Posts for thatguy

1 2 3 4 5 6
20 21
Posted: 12/1/2016 9:12 PM
Quote
thatguy
Editor, Experienced Forum User
Joined: 11/3/2013
Posts: 506
Well, maybe they just want to get off the planet for the duration of Donald Trump's presidency ;)
Posted: 10/13/2016 5:49 PM
Quote
thatguy
Editor, Experienced Forum User
Joined: 11/3/2013
Posts: 506
A cute problem I came across recently - I believe I have a proof, which took me about an hour so some of the guys on here will easily get it in a coffee break. It concerns a remarkable but little-known property of the golden ratio, phi. Draw up a table with two rows and indefinitely many columns. Populate the two rows like this: Fill in the first row such that the nth entry in the first row is the integer part (aka the floor function) of n*phi. In the second row, put all the positive integers that do not appear in the first row, in ascending order. Prove that the difference between the two numbers in the nth column is n.
Posted: 9/25/2016 12:05 AM
Quote
thatguy
Editor, Experienced Forum User
Joined: 11/3/2013
Posts: 506
Quick question for hyper: why did you not use FFM? Was it a stylistic choice because you prefer to start from clean RAM? Or was FFM discovered after you had started the TAS? (Or maybe it was both?)
Posted: 6/30/2016 10:38 PM
Quote
thatguy
Editor, Experienced Forum User
Joined: 11/3/2013
Posts: 506
What are the odds that Fractal would ninja me by 17 seconds?
Posted: 6/30/2016 10:37 PM
Quote
thatguy
Editor, Experienced Forum User
Joined: 11/3/2013
Posts: 506
The number b is going to be something along the lines of b=a11a22a33... For the case where all the a in A are prime, this would be sufficient, as follows: d(b) = 2x3x...x(N+1) = (N+1)! where N is the number of elements in A. d(anb) = 2x3x...x(n-1)x(n+1)x(n+1)x(n+2)...x(N+1) d(anb)/d(b) = (n+1)/n, which is different for every n and therefore d(ab) allows you to work out a. I imagine a similar technique would allow you to generalise but I can't think how exactly.
Posted: 6/20/2016 11:14 PM
Quote
thatguy
Editor, Experienced Forum User
Joined: 11/3/2013
Posts: 506
Nice find! Obviously no speedrunning application but a fun little glitch. Considering how easy it is to perform in real time, it's pretty amazing it went undiscovered so long.
Posted: 6/20/2016 5:08 AM
Quote
thatguy
Editor, Experienced Forum User
Joined: 11/3/2013
Posts: 506
sheela901 wrote:
I promised myself to not watch a FF6 run because I want to play it before watching it.
I doubt this TAS will have spoiled too much of the story ;)
Posted: 4/19/2016 6:02 AM
Quote
thatguy
Editor, Experienced Forum User
Joined: 11/3/2013
Posts: 506
Obvious star material here. Probably speedy TAS of the year.
Posted: 4/10/2016 11:33 AM
Quote
thatguy
Editor, Experienced Forum User
Joined: 11/3/2013
Posts: 506
Another type of game in the "trivial" category would be games where everything takes a certain amount of time no matter what you do. A good example of such a game would be Rollercoaster Tycoon (my childhood!) where the goal in each level (or "scenario" as the game calls them) is improve your park to reach certain objectives by a certain date. Even if you reach these objectives early, you still have to play through to the date specified in the scenario, maintaining your scores (it is possible to lose money, guests or whatever else you are aiming for), and there is no fast forward setting. Thus completing a level would take the same time for anyone, as long as you actually pass the scenario. The only advantages a TAS could confer in a game like that is frame-perfect menus between levels.
Posted: 3/29/2016 5:08 PM
Quote
thatguy
Editor, Experienced Forum User
Joined: 11/3/2013
Posts: 506
I guess I have a particular TASer in mind for both of these but: $10 for a Pokemon HG/SS TAS, with another $10 for a time below 1:50. $10 for a Mario & Luigi: Superstar Saga TAS, another $10 for a time below 1:05.
Posted: 3/6/2016 10:30 PM
Quote
thatguy
Editor, Experienced Forum User
Joined: 11/3/2013
Posts: 506
Okay, saw this just as I was about to go to sleep. RIP getting to bed at a reasonable time. Sorry boss, I might be a little bit slow tomorrow. The only thing that could make this more awesome is if you had somehow managed to complete the co-op diploma using the same input on both games...
Posted: 1/31/2016 10:10 AM
Quote
thatguy
Editor, Experienced Forum User
Joined: 11/3/2013
Posts: 506
Is this the first game on tasvids to feature a FMV cutscene?
Posted: 1/27/2016 6:52 AM
Quote
thatguy
Editor, Experienced Forum User
Joined: 11/3/2013
Posts: 506
Nice work again solarplex... but it could do with having an encode that removes the loading screens, they probably take up half the runtime.
Posted: 1/21/2016 9:03 PM
Quote
thatguy
Editor, Experienced Forum User
Joined: 11/3/2013
Posts: 506
I think the likelihood approach is better - purely because Fractal is a far better mathematician than me.
Posted: 1/13/2016 7:13 PM
Quote
thatguy
Editor, Experienced Forum User
Joined: 11/3/2013
Posts: 506
I don't always agree with the agreements individual communities come to for specific games, but I do admire the democratic process by which those decisions are made.
Posted: 1/8/2016 10:30 PM
Quote
thatguy
Editor, Experienced Forum User
Joined: 11/3/2013
Posts: 506
Pokemon glitch exhibition was quite amusing. I really want to know what the setup at the end was all about...
Posted: 1/8/2016 10:26 PM
Quote
thatguy
Editor, Experienced Forum User
Joined: 11/3/2013
Posts: 506
p(green) = p(green|correct guess)p(correct guess) + p(green|incorrect guess)p(incorrect guess) = p_right*p(correct guess) + p_wrong*p(incorrect guess) = p_right*rand + p_wrong*(1-rand) randgreen = p(correct guess|green) = p(green|correct guess)*p(correct guess)/p(green) = p_right*rand/(p_right*rand + p_wrong*(1-rand)) Similarly, randred = (1-p_right)*rand/((1-p_right)*rand + (1-p_wrong)(1-rand)) For a sanity check, when p_right = p_wrong then randgreen = randred = rand which is what you would expect.
Posted: 1/8/2016 6:54 PM
Quote
thatguy
Editor, Experienced Forum User
Joined: 11/3/2013
Posts: 506
Here's what I would want my code to do: At each stage, pick a random four-digit number. The probability for picking each digit at each position would be weighted so as to reflect the information already received: if a digit has been tried before and has flashed red, it is less likely to be picked by the algorithm, and if a digit has flashed green, it is more likely to be picked (assuming p_right>p_wrong). Modify the algorithm to stop it guessing the same combination twice. Some pseudocode: int (rand(a), rand(b), rand(c), rand(d)) = {random integers uniformly distributed between 0 & 9}; int (a, b, c, d, n); a = rand(a); b = rand(b) c = rand(c); d = rand(d); n = 1; rand(a)green = {Bayesian formula for probability distribution of digit was of light turned green based on rand, p_right, p_wrong}; if (a = green), rand(a) = rand(a)green; rand(a)red = {similar definition}; if (a = red), rand(a) = rand(a)red; //similar expressions follow for b, c & d a(n) = a; b(n) = b; c(n) = c; d(n) = d; n = n + 1; bool (check); {check returns TRUE if code is correct, FALSE otherwise} for (i = 1, i++, n) if (a = a(i) and b = b(i) and c = c(i) and d = d(i)) goto line 2 else (if (check = TRUE) end else (goto line 2)); {the first if clause stops the program guessing the same code twice} I apologise if this code is unreadable or just doesn't make any sense; it's been a while since I did any coding and I was always rubbish at it.
Posted: 1/8/2016 6:03 PM
Quote
thatguy
Editor, Experienced Forum User
Joined: 11/3/2013
Posts: 506
I wish I had never mentioned the word "infinitesimal" now...
Posted: 1/7/2016 9:17 PM
Quote
thatguy
Editor, Experienced Forum User
Joined: 11/3/2013
Posts: 506
I really enjoyed Splatoon actually out of the few that I have seen. I struggle to think of any other game released so recently which already has such a high level of optimisation, and the run itself was of remarkable quality for a marathon setting (52:58 against a WR of 52:44). Looking forward to Bloodborne too.
Posted: 1/6/2016 9:56 PM
Quote
thatguy
Editor, Experienced Forum User
Joined: 11/3/2013
Posts: 506
Warp wrote:
"Infinitesimal" would imply that it has non-zero size.
I'm not normally pedantic but since you are started it... "Infinitesimal" means "greater than zero, but smaller than any real number greater than zero". A point of zero volume would have zero mass even it had infinite density. An point of infinitesimal volume and infinite density would have a non-zero finite mass.
Posted: 1/6/2016 9:33 PM
Quote
thatguy
Editor, Experienced Forum User
Joined: 11/3/2013
Posts: 506
Also Hawking radiation is not as inconsistent as you might expect: it comes out a weak-field approximation of GR, where QM is perfectly well-defined (otherwise how would it ever have been theorised considering it is a fundamentally quantum phenomenon). It does give rise to the black hole information paradox though: there is no information in Hawking radiation, so if information-laden matter collapses into a black hole and then that black hole evaporates away to nothing via Hawking radiation, where did the information go? (Another Nobel Prize on offer there).
Posted: 1/6/2016 9:29 PM
Quote
thatguy
Editor, Experienced Forum User
Joined: 11/3/2013
Posts: 506
It's somewhat of a misconception that the singularity is this infinitesimal point with infinite density. In mathematics, the term "singularity" is really synonymous with "undefined" - ie we don't know what happens at the singularity. Most physicists believe that there is some underlying Planck-scale theory which reproduces both QM and GR and would be able to tell us what goes on at the singularity. And if you find it you'd better start drafting your Nobel Prize acceptance speech.
Posted: 1/2/2016 5:43 PM
Quote
thatguy
Editor, Experienced Forum User
Joined: 11/3/2013
Posts: 506
Throwing in this nomination, which I remember really liking but most people seem to have forgotten: [2812] GBA Rayman: Hoodlums' Revenge by DaJaWi in 13:44.53 Not expecting it to win and I probably won't vote for it but thing it's worth throwing into the ring.
Posted: 12/26/2015 7:43 AM
Quote
thatguy
Editor, Experienced Forum User
Joined: 11/3/2013
Posts: 506
Paper Mario 64 is a great run, but suffers from the same problem as several other broken runs in that it basically repeats the same exploit over and over again. TTYD deserves the star. It is long, yes, it would be comfortably the longest star on the site; but: - Everyone loves the game - and most who come to the site would be familiar with it. - The commentary adds so much to the run. - It performs a great many different exploits throughout the course of the run. - It shows the advantages of TASing compared to real-time better; the TAS is practically an hour shorter than any real-time effort. - We need to highlight GameCube games more - we have a bunch of very high-quality and entertaining GCN runs now.
1 2 3 4 5 6
20 21