Well, when this gets submitted, Definite Yes vote from me. I just like how you completely demolished the 2 auto scrolling levels, every brick destroyed, every enemy killed, every coin and power up collected.
just one little question: a few times in the autoscrolling levels you collect a mushroom and then let yourself be hurt to lose it again. are you sure that this doesn't cost time? because in Super Mario World every transformation did cost time, even in autoscrolling levels
apart from that, very nice :)
Joined: 7/17/2004
Posts: 985
Location: The FLOATING CASTLE
just one little question: a few times in the autoscrolling levels you collect a mushroom and then let yourself be hurt to lose it again. are you sure that this doesn't cost time? because in Super Mario World every transformation did cost time, even in autoscrolling levels
This game seems to have a completely different engine than other Mario games. Other games stop to show the powerup/powerdown animation, which loses time even when autoscrolling, but in this one Mario just keeps on going, transforming as he moves. In fact, look closely at 3-1, after I pick up the mushroom--I go under one enemy who should have hit big Mario, but because the transformation isn't complete nothing happens. I guess it takes time for shrooms to take effect after eating them.
In general, small Mario has much more to take advantage of in this game. He can hit any block without stopping while big Mario can only hit the last block of a row of blocks. In the shooters, small Mario's collision box lets me do more refined movements around enemies, like going between the octopus fireballs. It's also necessary to embed myself in the bricks further so I can sculpt those shapes. Finally, small Mario can move past a head-blocking obstacle faster. For example, at the end of 4-2, small Mario can run through that gap before big Mario. As small Mario, it's possible to finish 2-3 a few frames faster, but as big Mario I'm still able to hit the finish frame for the optimal bonus round, and the bonus animation is faster.
This game seems to have a completely different engine than other Mario games. Other games stop to show the powerup/powerdown animation, which loses time even when autoscrolling, but in this one Mario just keeps on going, transforming as he moves.
Super Mario Land 2 is like that too, isn't it? It reminds me of how Rockman World 2's engine seems very different from all the other Rockman Worlds, and it lets you pick up refills without stopping the game.
And now I think those other files I got from IRC just got a little corrupted at the end because I downloaded them in a nonstandard way, not from any fault of yours.
put yourself in my rocketpack if that poochie is one outrageous dude
I just finished watching your movie.
You didn't use a bug that I like.
It doesn't necessarily save time, but it
looks neat.
In Super Mario Land, in the autoscroll
areas (2-3 and 4-3), usually when
you are pushed off the edge of
the screen, you die. However,
if you do it in the right place
and hold up (or down), sometimes you
wrap around to ahead of the screen.
You might be able to abuse this to push the
button at the boss in 2-3 much sooner.
Also, in the long path in 4-3, I like to
use it to get into the outside area that
is supposed to be unreachable.
Then, you can use it again to get back inside.
It is more fun than just following the path.
I noticed that in 3-1 and 4-1 when you
had firepower, you didn't shoot at
all of the enemies that you could have
(without slowing down).
Is there a reason for that?
If your goal was points, that should
have gotten you a few points extra.
And Nach thinks you should have hit another invisible heart block in 3-1!
scubed wrote:
I noticed that in 3-1 and 4-1 when you
had firepower, you didn't shoot at
all of the enemies that you could have
(without slowing down).
Is there a reason for that?
If your goal was points, that should
have gotten you a few points extra.
hopper wrote:
You choose the the Fireflower in the bonus stage at the end of World 2, but in 3-1 you don't shoot some of the Pirahna Plants, even though you stated that your secondary objective was getting the highest possible score. It also looked like you didn't shoot everyone that you could have in 4-2. Would that have slowed you down or negatively impacted the run, or would the shot have gone through the Pirahna Plants because they weren't totally out of the pipe yet? This definitely gets a Yes vote, but I'm sure people will wonder.
Bag of Magic Food wrote:
Yeah, since TheAxeMan's holding B all the time to run, he has to release B to throw a fireball for it to register as a new button press, and that lowers his speed a tiny bit.
TheAxeMan wrote:
BOMF is right about the fireballs, it seems to cost a frame each time. Near the end of 4-2 I shoot the last snake because his fireball would have cause more than a frame of lag. The snake I shoot with the bouncing shot was just in my way.
put yourself in my rocketpack if that poochie is one outrageous dude
Sorry, I missed that post.
Does shooting a fireball always costs a frame?
Does this include while jumping too?
(Since momentum is often different
in games when jumping.)
Joined: 7/17/2004
Posts: 985
Location: The FLOATING CASTLE
In Super Mario Land, in the autoscroll
areas (2-3 and 4-3), usually when
you are pushed off the edge of
the screen, you die. However,
if you do it in the right place
and hold up (or down), sometimes you
wrap around to ahead of the screen.
I didn't know about this. But I do let myself get pushed back until the last possible frame, otherwise I would die. Perhaps they fixed this glitch with the 1.1 rom. That would be a neat trick though, but I doubt it would help finish 2-3 any faster since there's nothing to push me back for at least a whole screen before the boss.
You seem to be correct.
I tried it on
Super Mario Land (V1.1) (JUA) [h1]
and it wouldn't work.
I don't know what version I have,
but here is its SHA1:
3a4ddb39b234a67ffb361ee7abc3d23e0a8b1c89 sml1.gb
It probably is V1.0.
My point is that it did not work
on a V1.1 ROM, but it did work
on the version that I have,
whichever version it is.
h1 just happens to be the
first V1.1 ROM I found.
It is not the one I typically use.
I was agreeing with what TheAxeMan
said about ROM versions.
I recently started a TAS of this game for fun but instead of aiming for the highest possible score, I aimed for the lowest while maintaining maximum speed. At the start of level 3-1 I am 14 frames ahead of the current TAS. Here is the .vbm of the run so far.
I recently started a TAS of this game for fun but instead of aiming for the highest possible score, I aimed for the lowest while maintaining maximum speed. At the start of level 3-1 I am 14 frames ahead of the current TAS. Here is the .vbm of the run so far.
This just makes it boring, honestly. Good speed, but not getting anything... :/
I recently started a TAS of this game for fun but instead of aiming for the highest possible score, I aimed for the lowest while maintaining maximum speed. At the start of level 3-1 I am 14 frames ahead of the current TAS. Here is the .vbm of the run so far.
This just makes it boring, honestly. Good speed, but not getting anything... :/
The run on the whole is less interesting but I think in the levels where there are many points to be avoided are more interesting in particular levels 1-2 and 2-1. Perhaps a tertiary goal of making as many possible jumps while retaining the previous goals could make it more interesting...
Previously I didn't think there were any mistakes in the current TAS but it seems as though 10 or so frames were lost on the bonus level of level 2-3. Not that I have any plans of besting it.
Joined: 7/17/2004
Posts: 985
Location: The FLOATING CASTLE
Ok, I've looked into this a little further and determined that the difference comes in the transition between worlds. In fact, xenos is actually a few frames slower than the published run as far as the level-playing. Perhaps some lag or movement tweaks, I didn't figure out exactly where in the levels it happened.
Anyway, at the end of the first world he is 5 frames behind. But at the start of world 2 he is miraculously 8 frames ahead. He loses some of these but still finishes world 2 three frames ahead of me. It also appears that he doesn't need to wait as long to get a good bonus result at the end of world 2.
Now due to the same transition effect, he jumps ahead further. So overall, he gained 13 frames from the first world transition and 11 from the second, plus 4 from better luck. However, he loses 14 frames in the levels and the overall result is 14 ahead.
So to really capitalize on this we need someone to check the disassembly to figure out what is happening at those level transitions and maybe the bonus manipution too. Perhaps smaller score = faster transitions?
I started another TAS of this game except with maximum possible jumping but still with the lowest score. I was one frame ahead of the current TAS but in order to get luck manipulation, I needed to lose that frame (the frame was gained at the very beginning by pressing the start button earlier). Anyway, the transition took longer than usual and I started 2-1 34 frames behind the current TAS even though I had the same score as my former run. So it probably doesn't have anything to do with score. Here is the vbm of that run.
Joined: 11/18/2006
Posts: 2426
Location: Back where I belong
I have been looking into improving this run, but am completely stuck. The world ending fade out makes very little sense to me, although I have discerned that there is a 46 (!) frame rule in effect. I save one frame throughout the entire run, due to pressing start one frame earlier. This 1 frame savings lasts until I complete level 1-3, where I suddenly gain another 11 frames due to a faster fadeout / faster "princess to fly" flying right animation. HOWEVER, this causes an absolutely horrible setup for the bonus game (ideal is x1 life w/o using the ladder). If I delay one frame to match the prior movies time, it somehow causes me to lose 35 frames (46 frame rule -11 frames saved).
This does not seem to be affected by score, nor amount of coins collected (as I have altered both vairables with the exact same results). It seems to be affected by subpixel location of mario when triggering the level end sequence, or possibly frames of slowdown (ie not holding B). I think I have figured out that the fadeout to the bonus screen is affected by the flies movement patterns, but have no way to predict when a favorable one will appear (I have also been able to delay 6 or 7 frames and achieve the same result as xenos's WIP above, but still with horrible bonus level set up.)
So, in short, I'm very confused, and I'm not quite sure where to go from here. Does anyone with more programming experience have any suggestions as to how I should approach this? Or a RAM address that may be helpful (I've tried looking with no success). Despite this being a GB thread, I'm hopeful I'll get some sort of response before the thread dies.
Here's my version with 12 total frames saved, but terrible bonus screen:
http://dehacked.2y.net/microstorage.php/info/1879184794/Super%20Mario%20Land%20%28W%29%20%28V1.1%29%20%5B%21%5D.vbm
Here's xenos's low score attempt with 4 frames lost due to various causes, but 8 total frame saved with ideal bonus screen:
http://dehacked.2y.net/microstorage.php/info/1126605424/SMLTASlowscore.vbm
