After compressing to lossless, you should use anri to properly compress the run to save space. Lossless is huge and it isn't going to get any smaller if you redo the run :/
Find a way to send me the wmv files, and I can encode it for you.
SDA Forum
No, I have no desire at the moment to find a way to convert and store everything. The beggining is a bit ugly, and I probably could've manipulated more Quizmo's into specific areas. Also there is a ~5 min time saver at the end that was found just after I completed the run. I hope someone plans to fix the run up, and anyone is welcome to my playlist as a guide and is welcome email me for elaboration/help. I don't have any plans to work on it for quite some time personally though.
If anything I know I can do better, NSMBW I've gotten pretty decent at with enough practice, and the 27 min run I thought was "good enough" to make me try something else, but now I just keep attempting to get a run that doesn't have a load of fail everywhere. I should've practiced more for Paper Mario, I was completely prepared with the route and theory, but I think the execution could step it up a notch.
Getting a sub 5:30 should be easy enough, but I think if I come back to this (or whoever takes it up) the 5:04 goal will be in trouble >:)
If anything I know I can do better, NSMBW I've gotten pretty decent at with enough practice, and the 27 min run I thought was "good enough" to make me try something else, but now I just keep attempting to get a run that doesn't have a load of fail everywhere. I should've practiced more for Paper Mario, I was completely prepared with the route and theory, but I think the execution could step it up a notch.
Getting a sub 5:30 should be easy enough, but I think if I come back to this (or whoever takes it up) the 5:04 goal will be in trouble >:)
Main Work!!!
If your ready to start your run back, I'm currently planning a SS any% of this. My goal is sub 3 Hours(But it will probably be around 3:20(it's now possible with the chapter 7 improvement of swimming on earth with Sushi in Shiver city and also a couple of better planning trick than the current segmented run)). at the moment I'm planning a couple of based luck thing but quite limited(they are more in the beginning, it would be sad restarting at bowser because I didn't get lucky with close call 
). I telling you this because I would like you to help me and I know that when there someone else running the same game as you but in a different category, we become more motivated(i.e. my amnesia SS run with Apjjm segmented run(my first goal was sub 45 and now I'm currently trying to sub 30 and this is all done by mutual help)) so I would be please to work with you Raiker if it's fine with you and if you have the motivation to restart because you planning was awesome. I know that it's hard to restart from the beginning but it will just show that you're devoted to this run and at which point you are serious in speedrunning. If you agree with this plan, I'll be posting my chapter by chapter planning once they'll be done. anyway I have a big interlacing problem so I have a lot of time for my planning until I find a issue for this.
Try asking 0xwas in an email, it's a Bowser Castle glitch, if not I'll make an attempt sometime this weekend, and make a video. I can't remember if it was a nicovideo, youtube, or ustream recording or not.
Andre: I'd love to help and see what you have planned, I'm not going to try fixing my 100% because I've been playing around with NSMBW lately, and I'm currently a little busy trying (and failing) to find a job, I recently moved and I need money more then I need to start up another 100% run of PM along with my fun-runs of NSMBW. I'm also amazed at the time you have planned for a SS, I really hope you have the chance to get sub 3.
Andre: I'd love to help and see what you have planned, I'm not going to try fixing my 100% because I've been playing around with NSMBW lately, and I'm currently a little busy trying (and failing) to find a job, I recently moved and I need money more then I need to start up another 100% run of PM along with my fun-runs of NSMBW. I'm also amazed at the time you have planned for a SS, I really hope you have the chance to get sub 3.
Main Work!!!
That's genius!
I don't know if 0xwas is lurking this thread but if yes neat find and thank!
And Thank to you too Raiker for recording this and posting it here.
Do you think it would be better if I start posting here or in a another old PM thread. If I start posting here, we could ask an admin to change the thread name unless nobody care about a PM any% being talk in a 100% thread.
I don't know if 0xwas is lurking this thread but if yes neat find and thank!
And Thank to you too Raiker for recording this and posting it here.
Do you think it would be better if I start posting here or in a another old PM thread. If I start posting here, we could ask an admin to change the thread name unless nobody care about a PM any% being talk in a 100% thread.
Main Work!!!
Because you can skip a part of the Shiver city, the snowman puzzle and I'm currently trying to skip the whole Ice palace. Also, because the current run is 3:38 and the planning is awesome but I found some little trick that can be better planned BUT he use a lot of luck base thing that make him faster and segment make him more precise. I though about Sub 3H but I said I would probably get nothing under 3:20 in reality. It's just that I know this game can be more broken than that. I.E. The sushi thing can only be used in Shiver city but the third jump can be more exploited than that, if there's a character that walk near a wall, you can jump at a precise moment and get through this wall and while you're traveling OOB with this, you can change zone, the only problem with this is that I only found how to use it in toad toad but it's sure that this can be exploit somewhere else. The new bowser castle trick save about 2:40 to 3:15. I'm sure other tricks will be found either by someone else or/and me during the planning.
Yes, I thought it was an amazing trick myself, and you'd be better off reviving an old thread then using this one, but if someone changes it to a PM general discussion who are we to argue 
Sushi thing saves A LOT more time in any%, you skip Starborn Valley completely (100% you still need to go there), you skip the backtracking to the mayor (again 100% you need to go there), the snowman mini cut scene, irritating enemy dodges, the Monstar fight, and the duplighost fight. I think in an any% it can save 2-5 min, I know it saved 1-2 min for me.
The spin jump thing could only be used AFTER Lakilester is gathered, and getting spin jump in chapter 6 is a waste of time unless you can sequence break like 10 min (which I doubt) with that trick there. Chapter 7 likewise has nothing for the spin jump trick that I could find after a quick check, but then again with the Sushi clip, its not saying much. I haven't explored the later part of Bowser's castle for it yet, but I don't think there's another instance where you can use this trick for a huge boost of time, I'd have to get back to you with that.
Now for whatever clip you found Andre, vid please, this sounds interesting, toad town is directly connected to chapters 1, 3 and 4, I'd like to see how you do this and what you can do with it.
Also random thing I found for a SS, you skip the white clubba text in crystal palace if you use kooper's overworld attack, you don't get the cheap shot, but if you use mega quake, and his power/fire shell there is a chance you can 1-turn them, and if you time it right you can level up after the first or second one, you can beat all three relatively quickly. (That or star storm + fire/power shell)
Sushi thing saves A LOT more time in any%, you skip Starborn Valley completely (100% you still need to go there), you skip the backtracking to the mayor (again 100% you need to go there), the snowman mini cut scene, irritating enemy dodges, the Monstar fight, and the duplighost fight. I think in an any% it can save 2-5 min, I know it saved 1-2 min for me.
The spin jump thing could only be used AFTER Lakilester is gathered, and getting spin jump in chapter 6 is a waste of time unless you can sequence break like 10 min (which I doubt) with that trick there. Chapter 7 likewise has nothing for the spin jump trick that I could find after a quick check, but then again with the Sushi clip, its not saying much. I haven't explored the later part of Bowser's castle for it yet, but I don't think there's another instance where you can use this trick for a huge boost of time, I'd have to get back to you with that.
Now for whatever clip you found Andre, vid please, this sounds interesting, toad town is directly connected to chapters 1, 3 and 4, I'd like to see how you do this and what you can do with it.
Also random thing I found for a SS, you skip the white clubba text in crystal palace if you use kooper's overworld attack, you don't get the cheap shot, but if you use mega quake, and his power/fire shell there is a chance you can 1-turn them, and if you time it right you can level up after the first or second one, you can beat all three relatively quickly. (That or star storm + fire/power shell)
Main Work!!!
I'm currently in vacation at my friends home and I'll be back home in three days. Then I'll get through the game again in my free time(outside of school and work), during this play through I'll record my prologue and chapter 1 plan in one segment and I'll record the jump trick and the sushi road if I succeed skipping every puzzles and all and I'll post them here.
edit: Finally I don't think I'll get anything under 3:30 so my goal will probably be sub3:30, also for the jump trick, when I'll record, I'll try to do it with the second jump. I didn't really try it so I don't know if it's doable with. If it's not, I doubt the utility of getting OOB in one section of toad town for now. Also, I don't know if it's the chapter trigger that can permit us to go to the next chapter but if it's not, it could be nice to skip a whole chapter with this.
edit: Finally I don't think I'll get anything under 3:30 so my goal will probably be sub3:30, also for the jump trick, when I'll record, I'll try to do it with the second jump. I didn't really try it so I don't know if it's doable with. If it's not, I doubt the utility of getting OOB in one section of toad town for now. Also, I don't know if it's the chapter trigger that can permit us to go to the next chapter but if it's not, it could be nice to skip a whole chapter with this.
THIS RUN CONTAINS SECRET WORLDS
Good luck skipping any chapter other than 7, since in literally every other chapter before then you get a partner. If you don't recall, in Bowser's Castle you're nominally required to use every one of your partners at least once. However, since the dark area is now skippable, one should think that Ch. 4 is skippable, right? Not necessarily; since Watt is an absolute requirement for Ch. 5 (recall the hunt for the yoshi kids), avoiding the Toy Box is pointless.
With that in mind, Chs. 2, 3, and 5 are all necessary for Ch. 6 (because of seeds), where Lakilester is (and since you need Lakilester for the lava area of Bowser's Castle, that means you'd have to find some sort of enormous break in the game to get into Flower Fields to avoid any of those chapters).
Here's a brief list of reasons why the partners you get from each chapter are all almost definitely needed later:
Ch. 1 - Kooper and Bombette get used again at least once in every chapter after this one. Some more notable examples are the escape sequence at the end of Ch. 5, blowing up the rock to access the Sun's tower, and the duplighost puzzles in Ch. 7.
Ch. 2 - Pretty much the same is true for Parakarry. He's needed literally everywhere there's a gap too large. Besides, Shell Shot is a godsend.
Ch. 3 - Bow doesn't get used as much, but she's needed to start Ch. 4 and during a critical moment in Ch. 8. Also, she's needed to get through tougher boss fights and scare away forced fights.
Ch. 4 - Watt is needed to finish the Yoshi Kid quest, even if (s)he is no longer necessary for Bowser's Castle.
Ch. 5 - Sushi is needed for early Crystal Palace, Ch. 6 (after Lily's water returns), and Bowser's Castle (i.e., that one sewer section no one likes).
Ch. 6 - Lakilester is needed for Ultra Boots and lava in Bowser's Castle.
In a nutshell, if you somehow find a way to skip anything before Ch. 7, you will almost definitely have to go back and do it later (except possibly Ch. 4, which has no bub-bulb seed; if you can get around that one goddamned Yoshi kid in the dark, Watt will become unnecessary for any%).
I also have the strange feeling breaking into Ch. 3 before 2 might be possible; the first "room" in Forever Forest is accessible as soon as you return from Ch. 1, and the only barrier is a guy who tells you it's too dangerous. Whether this is a good idea or not is questionable; since Parakarry is needed to reach Tubba Blubba's palace, the only things you could accomplish are getting the Ch. 3 seed early (which may be slower) and possibly get Bow early (which may be pointless since the Shy Guys won't show up until Ch. 3 is finished).
At this point I should suspect that Paper Mario's heavy linearity and reliance on events to let you make progress will ensure that this game isn't going to be nearly as breakable as SM64, DK64, or OoT.
With that in mind, Chs. 2, 3, and 5 are all necessary for Ch. 6 (because of seeds), where Lakilester is (and since you need Lakilester for the lava area of Bowser's Castle, that means you'd have to find some sort of enormous break in the game to get into Flower Fields to avoid any of those chapters).
Here's a brief list of reasons why the partners you get from each chapter are all almost definitely needed later:
Ch. 1 - Kooper and Bombette get used again at least once in every chapter after this one. Some more notable examples are the escape sequence at the end of Ch. 5, blowing up the rock to access the Sun's tower, and the duplighost puzzles in Ch. 7.
Ch. 2 - Pretty much the same is true for Parakarry. He's needed literally everywhere there's a gap too large. Besides, Shell Shot is a godsend.
Ch. 3 - Bow doesn't get used as much, but she's needed to start Ch. 4 and during a critical moment in Ch. 8. Also, she's needed to get through tougher boss fights and scare away forced fights.
Ch. 4 - Watt is needed to finish the Yoshi Kid quest, even if (s)he is no longer necessary for Bowser's Castle.
Ch. 5 - Sushi is needed for early Crystal Palace, Ch. 6 (after Lily's water returns), and Bowser's Castle (i.e., that one sewer section no one likes).
Ch. 6 - Lakilester is needed for Ultra Boots and lava in Bowser's Castle.
In a nutshell, if you somehow find a way to skip anything before Ch. 7, you will almost definitely have to go back and do it later (except possibly Ch. 4, which has no bub-bulb seed; if you can get around that one goddamned Yoshi kid in the dark, Watt will become unnecessary for any%).
I also have the strange feeling breaking into Ch. 3 before 2 might be possible; the first "room" in Forever Forest is accessible as soon as you return from Ch. 1, and the only barrier is a guy who tells you it's too dangerous. Whether this is a good idea or not is questionable; since Parakarry is needed to reach Tubba Blubba's palace, the only things you could accomplish are getting the Ch. 3 seed early (which may be slower) and possibly get Bow early (which may be pointless since the Shy Guys won't show up until Ch. 3 is finished).
At this point I should suspect that Paper Mario's heavy linearity and reliance on events to let you make progress will ensure that this game isn't going to be nearly as breakable as SM64, DK64, or OoT.
Wow, now Paper Mario is even starting to get a selection of speedrun-friendly glitches.
The issue with this is that their conversations are really quick, about as fast as the time it takes for Mario to stop to stomp on Kooper. It might save time on the last one, as you won't need to backtrack quite as much, but then you might be wasting time switching to Kooper.
Similarly, you can skip the cutscenes with the two forced Putrid Piranhas in the Jade Jungle by first-stiking from afar with Kooper or Bombette. Unlike the White Clubbas, these first-strikes will do damage. Interestingly, they also turn the battle into a "normal" one: ordinary music, and you can run away (this won't let you skip fighting them).
Quote from Raiker Z:
Also random thing I found for a SS, you skip the white clubba text in crystal palace if you use kooper's overworld attack, you don't get the cheap shot, but if you use mega quake, and his power/fire shell there is a chance you can 1-turn them, and if you time it right you can level up after the first or second one, you can beat all three relatively quickly. (That or star storm + fire/power shell)
Similarly, you can skip the cutscenes with the two forced Putrid Piranhas in the Jade Jungle by first-stiking from afar with Kooper or Bombette. Unlike the White Clubbas, these first-strikes will do damage. Interestingly, they also turn the battle into a "normal" one: ordinary music, and you can run away (this won't let you skip fighting them).
Micromanager extraordinaire
Hey everybody. Bumping this to see if I can get the 100% single segment category off the ground.
As we all most likely already know, the biggest difference between segmented 100% and SS 100% is the optimal method for gathering the ludicrous amount of money needed to gather all the badges. As seen in Raiker Z's fantastic segmented run, the apparently optimal method for segmented is save scumming the hell out of that asinine "Jump Attack" game.
Be warned: the rest of this post contains a lot of maths, so for those who don't care about the details or are lacking a background in either probability or statistics, the tl;dr of this post is that I'm going to try using advanced maths (mostly probability theory) to calculate the Jump Attack game's risk-reward and thus demonstrate whether this strategy has a shot at being successful in a 100% SS run.
Spoilered for length:
If there are any mathematicians, statisticians, programmers, TASers, or RPG buffs following this thread and interested in helping me investigate these problems and possibly others, feel free to speak up!
(For those not in the know, this is Sillygramcracker; I decided to change my user name a few months back, because my old one is kind of dumb.)
As we all most likely already know, the biggest difference between segmented 100% and SS 100% is the optimal method for gathering the ludicrous amount of money needed to gather all the badges. As seen in Raiker Z's fantastic segmented run, the apparently optimal method for segmented is save scumming the hell out of that asinine "Jump Attack" game.
Be warned: the rest of this post contains a lot of maths, so for those who don't care about the details or are lacking a background in either probability or statistics, the tl;dr of this post is that I'm going to try using advanced maths (mostly probability theory) to calculate the Jump Attack game's risk-reward and thus demonstrate whether this strategy has a shot at being successful in a 100% SS run.
Spoilered for length:
I intend to start off by calculating the expected return (average payoff), and coefficient of variation of returns (a common measurement of risk) for the Jump Attack game to prove mathematically whether or not it's worth our time when reset spamming isn't allowed. I've spent the past couple hours reviewing the mini game's rules and trying to come up with combinatorial probability distributions for each version of the game (something resembling the multinomial distribution perhaps?) to describe the likelihood of getting any given return in the range of all possible returns.
Some difficulties I've come across so far include:
The fact that there are apparently 13 blocks, but only 11 appear for Mario to choose from, means that even in the rare case where you succeed in 9-block mode, two blocks are left untouched so you don't get to see the whole sample. Apparently among the initial 13 there are always 4 1-coin blocks, 4 5-coin blocks, 2 5-multiplier blocks, and 2 Bowser blocks (if somebody knows I'm wrong about this premise please tell me ASAP!).
The fact that the game ends when you choose a Bowser block, and that a successful trial entails never seeing a Bowser block makes it unclear whether there's always two of them among the 11 that appear. It could be the case that sometimes there aren't any Bowsers, but we'd never know without save stating on an emulator! (By the way, if someone can verify this using an emulator, first making sure that the blocks are actually determined when they spawn rather than when Mario hits them, I would greatly appreciate it.)
The fact that the operations/functions your blocks perform (+1, +5, *5, and *0) on your no. of coins earned aren't commutative (i.e. the fact that the order in which you hit the blocks matters) means I have to keep track of permutations rather than combinations, on top of having to choose 5,7, or 9 from a set of 11 that was already filtered from a set of 13. The non-commutativity of these functions may make it necessary to construct an algebraic group containing all possible compositions of these functions (granted, many of them, i.e. any permutation containing Bowser, will be degenerate cases since they always yield zero coins).
After doing all this sophisticated nonsense, I will attempt to conclude whether this strategy is worth doing based on the following criteria:
Expected return: if the average payoff from paying this game repeatedly is positive, this trick might be worth doing. A high average payoff would be ideal, but considering it's a gambling simulation I expect the average payoff to be negative. If expected payoff is negative then this strategy is essentially worthless, since it implies that you lose money in the long run.
Coefficient of variation: the smaller this value is, the more consistently you get payoffs close to the average, and the less risky the game is. A high coefficient of variation would indicate that it's inconsistent/unreliable/etc.
Some progress has been made already on this XKCD forum thread; be warned, though, I think they made a few mistakes in their assumptions prior to calculating. This thread's especially useful for those with backgrounds in programming:
http://forums.xkcd.com/viewtopic.php?f=17&t=38054&view=previous
On another note, I'm also interested in trying to come up with a model to determine the likelihood of getting optimal battles in a single segment run, working with Raiker Z's existing route and modifying it for a single segment context. (Did he upload a spreadsheet to work with? I looked through this thread for one and only found his Word document.) Of particular interest are determining the likelihood of finishing off bosses in the fewest possible turns with Power Bounce, and of terminating unskippable non-boss fights with Bow's Outta Sight in the fewest possible turns. I'm thinking of modeling Power Bounce's effectiveness by assuming that the no. of successful jump commands with Power Bounce in a single turn follows a Poisson distribution with mean 5 or 6 for an ideal player, but the memorylessness property of this distribution may come into conflict with the fact that time is discrete rather than continuous in turn-based combat. The discreteness of time in turn-based combat also seems to sever the ties Poisson has with the exponential distribution. Should I consider using the normal distribution instead, considering how frequently Power Bounce is used? Yet another option would be to model each use of Power Bounce with a geometric distribution.
Some difficulties I've come across so far include:
The fact that there are apparently 13 blocks, but only 11 appear for Mario to choose from, means that even in the rare case where you succeed in 9-block mode, two blocks are left untouched so you don't get to see the whole sample. Apparently among the initial 13 there are always 4 1-coin blocks, 4 5-coin blocks, 2 5-multiplier blocks, and 2 Bowser blocks (if somebody knows I'm wrong about this premise please tell me ASAP!).
The fact that the game ends when you choose a Bowser block, and that a successful trial entails never seeing a Bowser block makes it unclear whether there's always two of them among the 11 that appear. It could be the case that sometimes there aren't any Bowsers, but we'd never know without save stating on an emulator! (By the way, if someone can verify this using an emulator, first making sure that the blocks are actually determined when they spawn rather than when Mario hits them, I would greatly appreciate it.)
The fact that the operations/functions your blocks perform (+1, +5, *5, and *0) on your no. of coins earned aren't commutative (i.e. the fact that the order in which you hit the blocks matters) means I have to keep track of permutations rather than combinations, on top of having to choose 5,7, or 9 from a set of 11 that was already filtered from a set of 13. The non-commutativity of these functions may make it necessary to construct an algebraic group containing all possible compositions of these functions (granted, many of them, i.e. any permutation containing Bowser, will be degenerate cases since they always yield zero coins).
After doing all this sophisticated nonsense, I will attempt to conclude whether this strategy is worth doing based on the following criteria:
Expected return: if the average payoff from paying this game repeatedly is positive, this trick might be worth doing. A high average payoff would be ideal, but considering it's a gambling simulation I expect the average payoff to be negative. If expected payoff is negative then this strategy is essentially worthless, since it implies that you lose money in the long run.
Coefficient of variation: the smaller this value is, the more consistently you get payoffs close to the average, and the less risky the game is. A high coefficient of variation would indicate that it's inconsistent/unreliable/etc.
Some progress has been made already on this XKCD forum thread; be warned, though, I think they made a few mistakes in their assumptions prior to calculating. This thread's especially useful for those with backgrounds in programming:
http://forums.xkcd.com/viewtopic.php?f=17&t=38054&view=previous
On another note, I'm also interested in trying to come up with a model to determine the likelihood of getting optimal battles in a single segment run, working with Raiker Z's existing route and modifying it for a single segment context. (Did he upload a spreadsheet to work with? I looked through this thread for one and only found his Word document.) Of particular interest are determining the likelihood of finishing off bosses in the fewest possible turns with Power Bounce, and of terminating unskippable non-boss fights with Bow's Outta Sight in the fewest possible turns. I'm thinking of modeling Power Bounce's effectiveness by assuming that the no. of successful jump commands with Power Bounce in a single turn follows a Poisson distribution with mean 5 or 6 for an ideal player, but the memorylessness property of this distribution may come into conflict with the fact that time is discrete rather than continuous in turn-based combat. The discreteness of time in turn-based combat also seems to sever the ties Poisson has with the exponential distribution. Should I consider using the normal distribution instead, considering how frequently Power Bounce is used? Yet another option would be to model each use of Power Bounce with a geometric distribution.
If there are any mathematicians, statisticians, programmers, TASers, or RPG buffs following this thread and interested in helping me investigate these problems and possibly others, feel free to speak up!
(For those not in the know, this is Sillygramcracker; I decided to change my user name a few months back, because my old one is kind of dumb.)
I can verify that the blocks are set when you pay the gold. I did run tests to see if there were set patterns for them, like in most Mario games, I hit about 100 unique patterns with no repeats when I quit. I didn't think to check if there were always 2 Bowser blocks per sample though in every one I know for sure I ended up hitting all blocks; there were.
Also sorry, but I didn't make a spreadsheet, I had the word document for basic overview of where I should be, and for more specific things I had a notebook with everything like drawings, notes, anything that I remembered from my test run that stood out and I might need at a glance, rather then checking my computer. 90% of it was just a list of Quizmo/Hidden Panels/Star Pieces that I had completed/collected.
Also sorry, but I didn't make a spreadsheet, I had the word document for basic overview of where I should be, and for more specific things I had a notebook with everything like drawings, notes, anything that I remembered from my test run that stood out and I might need at a glance, rather then checking my computer. 90% of it was just a list of Quizmo/Hidden Panels/Star Pieces that I had completed/collected.
yuri 4 life
I did ss some months ago with not well planned route and luck of experiment.
http://ja.twitch.tv/0xwas/b/308119026
I have 6:31 now but there are rooms of improvement. I sold Dried Fruit to buy badges but I think smash attack is probably faster.
In bowser, turned out HP5 mario works fine with badges and Chet Rippo. I don't need many life shrooms if I use.
http://ja.twitch.tv/0xwas/b/308119026
I have 6:31 now but there are rooms of improvement. I sold Dried Fruit to buy badges but I think smash attack is probably faster.
In bowser, turned out HP5 mario works fine with badges and Chet Rippo. I don't need many life shrooms if I use.
yuri 4 life
6:23:51 in SDA timing/ 6:39:03 in RTA timing.
http://ja.twitch.tv/0xwas/b/328955082 and http://ja.twitch.tv/0xwas/b/328965812
I had 2 resets in chapter 4, some unneeded backtracks and 3 wrong answers in quizmo lol. I switched dried fruit to smash attack to get coins, it worked very well. Most strategies are same as before but I used danger mario after chapter 7.
http://ja.twitch.tv/0xwas/b/328955082 and http://ja.twitch.tv/0xwas/b/328965812
I had 2 resets in chapter 4, some unneeded backtracks and 3 wrong answers in quizmo lol. I switched dried fruit to smash attack to get coins, it worked very well. Most strategies are same as before but I used danger mario after chapter 7.
You'd need all partners, but I guess you could do early ch 4, get up to Watt, then move onto Sushi and backtrack for Bow + ch 3 badges. Then I guess when you go to desert for the melody you can collect desert and ruins badges... Although I think beating ch 2 before glitching would be better if your still going to enter the ruins.
Didn't he change it to Hammer Attack qwerty? I can't remember all of what is in the ruins though, I guess reasonably we don't get ANY keys and just grab the badges, star piece, upgrade and Hammer in the ruins?
I guess one thing is you can skip ruins until after ch 5 then you don't need the super hammer, then stop by ruins on the way to the Mt. Rugged hidden panels, since your going to need to visit the desert anyway.
I guess one thing is you can skip ruins until after ch 5 then you don't need the super hammer, then stop by ruins on the way to the Mt. Rugged hidden panels, since your going to need to visit the desert anyway.