Experienced Forum User, Published Author, Experienced player
(608)
Joined: 10/23/2004
Posts: 706
I voted "yes." I had been eagerly awaiting an eventual TAS of this game. I have fond memories of playing this game with the Phantom Menace soundtrack CD playing in the background :) I really appreciated the commentary as it helped me understand what was going on in context. Perhaps my memory of the game is fuzzy but were there other courses that weren't raced for some reason? I'll have to go back to some old videos I had but there was at least one course I remember having a half lap skip that I didn't see in this run.
*edit* here is the shortcut I was thinking of:
Link to videoLink to videoLink to video
Experienced Forum User, Published Author, Experienced player
(608)
Joined: 10/23/2004
Posts: 706
Well dang, man. Nice find! Avoiding a pause is most impressive. Too bad this is sounding TAS only but I guess it makes the TAS even more interesting when it has different strats.
Experienced Forum User, Published Author, Experienced player
(608)
Joined: 10/23/2004
Posts: 706
I found it a bit odd that a guy who spends 3 hours on each level / difficulty combination for RTA spent far less than that on every level and TAS ever. I think there is a whole lot more that could have been covered, but it would need to be from one of the TASers more knowledgeable on the journey.
Experienced Forum User, Published Author, Experienced player
(608)
Joined: 10/23/2004
Posts: 706
WMJ, you're the man! Love this.
Regarding loading times. For what it's worth, Mupen64Plus also had much longer loading times for Mario Kart 64. If you care the final encode to the console verification. The loading times were exactly twice as long on emulator. I winder if there is some easy fix that could be implemented.
Experienced Forum User, Published Author, Experienced player
(608)
Joined: 10/23/2004
Posts: 706
I think a TAS of the Wii U VC release would be ideal. That's probably the most value added release Nintendo has ever made on Virtual Console. Including all of the e-reader levels (which were previously very difficult to actually track down) and even updating the ROM to support all of them. This would make a great TAS.
Experienced Forum User, Published Author, Experienced player
(608)
Joined: 10/23/2004
Posts: 706
Are you using an SSD? Read and write speed would play a major factor here. Possibly you could also make a virtual drive in RAM and just regularly copy to disc.
Experienced Forum User, Published Author, Experienced player
(608)
Joined: 10/23/2004
Posts: 706
Holy schnozberries, WMJ. Amazing contribution here. It would be awesome if you would link this info with some high level info in a TASVideos game resources page for F-Zero.
Experienced Forum User, Published Author, Experienced player
(608)
Joined: 10/23/2004
Posts: 706
Nach, you make a good point.
I think what would really help in this particular case is if the HappyLee run were submitted while MrWint's was on the workbench so the discussion could be more specific and less about general rules and rules of thumb.
My sense is that I would probably prefer MrWint's run based on the descriptions so far provided by HappyLee. The reason being it sounds like MrWint completed levels "faster" if we disregard the frame rule.
Of course, without an actual run to compare to, we fall back on debating general rules which is difficult. It's, of course, harder to define the general theory of something, than the specific application of it.
Unfortunately, it sounds like we wont' see HappyLee's run soon. But, when we do, the comparisons I'm bringing up can be more specifically discussed.
Experienced Forum User, Published Author, Experienced player
(608)
Joined: 10/23/2004
Posts: 706
Nach, I see that you're arguing the definition of what a segment is in the game. I contend that there are at least two categories here: Full game and Individual Levels. I think these are objectively defined and would be supported by nearly any speedrunner. There is a game to complete and there are levels to complete (with future levels to load). If you go the next level down and say there are horizontal bars of pixels to load, that could be argued as a tertiary goal before entertainment. My feeling here is that at that level, entertainment is more important. Also, in the case of an auto scroll level (which was your example) you're not loading those horizontal columns of pixels any faster. So, what you actually are referencing is a quaternary goal which could be considered before entertainment. Specifically, the quaternary goal would be the character's average horizontal position over time.
Clearly the line has to be drawn somewhere before entertainment should come in. To me, it's after the secondary goal of making the ILs as fast as possible. To me, entertainment should be considered no earlier than this point. But, again, that's me. What I saw with HappyLee's comments earlier is that he greatly values entertainment as a secondary goal. It seems based on the above comments that MrWint values entertainment as a tertiary goal. To me, that definition (entertainment as a tertiary goal, behind IL times and full game time) allows for on-going competition even as the primary goal remains stuck for some time.
Experienced Forum User, Published Author, Experienced player
(608)
Joined: 10/23/2004
Posts: 706
Addressing both feos' and scrimpeh's comments. Sonic is actually a game that came to mind as facing a similar issue.
We either call SMB largely a maxed out game (although future improvements may happen, they may also not happen) and focus on improving the entertainment or we prolong competition by focusing first on overall length and then on individual level length (from level n load to level n+1 load to ensure flag pole glitches and anything that reduces overall length are still executed).
The more objective goal is Overall length first, IL length second and entertainment third. If entertainment is second, it becomes harder to judge which run is better and changes the motivations of successive runners.
I'm far more interested in seeing progress towards individual levels approaching the next frame rule than I am in seeing Mario moon walk one more frame or break one more block or shoot one more fireball. Those entertainment pieces, to me, are of tertiary importance.
Experienced Forum User, Published Author, Experienced player
(608)
Joined: 10/23/2004
Posts: 706
Nach wrote:
Weatherton, with that logic, on auto scrolling levels, the player should just always hug the rightmost wall (or whichever way it's scrolling).
When you have free time to play around with, make the most entertaining use of it as possible. That's the next criteria after speed.
I think you missed my point. In an auto scrolling level, by all means jump around and be as entertaining as possible. However, still minimize lag and be sure that you finish the level as early as possible. For example, if pushing Mario forward to the right side of the screen allows you to finish the level earlier then be sure to do that over the course of providing entertainment.
To me SMB is similar to a racing game at this point. Each course can be viewed individually as well as part of a larger run. If by playing around in a race you end up losing some frames and reaching the end more slowly, even if a frame rule exists that would have negated the gained time, I would prefer that you complete the course as quickly as possible. Still be as entertaining as possible, but don't lose time during each individual course while being "entertaining".
Looking at the summary table at the end of this submission text, I prefer this run in all cases to the prior run where frames and sub-frames were saved, even if some little trick was left out because of a frame rule.
Again, I'm sure other people see this differently. This is why I asked for others to post their opinions :)
Experienced Forum User, Published Author, Experienced player
(608)
Joined: 10/23/2004
Posts: 706
I see a lot of references to speed / entertainment trade-offs and stylistics choices in this thread. I figured I'd throw in my two cents and see what other peoples' opinions are.
For a game as close to maxed out as SMB and which contains a frustrating frame rule that limits the benefit of "small" improvements. I much prefer seeing how fast each individual level can be performed and would prefer a run that completes a level faster but ends up having to "wait" a few frames longer at the end. To me, a better run is one that completes each level faster, even if that means waiting a tiny bit at the end. This keeps the competition going even when the pesky frame rule means the overall movie may be exactly the same length.
In summary, when two TASes complete a level in the same number of frames (from loading that level to loading the next) I'd prefer the TAS that completes the portion controlled by the player faster. I am less impressed by superfluous movement that ultimately slows the player down.
It's certainly possible I'm in the minority but I thought I'd post to see where others stand on this topic.
Experienced Forum User, Published Author, Experienced player
(608)
Joined: 10/23/2004
Posts: 706
micro500 wrote:
I am also considering trying to make a 3D 360° video. 3D 360° video is very immersive when watched with a headset, and it should be easy enough to move the camera slightly off center to create the parallax needed for 3D video.
Experienced Forum User, Published Author, Experienced player
(608)
Joined: 10/23/2004
Posts: 706
This is an amazing accomplishment. It's so cool to not only see the concept finally be a reality but also to finally be able to see things in my run that I couldn't before and with a level of immersion that's just amazing for a game made in 1996.
I'm looking forward to Mario 64!
Experienced Forum User, Published Author, Experienced player
(608)
Joined: 10/23/2004
Posts: 706
I thought you all may be interested to see that Beck Abney successfully hit 3/3 "Weathertenko" Choco Mountain lap skips: http://www.mariokart64.com/cgi-bin/yabb2/YaBB.pl?num=1509320389
He streamed his attempts since August of this year, when he really got serious about hitting 3/3. On his 26,462nd attempt of this grind and after hitting 2/3 96 times, he finally nailed a time of 16"38. This puts him almost 40 seconds ahead of second place. Truly a remarkable speedrunning achievement.
Experienced Forum User, Published Author, Experienced player
(608)
Joined: 10/23/2004
Posts: 706
It's great to finally see this replicated on emulator and with an understanding (roughly) of why it's happening. As I posted over a decade ago, this had happened to me on console as well :)
Weatherton wrote:
That sounds familiar, I believe the boss kills himself sometimes... I don't remember the details but something like that happened to me as well.
Experienced Forum User, Published Author, Experienced player
(608)
Joined: 10/23/2004
Posts: 706
I'm willing to provide commentary. It would help to use runs that have detailed submission comments to ensure we can review and give the TAS justice.
I also think it would be fun to do something like the Choco Mountain Gauntlet (but for another game). So, something where a bot sets up a difficult to complete scenario without total control.
