 0.7 seconds are lost in each section by line clear delay. Now since Line ARE and normal ARE are different, if you include the lower line ARE, it comes down to 0.3 or 0.4 seconds. Just have to multiply the instances of line clear with frames lost at each instance. If you just lower ARE by 1 frame it saves about 1 second from a section further. That's assuming the player can handle the lowered ARE.
Here's the google doc for people unable to open Microsoft Excel files: http://spreadsheets.google.com/pub?key=tYERLT-cmd5GRjDP-Kv5bPA&single=true&gid=0&output=html If wksrdg or anyone else wants editing priviledges, (you'll need a google enabled email account), let me know and I'll send you an invite. Yes, starting a section with a high stack and finishing low would mean that a lot less tetrominoes would be needed. It would be interesting to find out the mean and standard deviation of number of tetrominoes required, but that would require examining a lot more games. I thought 55 would be a reasonable lower end estimate though. I don't know, maybe 50 might be better.
Well there is no substitute for real figures. However, some simplified calculations can be instructive. For example, assume you have tetrominoes for perfect stacking at your disposal. If we use the level bonuses as 1,2,4,6 and assume optimal stacking, then we get the following figures. I have only calculated a few values here. 19x9 - 48 (4+40+4) 16x9 - 50 (4+40+6) 15x9 - 51 (3+40+ 12x9 - 53 (3+50) 10x9 - 56 (2+50+4) Edit: Too easy to make a mistake by 1 lol.
