Experienced Forum User, Published Author, Player
(198)
Joined: 12/3/2006
Posts: 151
Wow, that was very impressive. I did have a few questions about inconsistent turning but I suppose it was due to the luck manipulation on the mushrooms.
comicalflop wrote:
PS:
how do you change the author title in the movie file? the name of the file is correct, but in the Snes9x select movie thingy it only lists me as the author. is there a way to change that?
I think Maximus is busy adding this feature to his movie editor, but if it won't be completed in time, you could always make a new movie with the correct description and paste the entire movie in that new movie (although his does reset the rerecord count).
In any case, I'm looking forward to the final result :).
Experienced Forum User, Published Author, Player
(198)
Joined: 12/3/2006
Posts: 151
At the end of the game I think I have 30'ish Ice Balls and 10 Fire Balls left. This is because I wanted to make sure that I had plenty for every boss weak against Ice and Fire, so I had a strong alternate attack to manipulate luck. Unfortunately, I noticed (too late) that using Ice Balls or Fire Balls pull the exact same number of numbers from RNG as attacking normally, which especially later on typically does more damage. At this point I tried hexing the part where I bought them, but it turned out to be slower because of all the frame I had to add to synchronise it.
Edit: to clarify, the desynchronising was due to random lag frames, so getting the items in the correct order again would lead to even more lost time.
The Fire ring was grabbed to increase the damage of Fire magic used against the Rogue Flower and the White Dragons (both vulnerable to fire). Also, something I hadn't planned, but the 10 ATP from did turn out to make a (several seconds) difference on the Sinistral battles. In fact, if only one person had 1 less ATP during the fight with Daos, I would've needed to heal in the last round :p.
Experienced Forum User, Published Author, Player
(198)
Joined: 12/3/2006
Posts: 151
Bisqwit: It shouldn't, unless you're using the non-WIP version of Snes9x 1.43. If I recall correctly the official Snes9x 1.43 version does not have the "Use WIP1 Timing" option. That one is only available to the 1.43 WIP version, either modified by nitsuja or not.
Experienced Forum User, Published Author, Player
(198)
Joined: 12/3/2006
Posts: 151
Perhaps an easy way to implement this is to give posts a certain 'Vote flag'. Say I wanted to anonymously vote no on a movie because I don't like the author, I'd have to make a post in order to vote, revealing my identity.
Then again, this would perhaps discourage voting in general, and/or flood the topic with meaningless posts.
Experienced Forum User, Published Author, Player
(198)
Joined: 12/3/2006
Posts: 151
I already answered in the Submission forum topic, but I suppose it's better to discuss it here.
Some basic things about the RNG I found:
The RNG produces 55 number between 0 and 255 (or 00-FF) at a time. The initial values are always the same when starting the rom, and cannot be manipulated at all. These random numbers are stored in the memory addresses 7E0521-7E0557.
Everytime a random number is 'used' by the game, an internal counter, located at 7E0559, increments by 1. This counter starts at 0 and if it reaches 54 all random numbers have been used once. If another random number is requested, the RNG produces a new set of 55 random numbers. The way this happens is through bitwise XOR operations. random number 'i' becomes 'i' XOR 'i+31' and so on.
These random numbers seem to determine most of the randomness in this game; world map encounters, the way monsters move in dungeons, level up stats, how hard you hit etc.
In the case of the Ancient Cave I'm guessing these 55 numbers define a set of possible next floors. If I recall correctly there were about 6 different floors depending on the frame you entered the stairs, right? If that's the case there should be another memory address containing a counter between 0 and 5, which increments every frame.
That's what I can come up with right now. Perhaps next week when I run out of things to do (my life is so hollow since I finished the run :p) I'll try and find out.
Experienced Forum User, Published Author, Player
(198)
Joined: 12/3/2006
Posts: 151
I have found out a great deal about the randomness in the game, but I'm afraid I haven't found a way to explain the time dependand randomness of the Ancient Cave yet. But the thing that kinda discouraged me from trying is the fact that the weapons from the Ancient Cave aren't even that good. Mostly due to their 'hit all enemies with element X' IP attacks. So I'm not sure if it'd be worth.
While you're here I again want to thank you for starting your movie way back then. It really inpired me :).
Experienced Forum User, Published Author, Player
(198)
Joined: 12/3/2006
Posts: 151
Actually, behing the vines is a wall that needs to be bombed three times in order to break. The first bomb both clears the vines and weakens the wall once. So slashing the vines first isn't needed (and slower).
Experienced Forum User, Published Author, Player
(198)
Joined: 12/3/2006
Posts: 151
Meh, I started making some notes of the run, but it turned out to mostly "I do this to manipulate luck so I can do that", which probably isn't very interesting. So I guess I'll just point out the hard parts and possible flaws in the run.
Acmlm: That's some really useful information, but I'm afraid I will not be doing a level 0 glitch run soon/ever. Like you pointed out, the glitch causes some really ugly delays and I don't really have to level up anyways. On top of that, it would be less interesting watch probably.
erokky: It's true that grabbing the Pumpkin Jewel was kind of useless, but I don't think I could've predicted it. The reason why I didn't use it at Idura is because I don't need it. Yes, Holy Energy does only 3x damage where Head Attack does 4x, but it still takes as many attacks to finish him off. Not equipping the Pumpkin Jewel and using Holy Energy turned out to be faster. The same thing happens at the Tank battle. Using Head Attack does more damage, but not enough to make a difference. Also it would screw up my luck manipulation for that fight (which is close to perfect now).
The thing why grabbing chests affects luck manipulation so much is because regular actions, like swinging your sword, shooting an arrow, only pull one or two numbers from the RNG. Grabbing a chest however pulls a few hundred (!) numbers from the RNG.
Nitrodon: Now, that is some good news. I will watch it right away (or, after I submitted my movie) and comment it. I'm really curious about the tactics/luck manipulation you used.
Edit: I watched it and I'm really impressed. I can't believe I overlooked the Gloom Jewel like that. Also, nice luck manipulation on the Eagle Rock and Perish on the Lions. Next to picking up the Anger Brace for example, I didn't notice any big time savers (or maybe the Lizard Blow?). Great job :).
Experienced Forum User, Published Author, Player
(198)
Joined: 12/3/2006
Posts: 151
Aaaaand I'm done! The Doom Island battles took me a while to optimise, but I'm really satisfied with the result (especially the Daos fight). Also, turns out I overestimated the game completion time a bit, so it will be significantly faster than 3:10.
That's all I'm gonna tell you :). You'll have to see it for yourself when this movie appears in the Submission list in a day or two (since it'll take me a while to go through the entire movie again, making notes for the submission text).
By the way, I decided to not pause at the final dialogue since I figured most people actually watching this movie would be people who played the game, so they know the ending. Thanks for the feedback regardless Dromiceius.
Alright, a final question. Currently I ended the movie input when the dialogue goes into auto mode. However, I was wondering if the reference dialogue to Lufia 1 after the credits (and also after 'The End" has appeared) should be included as well. If so, I'd need to extend the movie input by quite a bit (5 minutes or so). So, the question is, should this be included?
Experienced Forum User, Published Author, Player
(198)
Joined: 12/3/2006
Posts: 151
Although I don't really understand all of the code you posted, I must say this looks really interesting from a TAS perspective.
Unfortunately I'm a Windows user so I can't try/use this yet, but I'm really looking forward to a Windows build :)
Experienced Forum User, Published Author, Player
(198)
Joined: 12/3/2006
Posts: 151
I'm glad you liked it!
Fortunately I was able to work a bit more on this run than anticipated and I'm already at the Gratze basement, meaning I have only one boss left before I can go and fight the Sinistrals. With the clock at 2:20'ish now, I'm guessing 3:10 should be the final completion time.
Also, I noticed that after you finish the game and the cinematics start, you first have to scroll through some dialogue before the input can be ended. The obvious choice would be to turbo through all this text, but in my opinion that will be a bit of an anticlimax. I was thinking of waiting a short time so you can actually read the text, but I'm not sure what the general opinion of this would be. After all, the main goal of this movie is to finish the game as fast as possible. It would be great if I could get some feedback on this one.
Experienced Forum User, Published Author, Player
(198)
Joined: 12/3/2006
Posts: 151
Anon wrote:
It sounds interesting enough for an alternate TAS (much like the various glitched The Legend of Zelda runs), but seeing as it changes a major element of the game, there would more than likely be two published runs, instead of one replacing the other.
That sounds about right. I have gotten the glitch to work now, but it seems a bit random when it happens. The result was a character at level 0 with all stats at 999, except for GUT, since it's max is 199. Pretty much like wearing an egg ring.
That said: Update!
This WIP ends at Dankirk just before the Labyrinth and is 1:45:50 (380974 frames) long. I suppose a game completion time of about 3:30 should be possible :).
EDIT: Moved WIP to Microstorage.
Experienced Forum User, Published Author, Player
(198)
Joined: 12/3/2006
Posts: 151
Ah, that makes sense. I guess in my case the damage increase cause by the lower MGR was compensated with a lower 'fluctuation' variable. I didn't know Ironknuckle's FAQ contained this kind of formulae, I guess I have to re-read it again :).
As for the Camu Jewel, I can only sympathise with you. Manipulating drops has got to be the worst part of this game, toolassisted or not. I'm not sure if this can help you, but here is some things I noticed about luck manipulation in this game.
For luck manipulation purposes I monitor the following two memory addresses: 7E0559 and 7E0522. Basically the value of these two addresses seem to determine every little randomness in this game. This is what makes it really hard to manipulate luck in this game; every random thing is connected to each other, you can't change them individually. I learned this the hard way when I tried the Lake Cave for the first time. I think I had to redo this part over 10 times to make sure I got a fast Eagle Rock drop, Catfish Jewel, was faster than an eagle, got single enemy encounters in the Secret Skills Cave and much more.
The value of 7E0559 starts at 0 and goes all the way up to 55. This happens when take a step on the world map, swing your sword, open a treasure chest etc. When it reaches 55, it goes back to 0 and the 7E0522 value changes. These new values seems to be predefined.
Anyways, this helped my find some action that do and don't influence luck manipulation.
Do affect luck manipulation:
Taking steps on the world map.
Taking steps in dungeons while near to enemies with 'random' movement ( jellies, eagles).
Swinging your sword, shooting an arrow etc.
Cutting down bushes.
Healing through potions or spells in the menu.
Spending time in towns.
Using different battle commands.
Entering dungeons/taking the stairs.
Don't affect luck manipulation
Taking steps near enemies with a fixed movement (baby frogs, scorpions).
Buying different types of items in shops.
Spending time on the worldmap, menu, shops.
Refilling through the HP/MP tiles.
Equipping weapons/armor/rings/jewels.
Doing different amounts of damage with the same action (when having a higher STR or INT for example)
Spending time in conversations with NPC's.
Using Warp/Escape.
I think that's most of them, at least the ones I use to mainpulate luck. I hope these are of any help to you and good luck with that jewel.
Experienced Forum User, Published Author, Player
(198)
Joined: 12/3/2006
Posts: 151
erokky wrote:
Holy bugger. I can't wait to see where these items you get save time.
Power potion, Hi-potion, Witch Ring, Jute Helmet, Mind Ring; they're all gonna save time eventually, right?
Power potion was for luck manipulation, in order to get an easy drop on both the Eagle Rock and the Catfish Jewel. Witch Ring turned out to save one round on the Regal Goblin alone (ups the Mega Quake damage by ~10 points). Together with the Mind Ring another 2 rounds can be shaved off the clown fights in the Treasure Sword Shrine. Also, the Mind Ring's IP is used a few times for luck manipulation.
At first I was going to pick up the Jute Helm for it's IP attack (-MGR on an enemy) but it turned out to have very little effect on the spells I use. It seems that when a boss' weakness is for example lightning, all lightning spells ignore the boss' MGR. But as you can seem the Camu fight is near perfect and skipping the Jute Helm would change the boss' behavior and would turn out to be slower.
The Hi-Potion is pretty much the same story as the Jute Helm, not picking it up would turn out to be slower due to the effect on the boss' attack pattern.
Atma wrote:
The reasoning for using the E version over the US version was that there are graphical glitches present in several places during the game, which would detract from the viewer entertainment (at least in my opinion).
Also, since you're going to use the US version, why not just use the level 0 glitch?
I know why the E version was used initially, but seeing how much faster everything happens in the US version sold me. The only visible graphical glitch in the run would be the Dual Blade Shrine since all the menus are handled so fast and I don't think it's that bad.
The level 0 glitch sets your level 0 (duh :p) and causes you to do insane damage, but I'm not sure what happens when you get hit. I heard about level 0 glitch, and I tried to reproduce it, but I failed. But on second thought the glitch would take quite some entertaining elements away from the run like the near death boss fights and IP managment.
Experienced Forum User, Published Author, Player
(198)
Joined: 12/3/2006
Posts: 151
I've been hesitating for a while to post here, since I wasn't sure if quietkane is still working on his run and the last thing I wanted to do is demotivate him. But seeing that it's a year and a half since the last update I'll do it anyways.
http://dehacked.2y.net/microstorage.php/info/1126162887/Lufia2_WIP1.smv
Inspired by quietkane's work I've started making my own speedrun of this game with the knowledge of major timesavers, almost all discussed at the SDA forums (props to Dragondarch's near perfect strategy planning!). These include:
More Boomerangs/Items in general.
Skipping Insect Crush + Fire Dagger
Using Elemental attacks on several bosses (Gades, Idura, Lions)
Using the US ROM.
Technically, the timing difference between the roms shouldn't matter at all, since the runs are timed by frames. But in long runs like this one, the difference between 50 and 60 fps can make a total difference of ~30 minutes, which makes watching the movie more bearable. Also, watching Maxim move at high speed is (in my opinion) generally more impressive.
This run ends at Narcysus and I will not be going through the Ancient Cave seeing what it did to quietkane :(. It's just too damn scary.
Most of this run was recorded during the last two weeks, and the next four weeks will be pretty busy for me. So progress will be a bit slower, but I promise I'll finish this one.
EDIT: Moved WIP to Microstorage.
Experienced Forum User, Published Author, Player
(198)
Joined: 12/3/2006
Posts: 151
adelikat wrote:
Congratulations on having your first run accepted.
Well thank you :).
Although it will probably take a while for this to get published, I'd like to point out that the movie file hasn't been updated yet. I'm not sure if anyone else besides Bisqwit can change the movie file in the submission, but if so, the updated movie can be found here.
EDIT: Moved fixed SMV to Microstorage.
Experienced Forum User, Published Author, Player
(198)
Joined: 12/3/2006
Posts: 151
Zalien wrote:
Alright, after reading some replys I'm leaning towards yes. But, my one question is, is it faster to run from point A to B normally, or to turn around and shoot and use the recoil to move quicker?
Both are equally fast, since horizontal speed is capped at 3 pixels/frame. While walking, turning around does not influence your speed if you immediately jump or shoot after it. In the air however, turning around does slow you down a bit, so doing this is avoided.
Experienced Forum User, Published Author, Player
(198)
Joined: 12/3/2006
Posts: 151
Bob A wrote:
Also, one thing i noticed was that the 3-3 boss could have been done at least 60 frames faster without using the blitz attack (or whatever it's called). Though, whether this can be made worth it in the long run i'm not sure.
Oh, I forgot to reply to this one in my last post. As said, I finish the game with only one bullet remaining for the bomb gun and none for the anti-matter gun. Using special ammo on this boss would make me run out of ammo during the final boss. The nutty attack (I think it's called that way) is used on this boss because he is one of the few bosses without invulnerability time after a hit, which makes him take 1 damage every frame.
Experienced Forum User, Published Author, Player
(198)
Joined: 12/3/2006
Posts: 151
Alright some explanations:
Only thing that stuck out to me was near the end, a couple hits were taken that caused some pausing, but I don't know the game at all so I'm guessing it was faster than whatever alternative there is.
I'm pretty sure you mean the times I 'fall' on the green laser beams. On the first beam, I had the alternative to turn around a few frames and blosck the beam, and then move further, but as you can see, right after the point I hit the beam, there's a martian walking around that cannot be dodged. So it actually saves time to eat the laser beam and abuse the invulnerability time to walk through the martian.
The second laserbeam is less obvious, but right after I hit the beam, I have to change my walking direction. Fortunately, getting hit by a laser beam instantly resets your horizontal speed to zero, so this allowed for an extra quick turn.
The third time I get hit in the last level is right after the point where I fly underneath the level and jump on the martian. This is one of those enemies without a ceiling above them, so I couldn't use the dodge glitch. Also I did not have enough fuel to fly further underneath the platforms, so that wasn't an option either.
I wonder if this could be avoided through luck manipulation
There is no luck manipulation in this game. Everything seems to be predefined. The time enemies spawn (and start shooting) is triggered by getting within a fixed distance with them, so the only option to influence the spawn time is to slow down on the way to them. The only really 'random' thing appears to be the lag frames, which makes the game really hard to hexedit :(.
Also you seemed to stop up every time you wanted to use jetpack. Game mechanic?
Yep. Standing still gives you a negative horizontal acceleration of 0.5 pixel/frame/frame, where walking the opposite direction only gives you ~0.22. Also note that while using the jetpack, it is impossible to change direction and the horizontal acceleration is capped. Therefore it is necessary to achieve the required horizontal speed prior to using the jetpack.
I voted meh on account of that it violates the rule that the game should be played on the hardest difficulty when doing so is more entertaining.
Fair enough, though the submission text actually states normal mode was chosen for extra entertainment. As for the differences between hard and normal mode:
As you can see, the main weapons of this run (normal, antimatter and bomb gun) do significantly less damage. Also, if you watch closely you can see that this run ends with only one bullet left for the bombgun and the final world is entered with no cash remaining. Doing a run on normal mode would lead to using suboptimal weapons which would be extremely hard to plan. Hopefully this will shed some light on the normal mode issue.
Also, thanks for the positive feedback so far :).
EDIT: Also, the rule about playing at hardest difficulty level isn't really a rule, but rather a guideline seeing that it's posted here: http://tasvideos.org/Guidelines.html ;)
Experienced Forum User, Published Author, Player
(198)
Joined: 12/3/2006
Posts: 151
Unfortunately, the last frame of input was somewhere lost in the progress of submission. I've PMed Bisqwit the updated movie file, so hopefully it'll be updated shortly.
Oh, and I hope you enjoy the movie :)
Experienced Forum User, Published Author, Player
(198)
Joined: 12/3/2006
Posts: 151
Are you by any chance using the European ROM? I noticed that on my (European) console I'm not able to make the jump as often as on the US ROM (realtime emulated), so there may be some timing differences between the two.
This is also seen in castle 1; on my console I'm able to jump over the long lava pit just before the door, without grabbing the fence, while on the emulator with the US version, I've never succesfully done it.
Experienced Forum User, Published Author, Player
(198)
Joined: 12/3/2006
Posts: 151
I did buy 20x Bomb Gun Ammo since I need 12 of them at the boss and they go in stacks of 10. The 10 Anti-Matter Gun Ammo I buy is not used in the first level. But, since buying one stack of ammo can be done without losing time and I need more than 10 during the second mission, I decided to buy one stack in advance to save one frame.
About the hard-mode issue: it's true that it will be more challenging to TAS, and maybe more entertaining to watch, but I'm still quite new at this myself, so I think I am going to finish this run first. I already did some research on the ammo usage, and playing at hard-mode will require alot of planning since you simply do not have enough cash to buy special ammo required. Comparing the time gained versus ammo spent on different bosses will be easier while having a normal-mode run done.
Edit: Actually, the Bombgun does 48 damage, while the Anti-Matter gun does 44. This is compensated by a large 'cooldown' between shots on the Bombgun, but since the first boss' invulnerability time after a hit is longer than the cooldown, the Bombgun does more damage here.
Experienced Forum User, Published Author, Player
(198)
Joined: 12/3/2006
Posts: 151
Zurreco: Thanks for the link, though it's a shame it didn't contain any usable information. Also, could you elaborate on what you meant with 'super speeds'? As far as I know, the vertical speed is capped at 8 pixels/frame and horizontal speed at 3 pixels/frame (unless you're standing on a moving platform, or sliding down a slope).
Bob A: The only impact 'hard mode' has on the game, is the amount of damage done by Daffy. So when playing hard mode, I thought boss battles might get more boring instead of interesting, since you're basically doing the same thing, but taking more time.
The short answer to the up/down trick is whenever you press up or down, your speed remains the same for 2 frames. However, it also takes 2 frames to change Daffy's sprite from standing to crouching/looking up. So when up and down are pressed every other frame, Daffy keeps moving at the same speed, while his sprite is stuck either looking up, or crouching. The latter can be nice for dodging object, but, of course, it looks kinda cool too :).