Experienced Forum User, Published Author, Player
(241)
Joined: 5/1/2004
Posts: 4096
Location: Rio, Brazil
i got my hands full right now, and this game is kind of unforgiving (i did my best, i thought it was perfect, and still some levels were very improveable... even comparing to real times). so i don't know about working on this game...
i might try finishing it, but... you know, like i said i got other priorities right now (it may take an year before i want to get back to this game)
Experienced Forum User, Published Author, Player
(241)
Joined: 5/1/2004
Posts: 4096
Location: Rio, Brazil
So my guessing was correct, this game limits the speed with a simple code. I located the memory address that does this and could force it to never limit the speed. So the game plays freely at the max speed the SNES can do, at any given moment, so the speed varies depending on amount of enemies etc. My guess is they programmed the game like this then limited the speed to the lowest value they found while playing, so it would be consistent.
I don't know if this game would be interesting at all to see played normally, that's why i'd like to see a TAS of the japanese one (Return of Double Dragon) played on hard, 2 players (either mode A or B is fine), and using this speed cheat. I don't know how to convert it to a gamegenie code, but the memory address is this one:
"7E0018" and the value must stay as "00"
hum.. i guess the "pro action replay" code for this would be just "7E001800"
zsnes also gives this "PV" which i don't know what is, but is showing "01"
The japanese game is more complete, you can choose hard mode, and the game has one more level. But it would be faster than super double dragon because even on hard mode, the enemies have less health.
Experienced Forum User, Published Author, Player
(241)
Joined: 5/1/2004
Posts: 4096
Location: Rio, Brazil
just watched this, i found it very funny how you played around with the generic ninjas, like they were nothing at all :) good job.
feel sorry for the lag though, but yes vote.
Experienced Forum User, Published Author, Player
(241)
Joined: 5/1/2004
Posts: 4096
Location: Rio, Brazil
AnS wrote:
Good old game, well-made run. But it could be much faster if you find a way to avoid lags in places with many objects on screen.
Here I fixed your file with hexeditor.
http://shedevr.org.ru/temp/takahashi-fixed.rar
It took only 10 minutes and 2 bytes to modify. Also I cut some useless bytes at the end of file, and now movie is 1 second less. :)
Interesting fact - fixed movie works perefctly with good and bad rom.
wow..... i am speechless. thanks!
I fail to understand how could the movie stay in sync with both roms!? what did you do? did you just remove a frame? or did you edit the header? O_O
I haven't detected any lag in this game. Instead of lagging, it handles too many sprites on the screen with flickering. So it doesn't affect time.
The 1 second at the end, you're totally right, i don't know how i let that happen. Following adelikat's suggestion I'm gonna resubmit this just to clean everything up.
Again, thanks a lot.
Experienced Forum User, Published Author, Player
(241)
Joined: 5/1/2004
Posts: 4096
Location: Rio, Brazil
how is it that my route goes to the basement twice? i'm refering to basement as just the place where mips is. like i said, the distance from the vanish cap entrance slope (near 2nd bowser door) to the lobby is much shorter than that to any of the 3 levels (SSL, LLL, HMC). That distance is what i'm trying to avoid walking twice. I'm going to go there only once, then i'll exit it and be back in the lobby, instead of walking it back.
edit: nice analisis of the times, but i have to say that stopping the clock at the moment i touch the star is not the best, since i do not touch the star as soon as spossible, but i try to make mario stop his celebration dance as soon as possible. so, stopping the count by the moment he touches the ground after grabbing a star would be better than just as he touches the star.
Experienced Forum User, Published Author, Player
(241)
Joined: 5/1/2004
Posts: 4096
Location: Rio, Brazil
Luke wrote:
So I guess in this version of Excitebike, the game won't let your speed exceed 04FF (where 0340 is your unassisted top speed), or about 5 pixels per frame. So it'll be a lot slower than the regular NES version, where your speeed would just keep increasing every time you go over a ramp.
oh really.... well then it's not as interesting.... ¬¬
Experienced Forum User, Published Author, Player
(241)
Joined: 5/1/2004
Posts: 4096
Location: Rio, Brazil
i think it has been possible to download them at emule for a longtime now. i'm just not sure if someone intentionaly put them for upload or if it was some automatic media search.
Experienced Forum User, Published Author, Player
(241)
Joined: 5/1/2004
Posts: 4096
Location: Rio, Brazil
Bag of Magic Food wrote:
What memory would you watch?
The character's speed address, which doesn't have a clear maximum value. Sometimes it stays at 80, sometimes at 81, sometimes at 82... I believe that's highly manipulatable to use always the maximum value.
Experienced Forum User, Published Author, Player
(241)
Joined: 5/1/2004
Posts: 4096
Location: Rio, Brazil
Spacecow wrote:
It looks like it's just two people playing the game as fast as they feel comfortable with (not especially fast compared to any real run), with Less Than Jake playing in the background and people randomly cheering for who knows what reason.
I'm kind of confused as to why that video is in any way interesting or noteworthy (and on any sort of video sharing site) to be honest.
It's just a race played once. How is that hard to see?
Experienced Forum User, Published Author, Player
(241)
Joined: 5/1/2004
Posts: 4096
Location: Rio, Brazil
Wouter Jansen wrote:
shouldn't you be getting the key on the first possible frame? I don't think there's a last possible frame anyway, since you could just take however long you want before you pick up the key. :p
he meant taking it on the latest possible frame without losing time.
Experienced Forum User, Published Author, Player
(241)
Joined: 5/1/2004
Posts: 4096
Location: Rio, Brazil
I dont know if you understand, but this last bowser fight was not the definitive one. I remade it but i didn't release a WIP. It was bad because bowser didn't flail his hands madly at the beggining of the fight, so it was slower than spezzafer's bowser fight.
Experienced Forum User, Published Author, Player
(241)
Joined: 5/1/2004
Posts: 4096
Location: Rio, Brazil
ok you have the probability based on the amount of frames to try? because you had 1000 for probability, and you said it does ramdom search for 1000 frames... so i don't think "probability" is the best word for it, maybe you could say "appearance rate" because it's the exact amount of frames that button will be used? but in this case i don't know how usefull it would be, maybe for numbers like 1 or 2 but you can't predict if it will be using the button 40, or 41 times. maybe this could be loosened up, like add a margin:
frames to try: 60
"A" button appearance rate: 20-50
"B" button appearance rate: 60
"S" button appearance rate: 0
"up" button appearance rate: 0
etc.
Experienced Forum User, Published Author, Player
(241)
Joined: 5/1/2004
Posts: 4096
Location: Rio, Brazil
FILIM0NAS wrote:
0)Generally: At the beginning you always long jump in a slightly wrong direction so that you don't go 100% straight ahead.
When it matters, i try to get the angle the best i can do, but sometimes it doesn't matter (like when going for the elevator).
FILIM0NAS wrote:
1)Big Bomb Omb. It seems that your test run was faster, since you didn't get stuck in some places before going up the mountain. You start the BLJ later so that you don't gain the max speed from the elevator.
on the test run i acquire much more speed so i have stop on various points. on this new version, i don't acquire as much speed (on purpose) so i can go without stopping, but this way i couldn't be sent straight to the rolling balls part, so i had to travel like that. Don't worry, i checked and this was 2 frames faster.
FILIM0NAS wrote:
2)100 coin route. You use 2x diving instead of long jumping. Are you sure this is the fastest way? Especially when you go up the mountain.
2x diving or longjump are practically the same in speed, i only choose one or the other because they end in a different length. if i used longjump there, i'd have to run for some seconds untill i reached the corner so i could turn and do another longjump.
FILIM0NAS wrote:
3)100 coin route. You use the cannon to blast to the pole after getting the 100 coins - star. Try to triple jump off the red ! block on the floating island. It may be faster this way.
i tried all the possibilities, and this was the fastest. (triple jump from ground, triple jump at the ! block, and the cannon.
FILIM0NAS wrote:
4)Red coins. Nice idea to fly down to the 5th red coin. Why don't you fly all the way down to the 6th coin?
in my testing, sliding was much faster than flying, so i landed as early as i could on the slope
FILIM0NAS wrote:
5)Bowser. The camera in the beginning was IMO better in your test run. Press the C down for a farther better view. The way you got the key in the test run was also better (you touched the key in the very last moment after diving)
the bowser fight is different now on my unreleased WIP, and it's very close to both mario and bowser. it looks stylish instead of informational. i will do a different camera angle for the 3 bowsers. (i got the key on the last possible frame too, on this unreleased WIP)
Experienced Forum User, Published Author, Player
(241)
Joined: 5/1/2004
Posts: 4096
Location: Rio, Brazil
Phil wrote:
Also, in many games, you need to press some buttons for 1-2 frames for the action to work then you need to wait x frames to redo that action. So, what I want, is something that is configurable.
Ex. If bot press A, then wait 20 frames before it tries to repress it. It should eliminate many and many possibilties.
that sounds a little complicated without planning on how to set it up. let me think...
ok how about this: a field for each button so you can write how many frames it has to remain pressed when it's pressed. And another one saying how many frames it must remain unpressed, after it's pressed. like this:
"A" - 1 - 25
"B" - 2 - 0
"S" - 1 - 5
"up - 5 - 0
etc.
maybe this set of rules is troublesome to implement though, i'm not sure.... if it's doable then at least it gives something to work with.
also, options that i think there should be on the basic bot, (besides what it already has, which are nice):
- goal as a memory address value, or maybe 2 addresses values reached.
- set max amount of frames to try achieving the goal. No need to move for 2000 frames for the first thousand tries if you know it can be done in less than 50 frames, for example.
edit: i like the probability thing though. if i put 0 or 1000 it will respectively never press the button or keep it pressed, right? i think it looks good.
Experienced Forum User, Published Author, Player
(241)
Joined: 5/1/2004
Posts: 4096
Location: Rio, Brazil
maybe he's at battletoads level 3.
danny, do you own a real nes? can you really tell that fceu is running too fast, or are you going by memory? in-game clocks are most of the time unrealistic anyway
Experienced Forum User, Published Author, Player
(241)
Joined: 5/1/2004
Posts: 4096
Location: Rio, Brazil
Phil wrote:
I need someone to tell me what are the values for the mummy bosses's energy.
Found these:
01AA or 01A9 - both mummies summed
058B - mummy that starts on the left
058C - mummy that starts on the right
good luck.
note: 01A9 and 01AA behave differently... 01A9 gets deducted right away according to the amount of damage dealt. 01AA on the other hand, is lowered by 1 per frame untill it matches 01A9. maybe this could tell more about the critical hit. or not.
