Best demoscene production of the year, nominated for The Meteoriks - Outstanding Technical Achievement.
Sorry, I don't know what happened in the tas because I was looking at the background effects.
By the way, could you please explain how you manipulated the RNG? The video itself does not reveal anything about it. It looks like you start the game and magically have perfect RNG, lol.
I would say that control 1 knight and any% makes most sense, because that is how most people play the game. Playing with 4 knights is comparable to making a cooperative TAS in Doom with 4 players. 4-player run is of course very interesting, but more rare and exotic? Maybe 100% run would make more sense as a 4-player run?
No, wait, maybe it is better to control 4 knights to complete the game faster? 4-players makes tasing easier because you don't need to manipulate enemy movements.
No, wait, maybe enemy knight can bring us some key item fast because their action scenes are skipped?
No, wait, maybe a 2-player run is the fastest? Oh, it is more complex than I thought...
I would say anything goes!
> Would a less than ideal RNG roll be acceptable for sub?
I think it is impossible to prove that some RNG setup is optimal, so I would say that anything goes. Maybe the Moonstone speedrunners can help find some good RNG?
> Id really like to TAS this game but the complexity is so daunting
Me too. The game is amazing, but out of my league technically. I think for any% the initial key item locations is the hardest part, so if you can figure it out the rest of the TAS will be relatively easy and can be done in sub 5 minutes.
I think 100%/low% items is not interesting because most items are meaningless. 100% kills makes more sense, i.e. clearing all lairs and killing the dragon and rival knights (unless we control all 4 knights which we probably do). I think the optimal initial RNG for 100% run is literally impossible to find, ever. The difference between the best and average RNG is probably not much timewise, anyway.
TL;DR Most sensical combination: Control any number of knights and any%.
> you could start the game faster by holding start
Now there is another incentive for somebody to improve the run! It is a fun game to TAS. Plenty of moves to consider at each point. I wish somebody improves it by a lot someday!
A great TAS from HappyLee. I appreciate speeding up loading times.
IMO, if a loading screen has duplicated content, for example black screen and no sound for multiple consecutive frames, then all duplicated frames could be removed from the standard encode that goes for the general audience. The site could still archive the "true encode" with all loading times unaltered. But really showing black screen and no sound for multiple frames is just random noise and serves no purpose for any audience. That is because the general audience does not care or does not know, and the expert audience knows that the loading time was skipped and nothing was lost. If loading times are removed, then everybody wins.
On the other hand, maybe a glitch is found that speeds up loading times? Now the audience gets confused. So it makes sense to always publish unaltered encodes, so that there is no misunderstanding on what is what. If loading times are unaltered, then everybody wins.
A very long frame rule per level. Very short levels.
A very short frame rule per level. Very long levels.
A boss whose death animation lenght is determined by how fast you kill it. The faster you kill it, the longer it takes to die.
Thank you, I can navigate the site now, but there's just so much information that it is a bit overwhelming. Good job keeping it all together.
I think I found one way to improve it. Now most game pages have the screenshot duplicated, which looks confusing. I think there is no need for the game screenshot, because the same screenshot is immeditely seen below as the publication screenshot. If a game screenshot is necessary, then maybe replace it with the game cover image like they do at speedrun.com? You could just copy the images from there.
I think the game screenshot immediately catches the eye and moves attention to the right side of the page, so now I feel like I am reading from right to left. The same effect does not happen at speedrun.com where the game cover image is on the left side.
Bigbass wrote:
When you use the search box on the game list page (synonymous with using CTRL+F)
Doesn't the search box use site resources while ctrl-f does not?
This is very good, but now the game list links do not link directly to the game wiki pages like the amazing Super Mario Bros. wiki page: https://tasvideos.org/GameResources/NES/SuperMarioBros. I guess there is so much information on the site that I get confused and don't know where to click, LOL. I think the game list links could link to wiki pages where both the game resources and publications are combined, to minimize the size of the labyrinth of information. Oh, but then it makes the pages heavier, I see...
I understand that listing all games on the same page can be unnecessarily heavy, but vgmaps.com does it like that, literally all games listed on one page: https://www.vgmaps.com/Atlas/index.htm
This is very useful because you can use ctr-f to instantly find two things: the game you are looking for, or the fact that there are no maps for it available on the site. It is the speedrunner's method.
Hmmm. Now I understand, it can be done like this: https://tasvideos.org/Games/List?CurrentPage=1&PageSize=4386&Sort=Name
If it is not too heavy, consider making it the default view? Now all games are listed on a single page. Very nice! Thanks! Now I can use ctrl-f to instantly find the game and whether it has been tased or not. Problem solved.
The pure text list is often underrated and very useful for many kinds of things. For example, did you know that the word "ninja" is mentioned 58 times on game titles on the site? A site with 58 ninjas seems pretty dangerous, so I must agree with the thread title.
TL;DR I think "games" and "game resources" and "publications" could be combined into a single wiki page. Now the publication page does not link to game resources directly. The game resources also does not link directly to publications. It is a bit confusing for my limited brain capacity, LOL.
Thank you, RiTheMasher. Now that you are here, please make your own TAS of your favorite game. :)
I think the maximum goals run would be a fun thing to see. I estimate that somewhere aroud 50 goals can be made per round, so a total of 150 goals. The goal counter glitches out at 100, but as far as I know this does not affect the game.
Bug report: After using Basic Bot the messages disappeared permanently. Messages like "Saved state: QuickSave6" does not appear. Restarting emulator did not help. Deleting config file helped. (I also got like a thousand messages saying that something was wrong with savestates but I did not have time to read what the messages said, but it looked like a loop that tried to load a savestate that does not exist...? Using BizHawk 2.9.1. )
Feature request: Please allow using Basic Bot input with a hotkey in regular play. It could work exactly like autofire and autohold, and it could be called autorandom. When the hotkey is toggled, then autorandom input is applied, but it should not override player input, so that we can augment our manual input with random noise. (Now Basic Bot requires a movie, so I can't fool around with it while playing regularly or while trying to replicate glitches manually. I suppose autorandom could be done with a Lua script also, so if it is easy implement please include it as a default Lua script.)
Thank you for an amazing emulator!
The main hero Goldmoon does not give the boss-killing weapon away, but if she dies, then we can take the weapon. This could theoretically skip one character switch, but the problem is that it takes time to pick up the item, so almost all timesave is lost already. And now we have just another character in front who must waste a lot of time to avoid damage and survive to the end. Not working.
Edit:
I discovered two new improvements that probably work.
I assumed that the current WR speedrun had the optimal route, so I neglected to test an alternative route possibility. Around 5600 frames the old route goes to the right. We can go left instead to save 48 frames. Unfortunately the damage RNG is bad so the whole run needs to be redone to get this timesave implemented.
Another idea I had is that instead of switching the mage away after flying over the chasm we can let the mage get killed. I did not know it could be done so quick, so it is very probably faster to kill the mage but again it requires good damage RNG so that the mage dies from two hits. This will save about 100 frames due to not switching the dwarf to the front. This strat was also overlooked because human WR would not risk it.
I don't know what the total timesave can be in the end. Could be 48-160 frames. Not much different from the current run, but strangely complex. Not sure when doing a new run, probably not anytime soon. XD XD
Hi, Chamale!
More games that I know with very early input end possibility are:
Genesis: Hit the Ice (U)
Genesis: Hit the Ice (J)
TurboGrafx: Hit the Ice
TurboGrafx: TV Sports Hockey
I am not going to do those. Hit the Ice is fun, though. I did some test runs on Hit the Ice, and the enemy often made 9 goals while our computer player only made 1, so it looks difficult to win without making at least 1 goal before ending input, but then again for Hit the Ice there are probably billions of different ways to seed the match while losing fever frames than it takes to make a goal. I did not study the team compositions in these games at all, so maybe there is some overpowered setting. The Genesis version allows changing period length to 2 minutes, so it becomes much easier to solve.
If each game has a wiki page that the community can freely edit, then lots of problems are solved. Random featured articles should be on the front page, like Minecraft/Terraria/Doom TAS pages where the community will improve the article over the decades to incorporate literally everything that is known about the game and its TAS history. I think information would be more easy to find if each game page looked like a regular Wikipedia article.
Another idea: The front page might be good if it had an alphabetical list of all games immediately available. People often know what they are looking for so just give an alphabetical list of links to all games at once. Only add links to the list when a published TAS has been made for the game.
Wiki is a powerful tool for collaborative optimization challenges, one amazing example being the Busy Beaver challenge which is forever ongoing and very TAS-like: https://wiki.bbchallenge.org/wiki/Main_Page
Have a nice day!
Well, it turned out to be very easy to just record the whole thing from scratch in the PAL-M mode. Because I already knew the overall strategy and had common frame counts memorised. So I just quickly entered the same stuff. I did some different choices because the RNG was indeed different. The frame count is longer because for example this version loses 75 frames before the game is loaded. But it is much faster as the real time is 4:39 compared to 5:34.
Gameplay is almost the same, maybe 0-5 in-game frames difference timewise but since the situation is the same until somebody tests all combinations I am not trying to save a few in-game frames by testing a few more options for taking damage at a different spot without knowing the overall RNG behavior in-depth.
https://tasvideos.org/UserFiles/Info/638609820959280581
Sorry for not making the new version before the movie was published, because the epic publisher is a speedrunner at publishing. And I grossly overetimated how hard it would be to redo it in the other mode. XD
I would prefer the faster version to be the main movie and PAL version being offered as a side dish, but do whatever you think is the best way to go. Thanks.
Does PAL-M mode mean the same as this: Settings --> Sync settings --> Display type: Ntsc
Or is there some other setting that I need to change?
I can redo the whole run in the other mode someday, but I don't know when it will happen. Might take a year.
Thank you for the comment, fsvgm777!
Unless I made some mistake, I did it with the speed that the emulator set by default.
If somebody can make it work with the 60Hz version, then please do so. Maybe it synchs or maybe the RNG will be different?
Also, the genre may actually be "action-adventure" instead of RPG as I claimed. It feels very much like an RPG, though.
At around frame 3860 is it possible to face left after entering the doorway, avoiding a turnaround animation?
Some screen transitions have only one option, so the direction can't be changed.
ktwo wrote:
the main differences between the two versions
The overall game seems the same, but almost everything is different by a little bit. Most rooms have monsters in different locations, etc. Most notably the final room is longer in the SMS version. To me it looks like the NES version could save time by taking more damage and avoid using some spells, but I don't know about it since I have not played the NES version. In the SMS version I was able to avoid using any spells by planning the health and letting party members die. The completion time is similar for both games but that is a coincidence. The games are different. I was not able to save time by jumping in the SMS version, but maybe there are more tricks for both versions if we fool around more?
The biggest possible improvement that I know for the SMS version is to try different health management. This could save some frames. From the route plan you can see that I took 5 damage from the first human. The damage can be avoided by waiting for 8 frames or 1 in-game frame. I did not try this because I want to get damaged (just enough but not too much, so that the lesser heroes die and the main hero will be moved to the front before the final boss), but since this will change all future damage rolls and misses, it might turn out to be faster in the end. So if anyone wants to try it, your best bet is to start by not taking damage from the first human and see what your health situation is 4 minutes later. But then you can also choose to take or not take damage from the next dwarf enemy, etc. Taking damage also loses 8 frames, so we do not want to take too much damage. So it would be better to get lots of damage from a few hits, but again we probably must wait for 8/16/24... frames so in the end waiting for 16 frames to get 1 x 8 damage is same as waiting for 0 frames and taking 2 x 4 damage instead. Tricky, but far from the hardest problem you guys have solved. X-D
In the beginning of the video below the player says: "Okay, good, it doesn't look like I lost anything with that crash."
https://www.youtube.com/watch?v=bWyvjyOiMAI
Which makes me think the issue is random crashing and data loss.
This game may indeed be "speedrun proof" because nobody can finish a run without runkiller crashes.
But I think it is not "TAS proof" because TAS player can avoid crashes by playing differently. That is, if you can set it up.
I think the game may be nearly TAS proof because even if you can avoid crashes it is a big demotivator to retry random solutions just to avoid arbitrary crash. As for the game itself, it is difficult to find good random seed for procedural city generation because you do not know the long game without testing different setups. The game may not be "absolutely TAS proof" but just very very hard, which is practically the same thing, and the reason why I hope somebody will try to TAS it! :-D