Username:
B
I
U
S
"
url
img
#
code
sup
sub
font
size
color
smiley
embarassed
thumbsup
happy
Huh?
Angry
Roll Eyes
Undecided
Lips Sealed
Kiss
Cry
Grin
Wink
Tongue
Shocked
Cheesy
Smiley
Sad
<- 1  -   of 53 ->
--
--
List results:
Search options:
Use \ before commas in usernames
Good luck.  Time off is good. 
<(^_^)>
Good luck with IRL stuff Smiley we'll miss you, but it's important for you to sort things out first before coming back to speedrunning
Edit history:
andrewg: 2013-08-19 07:08:33 pm
andrewg: 2013-08-19 07:07:33 pm
andrewg: 2013-08-19 07:07:07 pm
andrewg: 2013-08-19 07:06:40 pm
andrewg: 2013-08-19 07:06:37 pm
andrewg: 2013-08-19 07:06:37 pm
andrewg: 2013-08-19 07:05:53 pm
andrewg: 2013-08-19 07:05:38 pm
andrewg: 2013-08-19 07:05:31 pm
andrewg: 2013-08-19 07:01:11 pm
Hi! I'm andrewg!
Warpless: Practice for SMB1 warpless will continue, but be more minimal than before. I will likely be running this at AGDQ and I want the run to go smoothly (well, very well practiced at least). I'm going to start finishing every single attempt I do. I'm going to create a spreadsheet of times as well so I can really optimize this.


Warps: I'm cutting back on this. I doubt there will be improvement to this for some time. However, this will always be a project I'm working on until I reach 4:57. It will happen some day.


Real Life: Pursuing career goals. Hoping to change many aspects of my life for the better. Working toward some of my other hobbies more. I decided that I just need to balance my time better instead of stop gaming altogether. thumbsup
Holy shit, I had a 8-4 where I only lost 4 frames (2 on the walljump, 2 on the turn). My previous best was 11 frames lost. Insane luck on the turn and then I didn't screw the swimming section up (90% of the time I lose frames on the exit). But I started from a savestate.
I still couldn't get a perfect 4-2. My best is two framerules behind.

Concerning complete runs I had a run die to Bowser, which would have been 5:07 with 8s lost in 4-2. Without the death and a perfect 4-2 it would have been 4:59.
Edit history:
andrewg: 2013-08-24 02:36:59 pm
andrewg: 2013-08-24 01:25:51 pm
Hi! I'm andrewg!


Getting there Smiley

I'm gonna get that 19:05 eventually...


...and a 4:57 maybe.
2:39.72 in -3 stage ending. Bad run, but tell me, if you think the route/tricks are good.
Congrats andrewg!  Impressive as always.
Improved by 0.09s. This teh urn. Kreygasm
Seriously, this run still sucks. But I'm streaming attempts now on www.twitch.tv/i_o_l and should get a good run soon.
Edit history:
andrewg: 2013-08-25 08:05:14 pm
andrewg: 2013-08-25 08:04:38 pm
Hi! I'm andrewg!
I THINK the flagpole glitch can be done underwater?

The 19:15 run I did seemed decent, so I was actually surprised I had only improved 1 second over my best. I guess I made a lot of small mistakes throughout the entire run. I still wager that my goal time of 19:04 is possible, but is going to definitely require luck and a large amount of planning (and predicting) castle patterns.

Improving that run by 11 seconds kind of seems impossible, but I still think it will happen in time.
Hey, I'd like to learn more about how monsters spawn in this game. Maybe you could link me to a good resource? Like, in 8-3, what determines where the hammer bros are when you first see them? RNG, or something else?

A clean run finally. Smiley

Quote from andrewg:
I THINK the flagpole glitch can be done underwater?

You're right, it's very hard and inconsistent, but reasonable and saves 5 framerules. Once I have 4:59 with warps, I'll go for 2:36 minus world.

Quote from willows:
Hey, I'd like to learn more about how monsters spawn in this game. Maybe you could link me to a good resource? Like, in 8-3, what determines where the hammer bros are when you first see them? RNG, or something else?

You're probably best of asking in the tasvideos.org forums. They're mostly consistent, when you make a savestate at the beginning of the level. But I don't think it is possible to manipulate them into favorable positions.
Here's how I do it:
the first pair:
Run the lower path. If there is one hammer bro at the bottom, jump on him and ignore the other one. If there are two, reset, but that's very rare.
the second pair:
Do a high jump against the top block just before the hammer bros. This will slow you to avoid the fireworks and get a perfect time. While Mario is slowed down, look where the bros are and try to avoid them.
the rest:
Literally just hope.

I'd like to hear how andrew and Kosmic do it though.
Edit history:
Mr. K: 2013-08-27 12:03:37 am
Mr. K: 2013-08-26 11:59:21 pm
Mr. K: 2013-08-26 11:58:12 pm
Mr. K: 2013-08-26 11:57:08 pm
Mr. K: 2013-08-26 11:54:53 pm
Mr. K: 2013-08-26 11:53:46 pm
Mr. K: 2013-08-26 11:53:07 pm
Yes
Speedrunning Super Mario Bros. isn't really my thing, but I figured out how the Hammer Bros. work and someone more knowledgeable than me is welcome to chime in with corrections and strats on how to actually deal with them.

Where are the bros. when they first appear?
The game's Random Number Generator which includes a network of successive addresses beginning at:
07A7 and ending at 07AD

What seeds the RNG?
It seems to be the product of some mathematical calculations and it changes drastically on ever frame, it looks like some sort of division by 2 and something else.  It ignores user controller input and in game actions.  It's running continuously from power on, during menus, and even during pauses.  It doesn't look like it can be manipulated in real time. 

When they throw hammers
Hammer throws are determined by a counter that may be at any one of the following RAM addresses:
03A2
03A3
03A4
03A5
How the address behaves:
When the hammer bros. gets far enough on the screen, a counter of 28 is set.  It counts down by 1 every frame until it reaches 0.  When the counter reaches 0 the hammer bros. will throw a hammer and the cycle begins again at 28.  Refer to the screenshot below.  On this frame, the 28 frame countdown starts for the Hammer Bro you see in the screenshot.

When they jump
When a hammer bro jumps is determined by a counter at one of the following addresses:
003F
003C
003D
003E
As soon as a Hammer Bro spawns, it is given a counter whose number is seeded by the RNG.  It counts down by 1 every frame until it reaches 0 and then he jumps.  After the counter reaches 0 the RNG determines the next number the jump will count down from.  The highest number the counter can be is 255.  The lowest number I've found in testing was 192.  Please note, a hammer bros. is considered spawned before he reaches the screen.  This is different from the hammer counter where the hammer bro has to be visible. 

Interesting quirk:  The game recycles RAM addresses and doesn't clear them mid stage.  This means if a hammer bros. is scrolled off the screen or killed, its jump counter may be passed on to another Hammer Bros. later on.  This makes it seem as if the counter is somewhat set at the beginning of the stage.  This is mainly relevant in 8-3 because of the number of Hammer Bros.  The jump countdown for all Hammer Bros. is cleared at the end of the stage and so it does not bridge over to the last Hammer Bros. in 8-4.

Edit: I've got a RAM watch if anyone is interested in fiddling around with this with FCEUX.

Attachment:
Quote:
What seeds the RNG?
It seems to be the product of some mathematical calculations and it changes drastically on ever frame, it looks like some sort of division by 2 and something else.  It ignores user controller input and in game actions.  It's running continuously from power on, during menus, and even during pauses.  It doesn't look like it can be manipulated in real time.

That would mean, if you start from power on and have a perfect run, you will always have the same hammer bros pattern, because of the frame rule.
twitch.tv/ccaakk
There's also a frame rule on the title screen, so you would have to press start immediately, or whenever gives you the best pattern.
INTJ
Maybe one thing you could do is find an earlier enemy, look at his movements and try to determine from that what the hammer bros are going to do.. That's a very far stretch, but sometimes games happen to coincidentally work out like that
Edit history:
Mr. K: 2013-08-27 09:19:00 am
Mr. K: 2013-08-27 09:17:54 am
Yes
So there's good news and bad news with RNG manipulation.

Good news: I made a state after pressing start on the title screen, began the game and messed around and checked the RNG on frame 1000.  I then reloaded the state began the game on a different frame and did absolutely nothing and the RNG was the same value on frame 1000.

Bad news: You'd have to have to be frame perfect about pressing start on the title screen and you'd have to have to hit the frame rules the same way (good or bad, let's say good?) in order for this to work.  In other words, if you are in the EXACT same spot on the same FRAME relative to a reset the RNG will be the same.  The RNG calculations skip a frame on lag frames so I'd also like to know if the game lags in the same way every time?  Is it always like a frame or so on stage transitions?  The problem here is that you wouldn't really know if you hit the right frame until you encountered a hammer bros. and it would be a royal pain to figure out if what you did for that run was ideal.  I still don't know how the game determines which hammer bros. is given which RAM address and how it decides to recycle them.  I can push it further if there is some interest, but at the moment, it's not looking good.
Edit history:
andrewg: 2013-08-27 10:21:50 am
andrewg: 2013-08-27 10:20:18 am
andrewg: 2013-08-27 10:17:00 am
Hi! I'm andrewg!
Kosmic and I have been using this to our advantage for both warp and warpless runs as of just a few months back. In warpless, you can know exactly how stage 3-4 will behave assuming you've played perfectly up to that point. The only requirement is that you've pressed "start" at the title screen within the first frame rule from power-on (probably about 10 frames of leeway).

It does work this way with warps as well, where you can actually know the exact pattern of the hammer bros, etc, but you need to know what frame rule you're on (and have held "right" the entire time during that stage). Any stage where you're making any slowdown whatsoever will prove to be random (For warps only 8-2 and 8-4 has any real randomness that could mess you up it seems).


There are certain enemies (like bullet bills), that don't always seem to behave the same like most other elements of the game. This is something I'd be interested in, knowing how the bullet bills work exactly.



For warpless, I'm in the process of figuring out various patterns based on what frame rule I'm on, but it's a bit difficult. Even knowing how the game will act sometimes doesn't help you if the pattern is unfavorable.


EDIT: Also, I'm not sure what causes those lag frames, and I'm assuming that's the cause of why things don't necessarily line up to exactly 21 frame intervals at times.
Yes
Regarding Bullet Bills, initial testing is showing that it is determined by the game's RNG.  Each cannon has some value or range such that when the RNG reaches it, it fires.  One cannon fires when the RNG hits a certain value, but that same RNG value won't make a different cannon fire.  It doesn't seem to have a countdown timer to fire like the hammer bros. do to jump.  Freezing the RNG address on a certain value causes the cannons to never fire.    I'll try and get you a more precise answer when I have the time.
Hi! I'm andrewg!
Dear 4:57, I hate you.

Love, andrewg
How can you hate something that doesn't exist (yet)?
Edit history:
Blubbler: 2013-09-11 11:36:34 am
Blubbler: 2013-09-11 11:32:51 am
Blubbler: 2013-09-05 06:41:41 am
Blubbler: 2013-09-05 06:41:32 am
Blubbler: 2013-09-05 06:41:15 am
Blubbler: 2013-09-04 01:33:38 pm
Blubbler: 2013-09-04 01:25:24 pm
Blubbler: 2013-09-04 01:23:18 pm
I made a quick rough list of different SMB1 versions to see which one is the fastest. But it's been rereleased so many times, I'm sure I missed some versions. With most of them I'm not 100% sure, if they may be faster or not.

the same game
Famicomsame speed
FDS versionsame speed
PAL versionfaster if flagpole glitches are possible and used, slower otherwise
Hong Kong NESslower due to framerate
with Duck Huntsame speed
with Duck Hunt, World Class Track Meetsame speed
with Tetris, Nintendo World Cupsame speed
Nintendo World Championshipsame speed ?
GBA Classic NES Seriesslower due to framerate ?
GBA Famicom Minis collectionslower due to framerate ?
Wii Virtual Console versionslower due to framerate
3DS versionslower due to framerate ?
GBA special edition?
Wii U VC?
Animal Crossing unlockableI don't think unlockables in other games should count. But due to the framerate it's probably slower?



unsure if port or different game
Super Mario All-Stars SNESslower due to different engine ?
Super Mario All-Stars/Super Mario Worldslower due to different engine ?
Super Mario All-Stars Wiislower due to different engine and framerate ?
Super Mario Bros. Deluxefaster due to framerate and different engine
PlayChoice-10 Arcade version?
VC version bundled with red Wii?
FamicomBox version?



different games with a similiar name
Mario Bros. II -  Commodore 64 ~ Based on Mario Bros. for Game & Watch
Game & Watch Gallery 3
Game & Watch Gallery 4
Super Mario Bros. (special edition) - yellow-cased version of Super Mario Bros. that came in a plastic box ~ part number is YM-901-S ~ Game & Watch
Nintendo Mini Classics "Super Mario Bros."
Vs. Super Mario Bros.
All Night Nippon Super Mario Bros.
Super Mario Bros. Special
Super Mario Brothers      For Nelsonic Game Watch  (Game watches/Wrist Watch Mario Game)
Super Mario Bros.            For YM-801 (Crystal Screen series/Mario LCD game)
Super Mario Bros.            For YM-105 (New Wide Screen series/Mario LCD game)



So except for Super Mario Bros. Deluxe, which should be counted as a different game imho, the NES/Famicom version should be the fastest unless the flagpole glitches on PAL are possible without left+right or there are some modifications on other ports that make the game faster.
What I didn't consider is, if some of these ports would allow left+right without hardware modification, but even then I don't think, it would save enough to counter the framerate difference, which in most cases would account for about a second in a warps run.
Versions you have missed

Super Mario Bros.            for the Family Computer Disk System

Super Mario Bros.            for GBA as a 20th Anniversary with a special edition

Super Mario Bros.            one of the trial games available in the "Masterpieces" section in Super Smash Bros. Brawl.

Super Mario Bros.            for Nintendo 3DS

Super Mario Bros.            for American arcade machine Nintendo PlayChoice-10.

Super Mario Bros.            for Wii U in Japan

Super Mario Bros.            for a Virtual Console remake with the question marks on the ? Blocks replaced with "25", exclusively bundled with a special, red Wii.

Super Mario Bros.            as Nintendo Mini Classic
Oh wow, Nintendo really can't get enough of this game.
Super Mario Bros.            for the Family Computer Disk System
That's in the list as the "Famicom version".
Super Mario Bros.            for GBA as a 20th Anniversary with a special edition
I read that on Wikipedia, but do you have any more info on this? What does special edition mean?
Super Mario Bros.            one of the trial games available in the "Masterpieces" section in Super Smash Bros. Brawl.
I didn't include it, because you can't beat it before the trial runs out.
Super Mario Bros.            for Nintendo 3DS
In the list as "3DS version".
Super Mario Bros.            as Nintendo Mini Classic
Isn't that one of the GBA games in the list?

I'll add the rest to it.
Caution: This user contains Kana ^_^
I just had to test that to be absolutely sure:
SNES All-Stars version is (slightly?) bugfixed. The one I noticed and remembered was the piranha plant in 8-4, first pipe Andrew goes down. Running into that plant in the SNES version kills you.
HELLO!
Did Super Mario Bros. FDS have any differences, other than possibly sprites?  SMB was originally released for Famicom, then re-released for FDS, unlike SMB2 which was FDS original.