Experienced Forum User, Published Author, Player
(209)
Joined: 2/18/2005
Posts: 1451
I see, welcome in the busy world.
Hopefully you can do at least a bit of progress this year though, so that you may realize that
moozooh wrote:
And no, pre-Torizo segment won't be copied in any case because I'm sure as hell that further improvements are possible. Besides, I promised to improve T&K's WIP in some way, and I am definitely going to do so.
won't happen. One can promise much, but sometimes we just have to face the facts. Everything that can be optimized already is optimized in the pre-Torizo (that is, till the exit of the mainstreet room) segment. Good luck to at least try though, just don't expect too much of it.
Experienced Forum User, Published Author, Player
(209)
Joined: 2/18/2005
Posts: 1451
Hero:
I'm currently working on SM-Redesign again, to finally finish the long awaited WIP5 for it. After it's released I will probably pick SM any% up again and work on both runs alternatively from that time on. As of now I didn't touch SM since my last public TAS demo in January, so I still can't tell how much time is truly possible to save over your, already very optimized run, and whether a 0:23 is really reachable with the realtime route. I still have some doubts about it even now.
evilchen:
I actually thought moozooh works at least on his low% v2, now that everything until Torizo just needs to be copied. There is no reason to wait for anything else.
Experienced Forum User, Published Author, Player
(209)
Joined: 2/18/2005
Posts: 1451
Damn, I at first thought that only the very first lap of my TAS would be somewhat improveable due to the better bump of the other car at the start. Seeing you managed a almost 0,1 sec improvement even in a normal 1-boost lap is just incredible!
I took a closer look at your smv, inichi, and it seems that the key for this unexplainable high improvement in the 2nd lap is due to a new boost trick in which you shortly release the gas at a few moments during the boost, which increases your speed for some unknown reason. Also the script to monitor all pixel values and directions seems to ease the corner-optimization work alot. I didn't have this benefit, and thus had to brute force every of them, which gets clearly noticable in the rerecord count.
As to the new boost location, it seems to indeed be slightly faster in the end. Although I didn't test it myself, it gets pretty clear since you lose around 30 km/h after the last corner, which costs a few seconds to regain again due to the slow acceleration of the Fire Stingray. The boost eliminates this problem immediately, where the shortcut through the sand is passable without any speed loss anyway, thanks to the tiny sand-free line along the wall there.
Other than that, good luck with further progress, bagger.
Experienced Forum User, Published Author, Player
(209)
Joined: 2/18/2005
Posts: 1451
Good - better - inichi!
A sub 3 hour before Lavos is just a unbelievable time, and the most shocking thing is, you managed to beat Lavos even without a GreenDream and Status protection (which I thought to be pretty much impossible on such a low level, even in TAS conditions)! 0_0
The new boss strategies are crazy as well. Seems like the Fire Whirl is extremely useful in a speedrun like this, and so it's the more surprising that nobody considered to use this relatively powerful tech for it's early stage before.
Even as a testrun, this looks flawless and very well optimized already. It will be very interesting to see how many seconds (or even minutes) you will still manage to improve over this, already great run. Good luck for the final version!
Btw, do you think this new Lavos strategy would be possible on console as well, or is the timing between your hits too critical in this case? Skipping Fiona's quest would allow for a pretty big improvement over your current SDA run after all.
Experienced Forum User, Published Author, Player
(209)
Joined: 2/18/2005
Posts: 1451
P.JBoy wrote:
...I honestly thought that that was completly obvious, I have done that in almost EVERY playthrough I've ever done. That is a really, really, really old trick. 8-years old infact. Are you telling me that NONE of you thought of that?
The glitch itself was known by some people back then (I accidentally found it myself in around 1997 as mentioned before), but surprisingly nobody ever got the idea to abuse it in a speedrun so that it saves much more time than just doing it the usual way. What I basicly showed is a new speed trick never done before, not just the glitch itself.
Experienced Forum User, Published Author, Player
(209)
Joined: 2/18/2005
Posts: 1451
Cpadolf wrote:
Well we had a way to bypass the screwattack de-equip without using the shinespark method... I guess we won't get to use that now as it is probably slightly slower ingame than not having it equipped :(
Very interesting. I didn't thought it's possible with it. But even with a additional pause screen, it's still much faster in both time categories in the end.
Experienced Forum User, Published Author, Player
(209)
Joined: 2/18/2005
Posts: 1451
Cpadolf wrote:
Also how can it save more realtime than ingame time if you need a pause screen?
Check the smv. You don't need a additional pause screen. Simply de-equip both suits right when you have to remove the Screw Attack for the Zebetite glitch anyway. The additional 22 realtime frames are probably saved due to better lag management.
dave_dfwm wrote:
I managed to do that trick on my actual console when I was really young -- however, I got "stuck", and the metroid never came off my head. Mother Brain never died, and I had to reset.
I too accidentally found that trick on my console during the 90's when fooling around. Though to that time I never thought it's possible to get MB down to 0 HP before even getting the Hyper, so I didn't test it further back then. In order to get stuck, you just have to make the Metroid getting trapped in a wall (which happens when you jump around in the wrong times during the phase when the Metroid tries to grab you, I think) so that it can never reach you. That way you will never be able to trigger MB's death no matter how many times you shoot at her because the game won't continue the fight script until the Metroid is dead.
Experienced Forum User, Published Author, Player
(209)
Joined: 2/18/2005
Posts: 1451
Although I planned this to be a surprise for my RBO run, which I would eventually finish someday, I decided to make a exception and show you the mentioned MB trick anyway:
Mother Brain Stand Up glitch
YouTube Video: http://www.youtube.com/watch?v=ol7c5WOFLIg
First time performed in one of my testruns in mid 2006, this trick obsoletes the Murder Beam in a 100% or a RBO run even when they aim for fastest in-game time. Compared to JXQ's 100% v2 run, the above demonstration saves 810 in-game frames (13,5 sec), or even 832 realtime frames, making a 0:36 100% a easy goal with it now. Go for it, guys. :-)
EDIT:
Added link to YouTube Video
Experienced Forum User, Published Author, Player
(209)
Joined: 2/18/2005
Posts: 1451
Spider-Waffle wrote:
Couldn't you use that MB zap trick do save a little time in a 100% run by doing some damage before getting hyper beam?
You can. In fact, it's the way to go in a 100% or a RBO run, because it's faster than using the Murder Beam, in both, real- and even ingame time. A obvious trick to use in any run if you have the required conditions for it, and I'm a bit surprised nobody ever got the idea for it before.
Experienced Forum User, Published Author, Player
(209)
Joined: 2/18/2005
Posts: 1451
Well, for only 86 rerecords and the slower car choice, the 1st version looks pretty good already. But yes, the purple "Fire Stingray" car is easily the best choice in every single course without exceptions. It has the greatest top speed and on top of that a much better corner handling (it doesn't slip out of line in them, even at a very high speed, saving you many frames in every corner in the end. Your 2nd movie was a huge improvement already, so just practice a bit more with the game and the tools, and I'm sure you will get close to perfect times with enough patience. Good luck!
Also, check this old F-Zero thread to get some more information:
http://tasvideos.org/forum/viewtopic.php?t=1518
Experienced Forum User, Published Author, Player
(209)
Joined: 2/18/2005
Posts: 1451
Bisqwit wrote:
Saturn: Well, you have to understand that it pisses off people to know that you've been sitting on this information all the time while not producing anything. When you suddenly trump yourself forth when others discover the trick, stealing their thunder, it only creates aggravation.
Bisqwit, please understand that I don't "steal" anybody's thunder. Taco got full credit in my vid to discover that trick as well, with a extra mentioning that he did it independently of me. Everything is as fair as it can be here.
It's really unfortunate that things turned out like that recently, and that a few people might be pissed here. You can be sure that it certainly wasn't my intention to happen.
As to the question of who I plan to surprise, everybody who knows Super Metroid well enough.
Thinking about everything you guys said, and the in my opinion not justified threats towards me, I decided to follow one of your suggestions and avoid commenting on new tricks in the future here (at least as long as I don't contribute a new trick myself), to not annoy you anymore. I will just quietly work on my run further as much as the time allows, and hopefully present you a finished one in the not too far future (if things go well maybe even this year already), without mentioning too many details of it until it's done, like the majority of TASer do here as well. Is this ok with you Bisqwit?
mmbossman:
I'm not hindering anything. Every person is free to TAS SM anytime, and nobody should be dependent on somebody else's run.
evilchen wrote:
i see thanks for answering saturn its sad you arent motivated very much for your tases but who is it :P?
Yeah, I know. Sometimes it's hard to be motivated, especially on a difficult game like this one. There is nothing we can do about it.
Tonski wrote:
Saturn didn't you already quit the RBO and any% projects? That leaves you only with SM:R TAS that I doubt you'll finish, you always seem to make excellent fast progress in the beginning of all your runs, but halfway or earlier you just stop.
No, I just cancelled the old ingame-oriented any%, but I will likely do a realtime one as the next run. And regarding my old RBO, even though it's obsoleted in the beginning by now, I still play with the thought to finish it just for fun, to see what time is possible to get with the route I used there. And no, I won't quit SMR as well. It's still top priority (along with the SM-any%) as of now.
Experienced Forum User, Published Author, Player
(209)
Joined: 2/18/2005
Posts: 1451
I mentioned it many times before, I just want to make a run that shows new stuff for the watcher to increase the surprise. Thats why I avoid spoiling tricks before finishing it. Unfortunately my freetime is very limited and my motivation to work on the run not as high as before. I know this way is not the best one, but you must understand that it would be pointless for me to work on the run further if there would be no new stuff to show off. I try to avoid giving up on my run (to show you all something new in the end), so I just have no other choice as you see.
How I feel about seeing a trick spoiled here? On one side happy that we have such smart guys like Taco here, on the other side of course a bit sad that things don't work out like I planned them to do (to show the trick the first time in the finished run).
Why I make videos of tricks? To at least prove everybody that I used them in my previous runs.
I hope you can understand my situation at least a bit.
Experienced Forum User, Published Author, Player
(209)
Joined: 2/18/2005
Posts: 1451
As moozooh said, F-Zero is a real pain to optimize. I would really like to see a full GP TAS of this game, so if you feel like going through all that, dartht33bagger, then go for it!
Experienced Forum User, Published Author, Player
(209)
Joined: 2/18/2005
Posts: 1451
Congratulations, Taco!
I knew it can only be you who will find this on his own, good job man. :-)
The funny thing is, I uploaded a demonstration video of this trick (the fully optimized version that I used in my any% as well) only 3 days ago, so everybody who is interested, check it out here:
http://www.youtube.com/watch?v=3LLwcB5-WcE
This trick saves pretty exactly 1 sec compared to the currently known bomb jump strategy. A pretty serious time saver right there.
Experienced Forum User, Published Author, Player
(209)
Joined: 2/18/2005
Posts: 1451
hero of the day wrote:
By the way, we can give a pretty accurate estimate on how much time the varialess norfair route saved. 12-14 ingame seconds will be the net gain, with an absurd real time gain on top of that. Seems I wasn't crazy after all when I thought of doing norfair varialess ;)
This is a realistic estimation if you skip the Wave Beam and therefore a large amount of energy refilling there. And yeah, a very clever route choice for this run after all, now there are no doubts about it anymore. :-)
What I wonder, did you got the Missile Pack in the bottom right corner of the green bubble room, or would the additional damage boost force another refill before Crocomire, and therefore likely slow the run more down than just getting it after Ridley?
As for the power rangers, they probably search for new Taco-optimizations before starting their run, which under some circumstances could make the same happen for them as it did with moozooh: not finishing or at least progressing in any runs for years in the end. To not let that happen again, I would recommend them to simply start the run, and just find the optimizations right along with the progress of it, like I did with my RBO and any% all the time.
Their RBO run is certainly one I look with much interest forward to see, especially due to the new route change and therefore the even harder conditions to pass in LN than in my run.
Experienced Forum User, Published Author, Player
(209)
Joined: 2/18/2005
Posts: 1451
I think I tested it without Space Jump. What I'm sure about is that you gain at least 100 frames in the mini Kraid room in both directions, so thats already 3,5 sec. The area exit and the traveling back through Kraid's room after Varia will together sure save close to one more second as well.
Experienced Forum User, Published Author, Player
(209)
Joined: 2/18/2005
Posts: 1451
Cool. As for the Speed Booster at Kraids area, I think it saves not 2 but ~5-6 seconds in total if used well. Thats what my very crappy quick-test showed at least. :-)
Experienced Forum User, Published Author, Player
(209)
Joined: 2/18/2005
Posts: 1451
Cpadolf:
Wow, this is some serious progress already! Now that you guys passed the most critical Norfair part (suitless) how many seconds did you (roughly) lose refilling in total? Also, was stopping at the pipe before Speed Booster avoidable?
Fano:
Too busy to work on any runs at the moment. SMR WIP5 will hopefully be done soon.
Experienced Forum User, Published Author, Player
(209)
Joined: 2/18/2005
Posts: 1451
hero of the day wrote:
that's pretty old news actually. Catnap was the first to post of this tech on his website about a year ago. I consequently posted his smv on this forum a while ago. Not only that, but I applied the same trick to go downward, which saved time by not going all the way around (to get to the bottom).
Yeah, I was aware of his smv. My method doesn't require the Springball though, and is still somewhat faster because you waste much less time in the air by intentionally cancelling the jump to ghost through the last blocks with immediate landing instead.
And yeah, the Blue Suit Fall is a pretty good timesaver here as well, though only worth to use when comming out of the sand fall room the 2nd time (not right after collecting the Missiles).
On another note, how is the progress of your runs guys? Everybody has been quite about it for a while now.
Experienced Forum User, Published Author, Player
(209)
Joined: 2/18/2005
Posts: 1451
Now that the Blue Suit Jump trick is out, I might as well share another serious timesaver you can get with it. Only applicable in a 100% run though, so don't miss it Hero and Cpadolf:
Snail Shortcut
Saves more than 1,5 seconds right there (compared to JXQ's 100% v2 run).
Experienced Forum User, Published Author, Player
(209)
Joined: 2/18/2005
Posts: 1451
Damn, yet another spoiler...
I knew about this jump long time ago already, and planned to use it in my RBO and any% (ingame-oriented with HJ-Boots) runs. This trick is actually based on the very same technique that is used in the Temporary Blue Suit glitch I posted a while back, that allows to get the Maridia mainstreet Missile without the Gravity Suit for example.
Great job to find this tech yourself, evilchen. :-)
Oh and to save your time, there is no need to try doing this jump without HJ-Boots. It's unfortunately impossible without them because the jump is just not high enough that way, no matter what you do. :-(
Experienced Forum User, Published Author, Player
(209)
Joined: 2/18/2005
Posts: 1451
moozooh wrote:
I am. Their earlier WIP is one frame slower there than yours and mine, since they didn't wait a frame before jumping. That's why Kriole said it saved two frames: compared to the WIP you posted it saves only one.
No moozooh, delaying a frame before jumping isn't faster than doing it immediately (which I did in my WIP as well). The double walljump is 2 frames faster than the best you can do if landing on the ledge, just compare it yourself with the best you achieved to see. It would be good if you try to be precise in your analyzes. As you see, guessing without a true confirmation isn't enough here, even if you think to be sure about it at first.
Experienced Forum User, Published Author, Player
(209)
Joined: 2/18/2005
Posts: 1451
Cpadolf wrote:
If his last any% was done completely without frame-advance I'm very impressed, sure it is improvable by a fair amount of frames even if only using what tricks where known at the time but it is about a 1000 times more optimized that anything I could ever produce with mere slowdowns.
I checked his latest any% carefully, and it indeed seems that he didn't use frame advance at all on it. You can see it clearly for example when he walljumps OTES up, because the delay-times when he releases the jump button and then re-presses it again for the next walljump varies between the optimal 1 and 3 frames all the time. Also the dash button during the short charges is longer pressed than necessary, thus increasing the distance to charge the spark. I don't think he would delay this critical moments unnecessary if he would use frame advance here.
And I totally agree, for a non-frame advance run, it looks extremely well optimized.
hero of the day wrote:
I don't feel like reading all this, but Cpadolf was certainly the first (out of me and him) to understand and use the momentum tech in a run. He used it in the red brinstar shaft well before me, props to him. After I learned about it, I just looked in a million different places for practical uses of it.
It's great to see that there are still a few people left who can admit such things in this days. My respect for your fairness and honesty, hero.