Posts for petrie911


Experienced Forum User
Joined: 7/16/2006
Posts: 635
BTW, for those of you who want to watch this run as it would be timed by the ingame timer, use this lua script
a=0 

while true do 

   if memory.readbyte(0x7e198e) == 128 then 
   gui.drawbox(0,0,255,31,"#000000") 
end 

b=memory.readbyte(0x7E1842) 
if a~=b then 
snes9x.speedmode("normal") 
else snes9x.speedmode("turbo") 
end 
a=b 

   snes9x.frameadvance() 
end a=0 

while true do 

   if memory.readbyte(0x7e198e) == 128 then 
   gui.drawbox(0,0,255,31,"#000000") 
end 

b=memory.readbyte(0x7E1842) 
if a~=b then 
snes9x.speedmode("normal") 
else snes9x.speedmode("turbo") 
end 
a=b 

   snes9x.frameadvance() 
end 
This script automatically fast forwards during any part of the run where the ingame timer is not running. I didn't make this script, BTW. I don't remember who did, though.
Experienced Forum User
Joined: 7/16/2006
Posts: 635
If memory serves, the Murder Beam generates tons of lag, so it probably wouldn't do any good.
Yes, but IIRC the lag doesn't count on the in-game timer. I could be very wrong about that, though.
Onyx3173 wrote:
Also a question, how do you get the super missile? I thought you had to mockball there to get past the gates?
http://tasvideos.org/GameResources/SNES/SuperMetroid.html#ShoulderPumping
Experienced Forum User
Joined: 7/16/2006
Posts: 635
Watched it. Voted yes. A high quality run. I do have a question, though. How much time do the High Jump Boots and Plasma Beam save? Would grabbing the Spazer for the Murder Beam save time?
Experienced Forum User
Joined: 7/16/2006
Posts: 635
twilightsamus wrote:
Okay, I went back to Taco & Kriole's run that's on the website just to make sure I wasn't missing something, and I apparently I am, because their run says "38:41.52" while this one is "39:15.3" but this one completes the game faster? How are these times listed in the subject lines calculated, exactly?
The subject line lists the length of the movie in real time. However, this movie improves Taco and Kriole's run as measured by the in-game timer. Because the in-game timer does not run during door transitions and cutscenes, the resulting movie is longer in real time.
Experienced Forum User
Joined: 7/16/2006
Posts: 635
Oh, good, it's this debate again. We certainly haven't done this to death before. Question: does any other game get this much crap about J vs U version?
Experienced Forum User
Joined: 7/16/2006
Posts: 635
OK, I can finally settle the question of whether to get the Ice Spreader. The answer is no, because I managed to find a way to do the Prime fight even faster. It's possible to hit Prime while he's charging at you. It's an extremely difficult shot, but it allows you to get two full charge shots and an uncharged shot on him before he changes color. The Ice Spreader does 225 damage to Prime, while 2 full charges and a normal shot do 210. Prime's phases have 220 HP each, and the extra 10 damage can easily be provided by the Wave Beam. This, combined with the freeze time of the Ice Spreader, should make the Prime fight go faster without it.
Experienced Forum User
Joined: 7/16/2006
Posts: 635
So, decided to investigate some things about jumping. Samus has a cap on her vertical velocity of 11.667/s. This cap is raised to 14.000/s when you acquire the Space Jump Boots. This is why Samus's first jump is higher with Space Jump Boots, as the last frame of upward thrust from her jump hits the lower cap. This is also why unmorphs go higher when you don't have Space Jump. Her acceleration due to gravity is -35/s/s. The Morph Ball has no cap on its vertical velocity. A bomb imparts an upward velocity of 23.516/s when it explodes. This is entirely independent of the Morph Ball's previous velocity. The acceleration due to gravity for the Morph Ball is 85/s/s when traveling upward and 67/s/s when traveling downward. This could also be thought of as an acceleration due to gravity of 76/s/s and an air resistance of 9/s/s. On the frame you unmorph, you lose about .1/s of your vertical velocity. This is somewhat inconsistent.
Experienced Forum User
Joined: 7/16/2006
Posts: 635
Great run. Neat levels, a slick design aesthetic, and of course great TASing. Loved the level at 4:15.
Experienced Forum User
Joined: 7/16/2006
Posts: 635
So, here's something fun 80C5D724 -- Forward Velocity 80C5D728 -- Upward Velocity This applies for all characters, from what I can tell.
Experienced Forum User
Joined: 7/16/2006
Posts: 635
And I've found the ingame timer DF33E0 -- In-game Timer DF33E4 -- In-game Timer Why 2 addresses? Because the in-game timer is stored as a double. This makes absolutely no sense, as a 32-bt integer would certainly have sufficed. The value of the timer is stored in seconds, and it tracks fractions of a second. The in-game timer seems to move around a bit in RAM. However, it's always 0x248 addresses before the Missile count, which is very easy to search for.
Experienced Forum User
Joined: 7/16/2006
Posts: 635
More addresses 80457D50 - Charge counter 80457D68 - Bomb refill timer (the individual bomb timers are after this) 80DF3728 - Missile count (Your inventory is all around this location) 8100B0CC - Omega Pirate health The charge counter starts at 0 when you fire your beam and counts up by 1/108 per frame. The damage of the charged shot is equal to the charge counter multiplied by the maximum damage of the charged beam. 62 frames is the earliest you can fire, yielding a shot with 57% of full damage. Omega Pirate's health might not always be in the same location. It's not too hard to find, though, as it always starts at 600.0. Anyways, having found the charge counter and at least one enemy's health, I can now update the damage table to be completely accurate. The flamethrower has an oddly inconsistent damage output. I'll update this if I ever figure it out, but it seems low priority. The estimates from before should still be fairly accurate. Also, while I haven't confirmed this, these values suggest that hard mode halves your damage output rather than doubling enemy HP. I'll see what I can do about finding the ingame timer. That's probably going to be annoying though, as like just about everything else it's probably stored as a float.
Experienced Forum User
Joined: 7/16/2006
Posts: 635
So, apparently Dolphin *does* have RAM watch, but it's only accessible in debug mode. Anyways, through some effort combining cheat search with RAM watch, I can now bring you the following. 8046B9BC - X position 8046B9CC - Y position 8046B9DC - Z position 8046BAB4 - X velocity 8046BAB8 - Y velocity 8046BABC - Z velocity The X and Y position addresses may not be the true addresses, as they don't seem to respond to changing their values like the others do. However, they should still be usable. Some other things --Samus is about 2 units tall --Speeds are measured in units/second --The acceleration due to gravity is -35 units/second^2 --There is no terminal velocity --When falling, the Morph Ball bounces off the ground with 1/5 of its impace velocity.
Experienced Forum User
Joined: 7/16/2006
Posts: 635
I got that second picture. I went with "Ugly". Possibly not the most accurate, but it was the first thing that came to mind.
Post subject: Re: Pokemon XD: Gale of Darkness
Experienced Forum User
Joined: 7/16/2006
Posts: 635
Would similar things apply for Pokemon Colosseum?
Toothache wrote:
What is this mode and how do you access it?
Experienced Forum User
Joined: 7/16/2006
Posts: 635
Awesome. Incidentally when Marcus hit level 20, I was wondering why you didn't go for any more stat points. I hadn't realized Paladin caps were that bad (seriously, 24 speed?). So, what's next for this game? HHM TAS, or HHM S-rank TAS?
Experienced Forum User
Joined: 7/16/2006
Posts: 635
Stop making those screenshots jpgs. They need to be pngs. If Mupen automatically saves as jpgs, then take the screenshots manually with PrintScreen. Also, as long as you're at it, turn up the antialiasing. Makes the game look real nice.
Experienced Forum User
Joined: 7/16/2006
Posts: 635
I'd like to see a fastest time TAS for Classic mode. Actually, I'd like to see fastest time for all modes.
Experienced Forum User
Joined: 7/16/2006
Posts: 635
There's also the problem that the scan dash isn't nearly as controllable. You can bend it a bit, but otherwise you're stuck in a straight line. Also, doesn't Geocore take significantly longer without boost since you have to reload the room?
Experienced Forum User
Joined: 7/16/2006
Posts: 635
I've been waiting for this to finish for some time. I'll watch the encode before voting, but I think the vote will be yes. Also, fun fact: on average, a Pokemon is caught every 80 seconds.
Experienced Forum User
Joined: 7/16/2006
Posts: 635
Of the ones Grunz posted, I like snapshot 151, though it might be a bit of a spoiler for one of the run's big new tricks. More concerning is the fact that the screenshots seem to be jpegs, and also don't have any anti-aliasing. Anyways, yes vote, and I'd say this should definitely get a star. We don't have any 3D runs starred, and this seems like a very good choice.
Experienced Forum User
Joined: 7/16/2006
Posts: 635
arflech wrote:
Another one turned out to be more fruitful, just raising that entire rational function to the power of n; then the envelopes of the normalized family turn out to reach their greatest extent in the limit as n->infinity, when the envelopes become exactly the same as for that Gaussian family, and the normalized family based on 1/(1+x2)n becomes the normalized Gaussian family based on e-x^2; I'm thinking about releasing a .nb file that can be viewed in MathPlayer to show this sort of thing, because it is truly astounding. Maybe the normalized Gaussian family is the most "open" possible family of functions converging to a delta distribution, in the sense of having envelopes with the greatest distance from the origin.
The fact that those functions approach gaussians as n->infinity can be readily seen from the fact that (1+x2/n)-n -> exp(-x^2) as n-> Infinity. This result isn't just true of 1/(1+x^2); it actually works for any function that is twice differentiable at the origin. This fact is the reason behind the Central Limit Theorem.
Experienced Forum User
Joined: 7/16/2006
Posts: 635
Given three non-collinear points A, B, and C, and two other points X and Y not coplanar with A, B, and C, find a set of conditions to determine whether the line segment XY passes through triangle ABC.
Experienced Forum User
Joined: 7/16/2006
Posts: 635
Spectacular run. Obvious yes vote. One question, though. Is it faster to use the female character over the male one? It seems to me like the time differences favor the male protagonist. Also, I must say there's some interesting things going on in the suggested screenshots. First, we see that everyone hates our main character... It seems Steven is in to S&M... Who knew that Archie was British? And, in all honesty, I think we have the TAS summed up in one picture here...
Experienced Forum User
Joined: 7/16/2006
Posts: 635
Any particular reason for the switch? Or is it the text thing again?
Experienced Forum User
Joined: 7/16/2006
Posts: 635
Does the Japanese text really make up all 8 min of improvement? I thought the time difference was less.