Experienced Forum User, Published Author, Player
(208)
Joined: 7/7/2006
Posts: 798
Location: US
We have finally decided on an acceptable Gorsia fight.
What is left:
1: Make a small healing change before the Goron fight, execute as planned.
2: Manipulate RNG to Griffan
3: Heal before Griffan
4: Fight Griffan as planned
5: Talk to Saro
6: Manipulate dying to Black Knights in 1 turn
7: Buy 2 B Agility and 1 M Water
8: Buff RNG pretty intensively to Gorfun
9: Grab/Use P Seed (This saves about 10 seconds off the Gorsia fight.)
10: Execute Gorsia Fight as planned.
11: Submission Text
12: Maybe a Game Resources Page somewhere along the line.
Thanks to everyone for your support and patience.
Experienced Forum User, Published Author, Player
(208)
Joined: 7/7/2006
Posts: 798
Location: US
Nahoc wrote:
Here is (what was supposed to be) a complete test run for Tony Hawk's Pro Skater (N64):
http://www.youtube.com/watch?v=NqlzAX94kxE
Sadly, I forgot the "Pro Score Tape" in School Miami, but it's no big deal. I am already working on the final version and will submit when I'll be done.
Epic Fail indeed.
I feel like it might be faster to get more special meter in 'School' before getting to the secret tape. It should make you a fair bit faster right? Maybe you tested this and it was slower. What are the goals of the run?
Experienced Forum User, Published Author, Player
(208)
Joined: 7/7/2006
Posts: 798
Location: US
moozooh wrote:
Kirkq wrote:
I'm sure 85-95% of the viewers of SMB TASes played either the (U) or (J) version and not the (E) version.
Well, isn't this a good bit of statistics. :D I was under the impression that at least 1/3 of us here were Europeans, possibly indicating that around 1/3 of SMB watchers have played the European version first.
Experienced Forum User, Published Author, Player
(208)
Joined: 7/7/2006
Posts: 798
Location: US
Games where the goal is to acquire a password are acceptable uses of a password without having acquired it (NES: Gauntlet Legends)
Games where using a password increases entertainment or entertainment density are accepted on a case by case basis. You should read this list to get an idea of what is acceptable: http://tasvideos.org/Movies-C4020Y.html You are probably best off posting on the forums before submitting any password run.
Games where using the password is basically a free savestate to the end of the game are most often not acceptable uses of a password. Using a password to skip most of an already entertaining run is most often not acceptable. This goes for Metroid, Megaman, and anything else of this nature.
Experienced Forum User, Published Author, Player
(208)
Joined: 7/7/2006
Posts: 798
Location: US
sgrunt wrote:
Let me draw attention to another recently rejected run, submitted under similar circumstances where there's no substantial difference in game play between the NTSC runs (and there are several there as well) and the PAL run.
Does that submission warrant revival for publication as a separate branch?
I agree most with the arguments sgrunt has brought up in this thread.
The difference between 50 FPS and 60 FPS is different enough for each game that I'm under the impression that around half of the runs on the site could obsoleted by the PAL version. Calculations are rounded and incremented differently resulting in such things as differing movement speeds (resulting in slight modifications of glitches).
I'm sure 85-95% of the viewers of SMB TASes played either the (U) or (J) version and not the (E) version. It simply isn't the same game most of us purchased, so why feature an alternate version for no substantial reason other than slight calculation differences in the engine.
--------------------------------------------------------------------------------------------
I best see this as a strong argument for featuring optimized runs on the site.
Perhaps each game should have 'Featured Runs' and 'Other Runs'
SMB1
Featured: any%
Featured: No warps
Other: -1 ending (of course there would be arguments over whether this gets featured or not)
Other: Walkathon
Other: (E) any%
Other: (E) No warps
Other: (E) Walkathon
Other: Superplay
Unfortunately if this run is accepted (EDIT: primarily for being faster than the (U) version), it would set a precedent for obsoleting a very optimized (U)/(J) version with an unknown level of optimization on the (E) version.
I feel that this either has to be rejected or it needs inserted with a lighter categorization on the site. I think obsoleting the (U) version is a bad choice. If this gets accepted as a separate category I feel Saturn's run needs to go back to the workbench.
Perhaps more discussion is needed on new categorizations of runs.
Voting 'No' unless the site's classification of runs is changed in the future. I think the ultimate decision should make a statement on the choice of categorization instead of just throwing the run aside because it is a few frames suboptimal.
(Sorry this post was a bit unorganized, I wanted to post this quickly before the decision was made about the run.)
Experienced Forum User, Published Author, Player
(208)
Joined: 7/7/2006
Posts: 798
Location: US
This definitely proves that TASing is an art.
Awesome drawing skills. We're these done largely by 'hand', or was there any computer generation involved?
<_<
Definite yes vote as a highly artistic movie.
Experienced Forum User, Published Author, Player
(208)
Joined: 7/7/2006
Posts: 798
Location: US
LSK wrote:
The published version doesn't seem to have a final boss - one's mentioned in the submission text, but this one goes straight from the folder-arrangement for the final boss to the credits.
The submission text was intentionally ambiguous. (Why spoil the ending?) I assume the published version is correct.
I have made a video containing this 'alternate' ending to the run. It also contains the dialog originally skipped before the ending credits roll.
http://www.youtube.com/watch?v=dzsQJEP2398
Experienced Forum User, Published Author, Player
(208)
Joined: 7/7/2006
Posts: 798
Location: US
Short version of the submission text (Bag broke the record for longest submission text by quite a bit.):
- Need to collect 60 chip types in the library before a point near the end of the game.
- Need to get Megaman to level 30 (HPMemory, PowerUP, Armor) before a point near the end of the game.
- Entering a battle on a certain RNG will determine which chip folder placement (1-30) will be assigned to the 5 chips in round 1 of the battle. Since the folder positions are locked, one can rearrange to folder such that the chips wanted are in the correct folder slots to be assigned in the first round of battle.
- There are a lot of people mandatory to talk to to advance the storyline.
My comments:
My contribution to the run was pretty minimal. I figured out the method of folder stacking by poking around memory, which helped to simplify the run to measurable time gains. I cannot imagine how suboptimal and painful this run would have been had this not been found. I also prodded Bag a lot because he procrastinates about as badly as I do.
Even though this run consists of a lot of walking, a big portion of the run was the route planning. Bag timed out pretty much every possible chip type that could give a quick collection. One can weigh (1 new chip) = (3 repeat chips usable in the trader) + (time) = X Zenny + (time). Once all this had been weighed, Bag just had to execute his plan.
This run may provide more than most of the other MMBN games. The reason behind this is that the chip collection and level requirement allow some unique boss strategies to be used by choosing to spend time getting suitable chips available. The other games may not be able to effectively spend time getting good chips and powerUPs to beat the bosses faster.
I would go so far as to say that this run cannot be beaten by more than 30 seconds under the current general route and tricks. Luck manipulation can always be slightly better. The only thing in the run I would even consider visually suboptimal is the random battle for Dash G. Too many uncontrollable random factors made this a couple seconds slower than it could theoretically be. Although I'm sure no one would've known the difference if we hadn't pointed it out. =p
I would like to personally thank Bag for running a favorite game of mine, and performing the planning and execution better than I could have.
Experienced Forum User, Published Author, Player
(208)
Joined: 7/7/2006
Posts: 798
Location: US
This is somewhat old, and known throughout the smash community. =p
It is also possible to get a character to have the wrong "stock icons" during a match.
Experienced Forum User, Published Author, Player
(208)
Joined: 7/7/2006
Posts: 798
Location: US
Critical five wrote:
I'm actually quite interested in [seeing] a Golden Sun run (1 or 2 whatever). Is there anybody still working on the route for GS1 ?
I'm very interested in running Golden Sun 1 after I finish my current projects. Currently I have found a sequence break to the end of most dungeons I have encountered. It does not seem possible to sequence break the forest before Tret Tree, but there is probably a small sequence break up Tret Tree.
When I intend to TAS this, I will do a preliminary (suboptimal) run to test things like how many sacred feathers/ how much healing will be necessary. The only fact that needs to be strictly considered is the RNG incrementation.
By the way, walking straight in two directions covers exactly the same distance as walking diagonally. The only consideration that needs made is how to best approach corners to make sure not to lose ~2 frames.
I can only say I will probably start within the next year. If people want to cooperate and get started now, I'll discuss all of my speculations with you. A lot of what I've done has been directly addressed in this thread.
Experienced Forum User, Published Author, Player
(208)
Joined: 7/7/2006
Posts: 798
Location: US
To find the RNG one must either disassemble the game (hard), or gain some understanding of at what point the RNG changes (easier). Once you understand WHEN the RNG changes you can apply 'Equal' and 'Not Equal' to get a short list of addresses that may be the RNG.
(Using disassembly, one would only need to know what sort of action would cause the relevant address to be accessed. Many things like damage calculation can only be found through the game code, and these formulas directly address the RNG.)
See this page if you haven't already: http://tasvideos.org/MemorySearch.html It should answer most of your questions.
Also, please don't create new threads for games that have existing threads. Either post in the Newbie forum or post in the relevant game thread.
Also again, if you aren't aware, everything you do here is reported to the IRC channel. =p
Experienced Forum User, Published Author, Player
(208)
Joined: 7/7/2006
Posts: 798
Location: US
It would be helpful if you could better explain what you intend to analyze or alter. Luck manipulation is a really broad scope. If you want to find the RNG, you want to first understand what actions change the RNG and then try to use memory watch to extract the address. I suppose the Memory Watch page could use a section on how certain common RNGs work.
If you are trying to see what the boss will be doing or if attacks will connect based on the RNG, you will need to disassemble the game via a debugger.
Using a debugger to read 6502 code is beyond the ability of 90%+ of the users at this site.
I can detail more if you ask more specific questions.
Experienced Forum User, Published Author, Player
(208)
Joined: 7/7/2006
Posts: 798
Location: US
Mitjitsu wrote:
I find the pace of the movie to slow compared to Arkanoid, plus it doesn't complete the game 'No' vote.
It seems fairly arbitrary to only play 4 levels, and as a game it isn't a strong contender anyway.
Input animation and mild pause manipulation are the only things this game really has going for it. Why not improve the current Arkanoid run and add input animation at the same time?
I'm sure some people will appreciate this run, but as just another Breakout game, it doesn't really have much to offer.
Experienced Forum User, Published Author, Player
(208)
Joined: 7/7/2006
Posts: 798
Location: US
I did surprisingly well on finals this semester. I'm doing research under a professor this summer 40 hours a week, but I still am finding time to work on things.
"We have not forgotten." - I'm not entirely sure how to take this, but it probably is helpful to hound me once in a while, for sure.
Goron is nearing death. I have a successful kill that is usable on a reasonable RNG. I need to get to Goron on that RNG, and check a few different variations on the second half of the battle that may improve it by a turn or save a potion or something.
Goron is half dead, then Griffan, then Gorsia. This run is approaching completion.
Experienced Forum User, Published Author, Player
(208)
Joined: 7/7/2006
Posts: 798
Location: US
Since these haven't been answered yet I'll briefly address them.
1: "Open Source":From what I understand only about half the stars are hex-friendly. I believe the two main factors that cause randomness are enemies and coins. The ~60 or so stars that are hex friendly tend to be the shortest. The longer stars tend to have enemy randomness or happen to be the 100 coin stars. Roughly 1/3rd of the run can be done by hexing, and they are trying to keep track of these .m64 files at www.sm64.org (though they are currently redoing their official list.)
2: BitDW: I believe that run you posted is something like 1-2 frames faster than the previous best BitDW run. It is not a substantial improvement. 1-2 frames can be gained just by turning at 74 degrees instead of 75 or something that a viewer would not notice. PBLJ is not useful in BitDW because there is a limitation on speed around one of the bends or Mario cannot turn sharply enough to stay on the course. PBLJ is only useful for gaining a faster speed than a regular BLJ could have, which is not useful here.
Dead link, reposted here: http://www.mediafire.com/?wzzrmtnztk2
(I assume it was uploaded after the last backup before microstorage crashed or something.)
Experienced Forum User, Published Author, Player
(208)
Joined: 7/7/2006
Posts: 798
Location: US
Highlights: Slime Climb, Rambi Rumble, Castle Crush, Toxic Tower.
Extremely polished. Good use of new tricks and routes to produce a fresh new run. Nice entertainment choices as well. Obvious yes vote.
I still feel that visual improvements are much more impressive and worthwhile than saving 2 frames around the next corner. Regardless, I feel sorry for the next guy that tries to beat you by 2 frames. Spend your time where you want obviously, I'll still watch your runs. =p
