Experienced Forum User, Published Author, Skilled player
(1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
I vote no mostly because bad game choice. The game is okay, but not for TASing. The whole TAS looked like a normal play-through. However the reset trick is new to me. Atleast you improved the Pilot Wings WIPs with many frames. So keep up the good work, and next time I advise you to talk about the game you gonna TAS next time so others can help you and leave a comment about game choice / TAS future / help with planning etc.
edit: I forgot to mention the lack of entertainment (due to bad gameplay choice)
PhD in TASing 🎓 speedrun enthusiast ❤🚷🔥 white hat hacker ▓ black box tester ░ censorships and rules...
Experienced Forum User, Published Author, Skilled player
(1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
I think the exactly very same as moozooh after watching it on youtube. Also I'm pretty sure that this movie is improvable with many seconds, not to mention the lack of entertainment. Please watch other sidescroller shmup TASes to understand what I'm talking about.
I'm going to vote no. Sorry.
PhD in TASing 🎓 speedrun enthusiast ❤🚷🔥 white hat hacker ▓ black box tester ░ censorships and rules...
Experienced Forum User, Published Author, Skilled player
(1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
Okay, I'm tired of this. Here are the facts in order of "priority". I wrote some opinions in the details. Keep in mind that this isn't true only for NES, I'm just show that EVEN the NES has these properties I mention (also the number of pros and cons isn't important).
Pros
- It's a built-in and intended feature in the NES
I hope I don't need to write any details for this.
- NES games are capable of detecting that the game was reset through the reset button
For an instance, Zanac. The game has many cheat codes just like the level selection code:
Press 'RESET' button on your NES 13 times then press Start. Use Left and Right on the D-Pad to select the stage.
Here's adelikat submission: #1344: adelikat's NES Zanac in 03:26.22.
Also there's 2 more cheat codes that makes use of pressing Reset for a various time (9 and 1).
One of them: http://tcrf.net/Zanac. By the way this is a very useful webpage for TASers!
- Saves time in various games
Watch this page: http://tasvideos.org/Movies-C3045Y.html
Note that most of them uses an input combination (mostly on the 2nd controller) to reset the game which isn't the same as pressing the reset button!
Cons
- It's not 100% verified that every NES hardware executes the same reset routine
Check out http://nesdev.parodius.com for further info or http://wiki.nesdev.com/w/index.php/Nesdev_Wiki about their experimentations and results.
- Save game data corruption (with the reset button) has a chance to fail on the real hardware
I tried to search for dissassembled games with saving data but couldn't find anything. I'm not going to details, I think you know why should it fail.
If the problematic cons will be verified on real hardwares, I will say that use the reset buttons to play games as fast as possible. Have fun.
PhD in TASing 🎓 speedrun enthusiast ❤🚷🔥 white hat hacker ▓ black box tester ░ censorships and rules...
Experienced Forum User, Published Author, Skilled player
(1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
Warp wrote:
...for example, the game corrupting its own save data due to a programming error (which is triggered by some unusual untested situation) and the save data becoming corrupted because the console was shut down while the game was writing the data. ...abuses hardware (which IMO should not be part of TASing, or even speedrunning in general).
I understand that this is only an opinion, but really... unusual untested situation? 90% of the TASes makes use of this. The first thing came up in my mind reading this line was the Up+Down and Left+Right combination. It's impossible to press on regular gamepads, you have to dissassemble them and of course they will work, and it's heavily abused in TASes. One of the greatest "unusual untested situation". So this also counts as an invalid gameplay?
The save corruption is really questionable since it's depends on the hardware version (according to various technical documentation regarding NES consoles), but it's a part of speedrunning, this is a fact. SDA also uses it for example NES Clue.
Don't think that I want to change your opinion or something like this, I just found your thoughts interesting.
PhD in TASing 🎓 speedrun enthusiast ❤🚷🔥 white hat hacker ▓ black box tester ░ censorships and rules...
Experienced Forum User, Published Author, Skilled player
(1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
micro500 wrote:
If I were going to try to play back videos on the console using a reset, I would play back the movie up until the the reset, then stop playback and manually reset the console, and let the bot continue from there. If the bot held no buttons while waiting for me to reset the console, and started playback fine after the reset, how many movies would this work for? I'm assuming this wouldn't work for games that use save corruption, since this is based on precisely timing the reset to "corrupt" the memory into a favorable state. But how many games would it work for?
I suggest you to try TASes that "uses a game restart sequence" by pressing a combination of buttons. This means you don't need to manually press the reset button since the game will handle the button combination required for the specific game so it will sync perfectly theoritically. However, most of the games requires a 2nd controller for this.
Also... manually... I don't think you will have enough patience and extreme rhythm sense to pull this out, the press and release have to take for exactly 1 frame, not to mention TASes with multiple restarts in a row.
Now back to real life. As far as I know, the reset routine heavily depends on the actual hardware version. However, if you develop another NESBot (since I think you have to for another controller or modify the existing one, I don't know the correct specification atm), then try these, they are much likely to sync than other TASes (they use Up+A on the 2nd controller):
[1082] NES Zelda II: The Adventure of Link "warp glitch" by Inzult in 05:43.47[1144] NES Metroid by Lord Tom in 08:19.32
I couldn't find TASes using restart sequence without using a 2nd controller or the reset button.
edit: also sorry for this debate/off topic on the verifying thread.
PhD in TASing 🎓 speedrun enthusiast ❤🚷🔥 white hat hacker ▓ black box tester ░ censorships and rules...
Experienced Forum User, Published Author, Skilled player
(1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
BeefOwl wrote:
Here it is:
http://www.megaupload.com/?d=JWMQVBG1
However, I'm pretty sure it didn't record because no input comes up on Player 2's display during playback, unless for some reason it was stored but not implemented...
My real question is what do I do next time to make it record both inputs, though
It's unfortunate but yes, it's really unrecorded.
Try this, it's worked for me:
- open ROM
- Config > Input...
- select gamepad for port2
- click Configure right next to it to bind the keys
- click Set at Input presets 1
- Close dialog
- File > Movie > Record Movie...
Next time just simply open ROM and record movie and it should work. Try it by making a very slow 1-2 sec movie. Also the movie file will contains lines like
|0|....T...|.L.U....||
instead of
|0|....T...|||
PhD in TASing 🎓 speedrun enthusiast ❤🚷🔥 white hat hacker ▓ black box tester ░ censorships and rules...
Experienced Forum User, Published Author, Skilled player
(1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
Warp wrote:
MESHUGGAH wrote:
You mean TASes using the "reset" button or using a restart sequence by pressing various button combinations on the controller (1/2)?
I'm talking about the reset button on the console.
In my opinion "playing a game" consists of supplying input to the game which the game reads and interprets1. "Playing a game" does not consist of shutting down the console and restarting it2 (which is what resetting essentially is). Just because you can shut down the console, and just because many emulators have support for emulating this, doesn't make it a valid form of playing. (IMO it's no different from using gamegenie codes: Just because you can do it with a real console and just because many emulators support it, doesn't make it a valid form of playing.)
So according to your logic, it's valid to press various button combination to reset the game1 which should be nearly (or exactly) the same as pressing the reset button, but it's invalid to press the reset button2 ? Feel free to correct me, I'm just guessing. I also think that this is about "verifying the consistency of emulators/tases on the real valid console" and not "making the console to act as the emulator".
PhD in TASing 🎓 speedrun enthusiast ❤🚷🔥 white hat hacker ▓ black box tester ░ censorships and rules...
Experienced Forum User, Published Author, Skilled player
(1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
Warp wrote:
I'm still of the opinion that resetting is not valid input in TASes and should be banned. and this only strengthens that opinion.
You mean TASes using the "reset" button or using a restart sequence by pressing various button combinations on the controller (1/2)? Also "strengthens that opinion": I think you wrote this for the unaccessibility of the reset button through the controller, right? Well, I think the reset button is on the same "level" like the power on button/switch (since you have to control the console to start a game and not the controller), however I don't know the consistency of it on a real console.
PhD in TASing 🎓 speedrun enthusiast ❤🚷🔥 white hat hacker ▓ black box tester ░ censorships and rules...
Experienced Forum User, Published Author, Skilled player
(1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
Thank you for explaining the timing issue and also for sharing the code.
edit: for the "lack of sending the reset input": Thread #10783: Legitimacy of using resetmicro500 wrote:
Experienced Forum User, Published Author, Skilled player
(1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
Bisqwit is right, however it's interesting that it's really working. I'm also interested in your full code. Here are some pages where you could upload it (if we can kindly ask you in order to support you)
http://kpaste.net/http://codepad.org/pastebin.cahttp://notepub.com/
PhD in TASing 🎓 speedrun enthusiast ❤🚷🔥 white hat hacker ▓ black box tester ░ censorships and rules...
Experienced Forum User, Published Author, Skilled player
(1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
dnnzao wrote:
I have asked there, and they or not ansewer me, or tell me to ask here, maybe because Im a shit at TASing, and they probably hate noobs :)
I've read the last comments you wrote on the SMW thread and all of them has been replied with informative and helpful informations, you also thanked them. I don't know why did you wrote this, anyway it's a general "rule" (or atleast a behaviour) that everyone helps others. Don't think that you won't get help just because you are new to the TAS scene or anything. Not to mention that the "The Newbie Corner" threads are also frequently replied with constructive tips.
PhD in TASing 🎓 speedrun enthusiast ❤🚷🔥 white hat hacker ▓ black box tester ░ censorships and rules...
Experienced Forum User, Published Author, Skilled player
(1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
- Read all the already known and detailed glitches here: Wiki: GameResources/SNES/SuperMarioWorld
- Read the corresponding thread here (86 pages): http://tasvideos.org/forum/viewtopic.php?t=1571
Feel free to ask your questions at that thread. I don't want to be cruel with you, but I think you should know that there are already a few excellent tasers who tries to come up with amazing glitches, like the last one: swapping objects to the "ending bar". I think it's strongly advised to read all the pages in that thread (at least the last 10-20). Good luck and have fun.
PhD in TASing 🎓 speedrun enthusiast ❤🚷🔥 white hat hacker ▓ black box tester ░ censorships and rules...
Experienced Forum User, Published Author, Skilled player
(1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
Brandon wrote:
That said, where would you acquire the source code for a specific ROM?
Depends on ROMs, but mostly you will have to disassemble them with a debugger. It's a very tedious and long task and also it's illegal, I don't know what's the situation with old games' ROMs copyright issues.
Don't forget that even if you get the source code* (after spending many days), you will be still nowhere. Source codes are long, "obfuscated" and it isn't easy for human beings to read it. However, I think that this way has some advantages over the "trying all the possible combinations". (mostly the time factor as I mentioned earlier in this thread). And also, making an algorithm to find the shortest route to the final desired memory state... The idea is okay, but it's much harder than it looks/sounds. I will investigate it further later.
I mostly choose the "read the source code (even if it's outdated)" method, cause they are something like a technical document and you can come up with new glitches and techniques. I already found many funny glitches for mobile games (.JAR files, you can simply open it as .zip, decompile them automatically (with the proper programs), and read it obfuscated), just like playing on custom resolutions leads to unexpected results. Not to mention PC games, I know very much glitches for Counter-strike because I already investigated Quake and HL source code (the engines based on... off topic).
Links you should visit:
http://en.wikipedia.org/wiki/Disassemblerhttp://bisqwit.iki.fi/jutut/megamansource/http://eludevisibility.org/spacy-funky-bob-source-code/http://www.zophar.net/utilities/development.html
*: The code that you decompiled. Of course, the original source codes are not obfuscated, and easily readable.
PhD in TASing 🎓 speedrun enthusiast ❤🚷🔥 white hat hacker ▓ black box tester ░ censorships and rules...
Experienced Forum User, Published Author, Skilled player
(1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
Warp wrote:
It's not that straightforward because even with brute force searching you can use clever algorithms to somewhat cut down the search tree, or at least prioritize branches which are more likely to give you a result. Also, in average it would take half of that time. The complexity class is correct, though (that is, O(2n).)
I already thought and tried to come up with some clever ideas, but still the time needed to complete a game completely without human actions would take rather long. I tried to generate movie files by setting a "maximum number of inputs per frame", "enabled and disabled input combinations", "preferred inputs" (start generating movies with Left and A instead of Restart and Select), "memory investigation per frame", "possible end-game conditions". Of course it would be great for really really short movies, where everything is already known, and there are no obscure suprises. For example, I never thought that pressing the restart button warps me to the final 9th stage. However, I'm really curious that there are existing movies where possible odd thinking improvement exists like the one I mentioned before.
I would rather write a program that analyses the structure of a game's source code and tries to find the shortest route to the final desired memory state.
Warp wrote:
<nitpick>
Don't you mean "at most"? (And there's no need to specify that many decimals as they convey no useful information in this case. A 5.3*101613 would have sufficed to give a notion of the magnitude.)
</nitpick>
You are right, I made a mistake. It's the "shortest longest time a brute force would take at 1000% emulation time" not considering other non-gameplay tasks (generating movie inputs, loading them, analyzing "possible end-game" results, etc). I just simply copied the result from windows calculator, I thought that people without proper math knowledge will fear recklessly if they have to read so many numbers after the other.
edit: I failed to correct my mistake, amazing.
PhD in TASing 🎓 speedrun enthusiast ❤🚷🔥 white hat hacker ▓ black box tester ░ censorships and rules...
Experienced Forum User, Published Author, Skilled player
(1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
Brute force is the best solution to find the fastest completion input collection. Needless to say, the time needed for brute forcing a game is something like that:
[2 ^ (Number of buttons * Actual TAS record frame count)] / Emulation time.
For example, the shortest movie is [1145] Genesis King's Bounty by gia & Aqfaq in 00:09.93 so brute forcing it would take at least
2^(9*596)/10 = 5,3075823547897896501038675241442e+1613 seconds.
Feel free to correct me, I'm not good at math.
Of course, optimizations will reduce this time, but also increases the chance to be slower than the possibly fastest input series used to complete a game. The reason is that there are many bugs and glitches that happens "randomly" for normal human sense. Just like sub-pixel calculations, wrong order in the code (contra fadeout glitch), warping like a madman (castlevania aimmx), etc.
edit: formula...
PhD in TASing 🎓 speedrun enthusiast ❤🚷🔥 white hat hacker ▓ black box tester ░ censorships and rules...
Experienced Forum User, Published Author, Skilled player
(1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
The developers @ sun already made the Java(TM) Platform Micro Edition SDK 3.0 which comes with an emulator. I'm planning to do some research about the possibility to TAS j2me games.
edit: Okay, I've ended my research.
I assume you meant MIDlets that playable on mobile phones. I don't have much free-time, so I'm not sure I will start writing an emulator atm, however I wrote down some guidelines:
- The required APIs
Sun's newer and older APIs (midp 1.0 ~ midp 2.0 and others), manufacture specifics (from Nokia, Motorola, Sony Ericsson) and some others (OpenGL ES and audio relateds like mp3/ogg)
- Jar verification
Ask game developers for some checksums of their original games. Also make a backup of the original Jar file when you load it.
- Frames?
To make a movie file, I recommend using a special time-based input file where the smallest time unit is microsec. Can't come up with better atm. Here's an example movie file:
0 p5 //press 5 at 0 microsecs
352 h5 //press and hold 5 at 352 microsecs
678 p3 p7 //press 3 and 7, release 5 at 678 microsecs
- Keys?
Inputs can range from 19 keys to qwerty layout. Don't forget extra buttons like restart/resume/shutdown and volume controls etc.
- Emulator options I should want to use while tasing a game
Ability to play on custom resolutions (if the game scretches the game screen).
Memory viewer.
Code viewer.
Fancy animation when vibration occurs.
Here's an example of making a very basic emulator: http://www.longsteve.com/?p=4
PhD in TASing 🎓 speedrun enthusiast ❤🚷🔥 white hat hacker ▓ black box tester ░ censorships and rules...
Experienced Forum User, Published Author, Skilled player
(1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
gia wrote:
......You'd be fine with the statistics and /or credits looping infinitely since that's an end game state and you cant progress further if it loops infinitely.
I'm not sure if you'd be fine if the game was forced to loop infinitely drawing the last frame that says The End. You are at the very last instruction of the ending, and you are constantly executing it again and again, does that count as state? You haven't just drawn the end once and continued to play afterwards, the game is locked drawing it forever.
However if you created the code that will make this loop in some area of memory you have access to, and as such it is not the original's game ending code the one being executed, it is not the The end that should be drawn, its an identical clone you created, I'm not sure if it would fit your definition.
Don't forget that this is only my opinion. Now let's face it... Memory corruption (and other abuses) screws my definition, so let's revise it again.
In gaming, completing/beating a game means that you fulfilled the game's objectives. Depending on the game, either you get some statistics, congratulation screens, new features (new game+, new items etc) or just a plain the end image. This means you gain such an experience that only achieveable if you beat the game.
The game's perspective again: If you met the requirements for a set of conditions, let's start the ending sequence. Depending the code and features, completing a game unlocks features that can be accessed at a new game.
---> Direct/indirect memory corruption enables the TASer to change the part of/entire code. This means you can define new requirements needed to met to complete the game, changing values to immediately met the requirements, etc. This leads to a very-very different reception among tasers and viewers.
To define a game ending, someone from the bosses of tasvideos should detail furthermore the existing rules and the general acception of a movie.
Obsoleting my definition leads to another (ignoring viewers' reception):The movie must reach a point (optionally where it mets the game's requirements) that leads to a state which should only occur when the game is beaten.
Some notes should be added regarding newgame+. Also this is such a lousy and abusable definition... can't come up with better one that satisfies movies with or without memory corruption.
PhD in TASing 🎓 speedrun enthusiast ❤🚷🔥 white hat hacker ▓ black box tester ░ censorships and rules...
Experienced Forum User, Published Author, Skilled player
(1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
henke37 wrote:
How does this do with things like any % vs 100 %?
It also is very unclear with the regards to previous saves/passwords. It may not be it's job to deal with these matters, but I feel that a curiosity examination of the interaction with the matter is warranted.
Any% and 100% are different game endings. Both of them eventually leads to a positive end game state. Show an example if this doesn't happens in a specific game.
Well my opinion is that passwords are used if they make the gameplay harder or have such a benefit that makes the run faster without the intention of making the gameplay easier/shorter.
Possible passwords: a "hard difficulty" code, an intended "negative" code that abusable, making the run faster with a curse for example
Denied passwords: warp to level N, all weapons at the very start (new branch if it's significantly much-much faster than a normal tas)
But that's just my opinion...
henke37 wrote:
And more importantly, how does it do with games where you see how long you can last? (Assuming no killscreens or similar) For that scenario there is the problem of a lot of games having upper limits to how high they can track your progress (score counter overflows/maxes out). The game doesn't track further progress, but you can do it instead.
Do you know any games that could be TASed and can be played forever without reaching a point where no more change occurs in the gameplay? I merely assume that after a limited time (that would range from 2 hour to 2 days or even more) the game will have something like a repeating pattern (define end-point), a memory corruption (that should end in a freezing state/restart/death) or unwinnable situation (inability to move/attack/do anything due too many enemies on screen, lag, anything).
PhD in TASing 🎓 speedrun enthusiast ❤🚷🔥 white hat hacker ▓ black box tester ░ censorships and rules...
Experienced Forum User, Published Author, Skilled player
(1892)
Joined: 11/14/2009
Posts: 1349
Location: 𝔐𝔞𝔤𝑦𝔞𝔯
Warp wrote:
I have to disagree with the notion that if a game has a programming error that makes it impossible to reach the normal ending intended by the game designers, the last point reachable in the game could be considered a valid "ending" with respect to a TAS.
If a game cannot be completed, then it cannot be completed, and thus it's unTASable (at least with respect to the traditional goal of playing the game through). It's no different than eg. a game that is so buggy that it outright fails to start being completely unTASable.
Of course a special goal/category could be created for such a game, but it would be an exception to the general rule.
There's a very few (~9?) games in the world that can't be completed due programming error without patching the original game (if it's possible) so this isn't a serious threat, I just wanted to be specific for such a problematic definition. Also note that the "end point" shouldn't be the "last point reachable". Think about games with an ending like changing the map to "theend". It makes no sense to run through a level that has a text "THE END" made out of tiles and jump to the last reachable place. This mostly featured in DOS/PC games.
Guess what: Action 52 has many uncompletable games since there's no ending [positive game end state] or can't be completed.
And I didn't mention games that has absolutely no positive end game state only a game over screen.
edit: And not to mention games without end game states like SuperNova
PhD in TASing 🎓 speedrun enthusiast ❤🚷🔥 white hat hacker ▓ black box tester ░ censorships and rules...
