Posts for Flip


1 2
6 7 8
11 12
Experienced Forum User, Published Author, Skilled player (1535)
Joined: 7/25/2007
Posts: 299
Location: UK
Well there's 5 chars, 4 maps, so up to 20 different combos we could submit, which I think is too many for one game. At best, maybe you could link to various Mercenaries videos on the RE4 resource page, but I don't think they deserve a publication of their own.
Experienced Forum User, Published Author, Skilled player (1535)
Joined: 7/25/2007
Posts: 299
Location: UK
Not exist.
Experienced Forum User, Published Author, Skilled player (1535)
Joined: 7/25/2007
Posts: 299
Location: UK
Most nominations, but losing the poll :(. As a shameless attempt to promote myself, I'd like to remind all that the Warcraft II run was practically the 1st edition to the entire RTS genre, not just to that one particular game.
Experienced Forum User, Published Author, Skilled player (1535)
Joined: 7/25/2007
Posts: 299
Location: UK
Well to know what the perfect strat is for games, you'd have to know the expected values for these, IE what's the probability of him using each strategy, as well as the actual probabilities of switching during each strategy bringing an advantage. However, in this case that's super easy, considering it's either 33/66, or just 50/50. IE in case 2, switching won't make any difference, yet in case 1 it makes all the difference, so that alone shows you that it can only help, it'll never make things worse. But if that's not satisfactory, then say he has Probability p of doing first strat, (1-p) of the second. Exp(Switch)=(p)0.66+(1-p)0.50= 0.5+0.16p Exp(Stay)=(p)0.33+(1-p)0.5=0.5-0.16p Clearly for all p, we have E(Sw)>=E(St), so it's always beneficial to switch in this game, regardless of if he's trying to do it random or not. It'll either increase your chance, or equal your chance, but it can't ever decrease. Lets play another game, there's 10 cigarettes, one of which is poisonous. I offer you one, you either smoke that one, or choose a different one. If I told you my strat was that I want to kill you, I'm sure you'd agree the best strat is 'pick another one'. If I told you I grabbed one at random, I'm sure you'd agree there'd be no real benefit to switching. Surely it's obvious that depending on what the host's intentions are, it'll affect what decision you should make? With the host having different intentions, certain combinations will or will not happen, thus by excluding them from the set it'll affect the probability of the remaining situations occurring. Accept it.
Experienced Forum User, Published Author, Skilled player (1535)
Joined: 7/25/2007
Posts: 299
Location: UK
Let's say there's Goat A, Goat B, and Car, so can use 'ABC' notation conveniently enough. Let the first letter be your choice, the second the revealed door, and thus the third letter what you could choose to switch to. Obviously the basic combinations are: ABC ACB BAC BCA CAB CBA The problems stem from conditional probability, trying to keep track of which remaining possibilities/combinations we're actually talking about here. There's 1/3 chance of picking A first say, and thus 2 remaining doors to play with. However, this is where the two games diverge, since normally the host never shows you the car. This means that given A, there's a 0% chance of him doing 'ACB', and thus 100% chance of it being 'ABC'. If we choose C first, then he can show us either A or B, with probability of 50% each, thus: ABC=1/3*1=1/3 ACB=1/3*0=0 BAC=1/3*1=1/3 BCA=1/3*0=0 CAB=1/3*1/2=1/6 CBA=1/3*1/2=1/6 Good: ABC+BAC=1/3+1/3=2/3 Bad: CAB+CBA=1/6+1/6=1/3 Those give the standard figures. For your alternative game, we no longer deliberately discard 'ACB' and 'BCA' from the start, because he could accidentally show you the car now. This means that given A as your first choice, it's no longer 100% that he'll show you B next, but he may also show C, with probability 50% each. This changes it from the standard 100/0 split the game had before, to 50/50. This means the probabilities of each set occurring are no longer biased against one particular set, giving us a normal distribution of ABC=1/3*1/2=1/6 ACB=1/3*1/2=1/6 -Discarded after BAC=1/3*1/2=1/6 BCA=1/3*1/2=1/6 -Discarded after CAB=1/3*1/2=1/6 CBA=1/3*1/2=1/6 Good: ABC+BAC=1/6+1/6=2/6=1/3 Bad: CAB+CBA=1/6+1/6=2/6=1/3 Here we have P(Good)=P(Bad)=1/3, with the remaining 1/3 being discarded, IE restarting the game. So given that we actually had a successful game, we have equal chances of switching being a good or bad thing, thus the 50% figure you got. Why is that counter intuitive? Well instead of the host choosing to get rid of a bad prize, he just gets rid of a prize at random, then of course that's gonna lower your chance of winning isn't it? Would've thought that's expected.
Experienced Forum User, Published Author, Skilled player (1535)
Joined: 7/25/2007
Posts: 299
Location: UK
Well if "To the Vines" is slower, why does the Time column say 237, yet the Old Time says 454? To me, that sounds like the new version is faster, since it took 217 fewer frames.
Experienced Forum User, Published Author, Skilled player (1535)
Joined: 7/25/2007
Posts: 299
Location: UK
Yes vote obviously, but I'm confused at the values in the table. Isn't "Difference" supposed to be just "New Time minus Old Time", in which case you should have mostly negative values? If you want "Improvements" as the column, then obviously you do "Old minus New" instead, giving positive values for frames saved. However, there seems to be inconsistency in your column. EG "To the Vines" does New-Old, giving a negative number from improved frames, yet "Reaching Basement 1" does Old-New, this time giving a positive number for frames saved. And apparently for Reaching Ganon Cutscene, 201-197=3?
Experienced Forum User, Published Author, Skilled player (1535)
Joined: 7/25/2007
Posts: 299
Location: UK
In future you can ask at the Newbie Corner, or on N64 Emulators page for questions like these, since topics on this forum are generally reserved for discussion points for specific games. But to save your work, just simply go to Movie->Stop Moving Recording. That will stop the movie on the current frame, saving it for later. Obviously just 'play movie' to see what you've done up to this point. To continue recording your TAS, play the movie, but do NOT load it as 'read only' mode; otherwise, it'll be for viewing purposes only. Loading a savestate while in 'read only' will simply skip ahead to that point and continue playing the TAS, but if it's not 'read only' mode, then loading a save state will allow you to continue working from that particular point. And for TASing OoT, good luck. There's an ongoing discussion in the forum about what would even constitute 100%, let alone coming up with a decent route.
Experienced Forum User, Published Author, Skilled player (1535)
Joined: 7/25/2007
Posts: 299
Location: UK
How about this for a potential solution: Don't publish it as a new record for its own category, but instead have it linked as one of the movies that the current run obsoletes. IE each published runs form a chain where it links to the previous record, well how about wedging this one in the chain between the two publications whose times it beats/fails; as though it was a proper submission at the time. That way you could still get some credit, even if you have to search to find it. Then again, that would basically open floodgates and allow arguably any run to be accepted, regardless of if it beats the record or not; since you could just stick any new movie just before the one it couldn't quite beat. So maybe not, you'd need well defined tiers of when to officially accept a new run, since you don't want an arbitrary 10 star run being labeled as the one which obsoleted the 16 star route.
Experienced Forum User, Published Author, Skilled player (1535)
Joined: 7/25/2007
Posts: 299
Location: UK
Entertaining, but not publish worthy, as the current route is better. And as has been previously stated, why did you do rupee collection in Kokiri Forest if you got 100 rupees from BotW anyway? Arguably the current publication is a "Ganondorfless" route to get to the ending, so should we have another category which doesn't skip it? Neat as a concept demo to show off what the current method for BotW/Dodongos' Cavern is, but I don't think it's worthy of being published.
Experienced Forum User, Published Author, Skilled player (1535)
Joined: 7/25/2007
Posts: 299
Location: UK
Bumps an old topic for no reason? WTF?
Experienced Forum User, Published Author, Skilled player (1535)
Joined: 7/25/2007
Posts: 299
Location: UK
So we have a cube with N columns, trying to simulate one with n columns. For N=4, we use grouping G={2,2} to simulate n=2, and G={1,2,1} to simulate n=3, with obviously G={1,1,1,1} being the default N=n=4; Where |G|=n. Clearly this is just ways to split up our integer N into different sized sets, subject to certain conditions we hope to derive along the way. For example, we cannot use 4={1,1,2} as an emulator for n=3, because on the left side we have a face by itself, but we on the right we have the face paired off with an edge section. This means that when solving it, it'll eventually be split up. EG the two (Face, Direction) rotations: (Bottom, Left), (Left Up) would separate part of the 2x2 simulated cube initially in the bottom right corner. This is because of the lack of symmetry in the {1,1,2} expression. IE if we have tiles 'a' and 'b' in the 3rd emulated set (--,--,ab), after a rotation we have (a,b,--), which will then be separated upon moving either column. So we need each tile to be paired with just as many pieces on one side as the other, thus our given emulator distribution MUST be symmetrical. For all N>3, we we have N={1,(n-2),1}, which is of course symmetrical and has magnitude 3, thus an n=3 cube can always be emulated for any N. For N=5, we also have G={1,3,1}={2,1,2} for n=3 emulation. We cannot have n=2, because we have 5 columns to pair off to our 2 emulated sides. Since we cannot split this evenly, it is not possible. Now in terms of symmetry, it depends on if we have an odd or even number of columns, thus if we have a central column or not. IE for n odd, we need its central column to be emulated, done differently depending on if N is odd or even. If N is even, we can do this by having equal amounts of columns adjacent to the central divide as being treated as a single block. EG {1,2,1} gives us an emulated central column, so n=3 can be done inside N=4. Note here that the size of that central column (2) is even, but if we have N odd, then the central column in N does not get split up, but instead needs to be bunched with equal amounts of columns on both sides, thus we would have the size of the middle string being odd. EG for n=3 inside N=5, gives G={1,3,1}={2,1,2}, where both middle columns have odd numbers. Hence we have a solution for N odd. Either way, n odd can be emulated in both N even and N odd. But for n even, we need to split the columns of N into an even amount, one which is clearly impossible if N is odd, thus N also needs to be even. Thus N odd can emulate only n odd, N even can emulate n both odd or even. This is true for All N>n, for example: N=4. {1,1,1,1}, {1,2,1}, {2,2}. N=5. {1,1,1,1,1}, {1,3,1}={2,1,2}. N=6. {1,1,1,1,1,1}, {1,1,2,1,1}, {1,2,2,1}={2,1,1,2}, {1,4,1}={2,2,2}. {3,3}. N=7. {1,1,1,1,1,1,1}, {1,1,3,1,1}, {1,5,1}={2,3,2}={3,1,3}. N=8. {1,1,1,1,1,1,1,1}, {1,1,1,2,1,1,1}, {1,1,2,2,1,1}={1,2,1,1,2,1}={2,1,1,1,1,2}, {1,1,4,1,1}={1,2,2,2,1}={2,1,2,1,2}, {1,3,3,1}={3,1,1,3}={2,2,2,2}, {1,6,1}={2,4,2}={3,2,3}, {4,4}. So Odd cubes can only emulate smaller odd cubes. Even cubes can emulate all smaller cubes. Basically because the emulation will take the symmetric form of either G={---k---,---k'---} or G={--k--,m,--k'--}; thus giving total group size=emulator size of the form a+a=2a=even, or a+b+a=2a+b which can be made even or odd.
Experienced Forum User, Published Author, Skilled player (1535)
Joined: 7/25/2007
Posts: 299
Location: UK
There's 60C7 total hands, with (60-n)C7 of these containing no Land cards. The probability of choosing a hand with 0 Land cards is thus given as P(i=0)=(60-n)C7/60C7. For P(i=1), we want 6 of the (60-n) regular cards, and 1 of the n good ones. Thus (60-n)C6 * nC1 ways of constructing the hand. Totally probability P(i=1)=(60-n)C6*nC1/60C7. Generally, in a hand of 7 cards the chance of having i out of n good cards, is given by P(i)=((60-n)C(n-i))*(nCi)/(60C7) 1) Previously "n" was how many there are in deck, not the specified amount you want in your hand, so be careful of notation. 60-24-4=32 other cards. There's 24Cn ways of choosing which Land cards you want in your hand. There's 32C(7-n) ways of choosing the remaining (7-n) cards from the 32 boring ones. Thus P(i=n, A=0)=(24Cn * 32C(7-n) ) / 60C7. 2)Similarly, now there's 28 boring ones, thus P(i=n,A=0,B=0)=(24Cn)*28C(7-n)/60C7 3)So we want n Land cards in our hand, which had 24Cn combinations. Now in our remaining (7-n) cards we want A not to be equal to 0; where the probability is given by P(A>0)=P(A>=0)-P(A=0). We have (7-n) cards left, choosing from the remaining 60-24=36 cards. There are 36C(7-n) total hands, with 32C(7-n) of these not containing any A cards. Thus there are 36C(7-n)-32C(7-n) hands containing at least 1 A card. This gives: P(i=n, A>0)=(N selections * Remaining A>0 Selections)/Total Hands=(24Cn * (36C(7-n)-32C(7-n)) / 60C7. 4)As before, we have 24Cn combinations for the first n cards. We have 36C(7-n) total combinations for the remaining cards, with 28C(7-n) of these containing none of the 8 A or B cards. Thus this gives us 36C(7-n)-28C(7-n) combinations, which do contain at least 1 A or 1 B. Thus we have a very similar formula: P(i=n, A+B>0) =(24Cn * (36C(7-n)-28C(7-n)) / 60C7. 5)Needs more variables for the amount of red cards j, where 0<j<(7-n-1), giving at least 1 free space so we can have k>0 blue cards too. We have k<(7-n-j), giving at most (7-n-j-k)>=0 remaining cards to be drawn from the 28 remaining boring cards. Proper formulas inc another day.
Experienced Forum User, Published Author, Skilled player (1535)
Joined: 7/25/2007
Posts: 299
Location: UK
There's a video I saw many years ago, but sadly can't find now, of somebody doing all 4 sonics with the same input, IE 1+2+3+S&K. They only did 1 level, but naturally it's tougher than just those 2; since about 80% of the time you only really needed to control just one Sonic.
Experienced Forum User, Published Author, Skilled player (1535)
Joined: 7/25/2007
Posts: 299
Location: UK
There's a whole selection of "vanishing puzzles" [Url=http://www.marianotomatis.it/blog/index.php?post=blog/20110715]here[/Url], all showing different amounts of an object after rearranging the pieces. They're all based on the same principle though, of the new arrangements just simply being incomplete.
Experienced Forum User, Published Author, Skilled player (1535)
Joined: 7/25/2007
Posts: 299
Location: UK
Insanity is playing the same TAS over and over again and expecting different results. -Albert Einstein
Experienced Forum User, Published Author, Skilled player (1535)
Joined: 7/25/2007
Posts: 299
Location: UK
I'm sure 99% of TASing is done on ROMs from games you don't own. It's as illegal as downloading a song is, or taping a TV program instead of purchasing the DVD from the store. IE nobody gives a toss really, it's not as though if the government stumbles across this site we're all in for 20 years imprisonment is it? But obviously if people just download free roms instead of purchasing the actual game, it's not good business for the publisher; so it's not the kind of thing we'd directly say how to do. But it's not as though it's difficult to work out, Google will quite easily guide you in the right direction. I can assure you obtaining roms is not the big deal you're making it out to be.
Experienced Forum User, Published Author, Skilled player (1535)
Joined: 7/25/2007
Posts: 299
Location: UK
Well you could just be happy that a game you like has been pushed to its limits even further. If they made the movie from scratch, then surely that person can claim to have put in just as much effort as you did; if not an even greater amount, and as such should still get published. If they just hex'd in the improvement from the other author, then I would consider that unethical and lazy. But you don't seem to realise that the original author could do the exact same thing, and so change his original movie or create another submission. Then there'd be the question of "Original+Improvement VS Clone+Improvement", and whose movie should be published? Well clearly people would vote for the original author to be published, resolving any ethics issue. And if that doesn't work, then just simply use this lack of justice as motivation to create another TAS. Comb over ever nook and cranny of the game until you can finally find an improvement and get your name back on that TAS like you deserve it.
Experienced Forum User, Published Author, Skilled player (1535)
Joined: 7/25/2007
Posts: 299
Location: UK
Well in terms of obvious significant improvements, look no further than our favourite classics SM64 and OOT. SM64: Method for getting to Bowser 2. [Url=http://www.youtube.com/watch?v=ZV7nJwt2zhc&t=22m15s]70 star=30 stars+DDD Needed[/Url] (Stolen from BLJ-less TAS, as 16 star route was already known for very first movie) [Url=http://www.youtube.com/watch?v=4ZOBdU_5zlA&t=10m55s]16 Star=MIPS+DDD Needed[/Url] [Url=http://www.youtube.com/watch?v=vxJ1k7tBpNM&t=2m38s]1 Star=DDD Needed[/Url] [Url=http://www.youtube.com/watch?v=QDkxRA7yxLY&t=2m21s]0 Star=Skips all[/Url] OOT: Method for accessing Adult World [Url=http://www.youtube.com/watch?v=sVR32jXj68w&t=1h17m40s]Grabbing all 3 spitual stones[/Url] [Url=http://www.youtube.com/watch?v=0VEfQo8hXbs&t=11m25s]Door of Time Skip[/Url] [Url=http://www.youtube.com/watch?v=_d738pTAmOQ&t=13m]Wrong Warp[/Url] No doubt that's the thing you need, and I don't think any other games come close to having as many revisions as these 2 do.
Experienced Forum User, Published Author, Skilled player (1535)
Joined: 7/25/2007
Posts: 299
Location: UK
I'd like to see some tool assisted writing appear on screen, as there's a few movies which incorporates that. For example, if my recent Warcraft II runs doesn't get accepted for star material, then I'd at least like it to be featured in the promo video. Warcraft II Nightfire Goldeneye (Old) Brain Age Plenty of LOLs and TASs to shove in the promo somewhere.
Experienced Forum User, Published Author, Skilled player (1535)
Joined: 7/25/2007
Posts: 299
Location: UK
GC and or Wii emulation is still a huge pain, goodluck getting Wii U stuff to work.
Experienced Forum User, Published Author, Skilled player (1535)
Joined: 7/25/2007
Posts: 299
Location: UK
Yes, but it wasn't of much use in later levels, and I for one am not familiar with how to make Lua stuff so I couldn't make a better one. It often drew a wrong sized minimap over the existing one, which is not much help trying to locate stuff. It misses out at least half of the units on the map, plus draws wrong sized boxes over many of them. EG Critters have a 4x4 white box around them, despite being 1x1 units. Similarly, towers should be 2x2, but are 3x3. etc Plus since all the memory values weren't fully decoded just yet, buildings are labelled as 'type 5' etc, as opposed to us knowing their real names. If you wanted to check out what a particular building/unit actually was, you could just go send somebody over there and find out yourself, somewhat defeating the point of the script in the first place. As neat as it is, it's not much use in the end. It may tell you where are few things are, but any person hoping to TAS a level should know what to expect anyway. The only random elements worth tracking with such a script would be Gryphons in level 14, since all other units are well behaved; but this script didn't even locate them >_< . And to answer other questions, other emulators suck for this game, plus in this version there is no payload to clicking either critters nor units multiple times. Human missions will be TASes next, but not starting them just yet.
Experienced Forum User, Published Author, Skilled player (1535)
Joined: 7/25/2007
Posts: 299
Location: UK
Same ignorant BS from people who don't understand the concept of TASes, that it's somehow 'cheating'. Cheating implies deception, which there is none of in this community, we always point out that our runs are not real time, but are TASes. We're not trying to show off human skill, we're demonstrating how fast the game will allow us to complete it; which often involves searching much harder for solutions than a typical real time run would involve.
Experienced Forum User, Published Author, Skilled player (1535)
Joined: 7/25/2007
Posts: 299
Location: UK
Right now the best I can do is create a .7z archive, but apparently the submission instructions specify that it absolutely must belong to a .zip format, is that right? Edit: Never mind, just confused about files extensions, problem solved quickly
Experienced Forum User, Published Author, Skilled player (1535)
Joined: 7/25/2007
Posts: 299
Location: UK
-"Read & Write Mode" means you can resume recording on a movie by loading a savestate, instead of just watching it. Just watching is called "Read Only" mode, and if you load a savestate it will simply play the rest of the movie from this time, but not record anything new. -I don't think there is an option to rewind frames, you can only go forward in PSX. That means if you want to make a new savestate from a few frames earlier, you instead have to play the movie from the start, and then make a new state when you get to the right bit. Afterwards, then resume recording from this new save state. It is strongly recommended that you use way more than just 1 save state, otherwise it's highly unlikely that your TAS will have the standards required for this site. It's not hard to use more than 1 save state is it? Just save parts of your run to slots 2-9 too, use as many as you need to test each section. Using just 1 is bad, since it assumes that your current attempts will automatically be the best, and that you won't need to ever go back to an earlier section to redo something. Using more than one is far more efficient, and will make your TAS much better in the end. Basically, make a new save state each time you need to make a decision, or at anything which needs to be tested. You'll need to check to see if something actually works, and then reload to either accept that input or try another one; thus a save state is needed. There's no need to use 6 different save states when navigating through menus for example, I'm sure 2-3 would easily be enough. 1 could be the very start of the menu, and 2-3 just the individual states you use for testing each frame. Once you're sure the menus have been done properly, you can then change state 1 to be the start of level 1 for example, and then use 2-9 for whatever you want. But seriously, you won't get anywhere with just 1 save state, use more. -No idea about this problem, but it sounds like you're not resuming record properly? But that shows what I mean when you make a 'bad save state'. If you had a problem with save state 7 or so, you could just reload from 6, or 5 for example. With only 1 state, you'd have to start the movie from the beginning, and watch 10 minutes just to get where you want. If you were using more than 1 state, you could play the movie, and then just load 1, or 3, or 6 etc, and then you would only have to watch 1 minute of your run instead. Clearly more save states are better, and you can get to the bad section a lot faster. But just to make sure you're trying to resume record properly, this is what you should do. Play your movie (read only mode should be on by default). If you have the savestate you want to resume record from, then go into File->Movie->Read Only, (or press shift+8), and then load this savestate. This will resume recording from this point. If you do NOT have the savestate you want to record from, you need to create one. So play the movie (Read only), and then when the section is coming up, be prepared to make one. Trying to make a new savestate at the correct time while the movie is playing is kinda silly, so you'll need to slow it down to make sure. Try using pause/frame advance when the section is coming up, to make sure you get to the right frame, and then make a new save state. Then put Read Only mode off (do Shift+8, or click in menu) and load this new state, and you can resume recording from it.
1 2
6 7 8
11 12