I went ahead and started this but I'll probably get board before I finish
Just the first three trial levels for now, found a small exploit that makes it slightly less boring.
http://www.megaupload.com/?d=7973K8OC
normally the game will stop you and give you messages at several points but if you tell Maxwell to get an item he will still move when messages pop up.
There are other exploits that involve switches and switchdoors. A Starite can go through a switchdoor with a string-like thing or a fan. You can also launch yourself through ground and other solid objects by trapping yourself between two doors and releasing one of the doors. I think a pure speedrun should be done (if it is done fast enough with proper exploiting of items, it would be more entertaining than a run that only goes for low par, in my opinion)
it would probably be quite boring because the same thing would be done most levels (spring or vending machine glitch).
An idea: what about doing the whole game without repeating any items
not repeating items seems like an arbitrary restriction. Not to mention it would make a game already impossible to optimize much much harder. Just the planing for such a run would take decades.
Idk about a pure speedrun, every level would look the same.
is it a bad thing to make it harder to optimize? Would make it more interesting I think, and if the "route planning" (well actually summon planning I guess) is not perfect, nobody would notice. Also the game rewards you for using an item you haven't used before so maybe the goal is not that arbitrary
Going for low-par would almost always involve glitching the star to you through walls with a perfectly placed spring each time, except the levels you can complete with zero summons. That would hardly be very entertaining as there's so much diversity that would be untapped.
I think most of the levels can be beaten with zero summoning, glitching the star out should only be low par in a few levels.
forgot about the new item bonus
it can be glitched by adding letters to the front of the word and it counts as a different word. fan, a fan and b fan all count as different words.
but scribblenauts is all about the summoning! If you do most levels without summoning, we'd see so little of what makes the game so great!
maybe the extra letters glitch would have to be banned - perhaps "if two words summon the same item only one may be used"
My personal inclination is to not aim for speed as a goal at all in a Scribblenauts TAS. Instead, aim for creative solutions and unusual situations. Maybe your level solution requires swinging a soccer ball around on a vine, or a duel between Charles Darwin and a humpback whale. Above all, keep the viewer guessing. TASing in this case would not necessarily be required for all solutions, though efficient summoning and placement of objects would of course be a keystone.
How amenable is the game to hexing? If it can be hexed easily, then you could split up the game by level and have multiple contributors working in parallel to come up with lots of unusual approaches.
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
My personal inclination is to not aim for speed as a goal at all in a Scribblenauts TAS. Instead, aim for creative solutions and unusual situations. Maybe your level solution requires swinging a soccer ball around on a vine, or a duel between Charles Darwin and a humpback whale. Above all, keep the viewer guessing. TASing in this case would not necessarily be required for all solutions, though efficient summoning and placement of objects would of course be a keystone.
How amenable is the game to hexing? If it can be hexed easily, then you could split up the game by level and have multiple contributors working in parallel to come up with lots of unusual approaches.
I second this. Keeping the viewer guessing is probably a key part of the entertainment in this game.
Plus, the game should be extremely easy to hex edit since each level is very short and the level menu involves pretty much a few clicks of the touchscreen.
I like the idea of putting together a huge team and focusing on interesting solutions.
The intro levels are going to be the same regardless and are proving harder to optimize then you might think.
personally I don't like the idea of low par at all, whether I'm playing unassisted or for this TAS. Being rewarded for using less of the game seems backward for me - especially if most levels can be done with zero summons - it would just be a (bad) platformer for those levels.
Derakon's idea seems like it could work but I'd personally prefer a proper goal rather than just "be entertaining".
In that case, I advocate speed and entertainment rather than low par and entertainment. I remember some levels where it is faster to use a few items to complete than not using items.
"Fast and unexpected" should be a goal for a TAS run of this game.
I like Derakon's idea as well. Still not sure if I would like to see action only or to add the puzzle levels as well. Advanced levels are also a possibility.
I vote for going for fastest time with the additional condition of not using an object more than once for the entire run. I say object instead of word in that multiple words or phrases can map to the same object.
yes that's the problem with doing this run, there are a ton of different ways to do it.
Think about the planing required for that run you just suggested, seriously if you can figure out the ideal items for each stage without repeating and maxing speed, the actual run would be trivial
yes that's the problem with doing this run, there are a ton of different ways to do it.
Think about the planing required for that run you just suggested, seriously if you can figure out the ideal items for each stage without repeating and maxing speed, the actual run would be trivial
Well yes, that makes it more challenging. If you did allow the same items, most stages would be solved in the same way using the same exploit. How many stages can you do by putting hand cuffs on the star, putting the hand cuffs into a vending machine, dragging the star with them, then moving the vending machine to your character, and vending the hand cuffs to get the star. Its cute to see once, but it would get repetitive if you do that every time.
As far as route planning, you probably want an arsenal of a few fast combos like that, then try to do each stage without those items. The stages that take the longest, you substitute with the faster easy way instead.
except for some of the more complex levels the handcuffs in a box trick wouldn't be the fastest because you need to scroll all the way to the box two items and then all the way back. Quicker just to summon one item and go oob in most cases I suspect. Plus there are ways to trick the game into summoning the same item but counting as an item you haven't summoned before.
I actually think a pure-speed run is a good idea. Making a each-item-only-once run takes so much planning that making a pure-speed run would actually be a great help. I don't know how hex-friendly the game is, but it does seem to be a project that could be done by many people (1 level per week or so, fastest goes in the actual movie).
I would totally be game for a multiperson run if the levels can be hacked together easy enough. The main reason I haven't worked on this much is because everyone has a different idea on what the objectives should be. A pure speed run might be somewhat repetitive but it would be a good starting point.
Even if it's not hexable switching with someone every level or two would be a lot of help
Actually, if you keep one up-to-date file online which contains all levels that have been done so far, people only need do download that one and they can continue recording. If you'd decide on what point to end the next level, that would make comparing the different attempts a lot easier.
