I liked the one with the weapon switches. It sucks that the pauses take so long in real time. But with the pauses edited out afterwards it looks really awesome.
Build a man a fire, warm him for a day,
Set a man on fire, warm him for the rest of his life.
I think both should be done, eventually, as they each have their advantages: a real-time run would be shorter due to the lack of pausing, while the in-game-time run would get to show off the various weapons more.
Alright, time for an update. I found out the Triple Blade Glitch that I used in my Mega Man 10 run is also usable in Mega Man 9... in the form of the Magma Bazooka Glitch. It bypasses boss invincibility in much the same way the Triple Blade does. With this insight, I ran a real-time vs in-game time fight with Jewel Man in the time trials. The overall time isn't important since I didn't optimize the level. What is important, is the differences between boss fight times.
Here's a comparison video that showcases the differences between the two.
Link to video
To sum up the videoFavoring real-time
Real-time: 5.10 seconds
In-game time: 5.10 seconds
Note: I wasn't even able to switch fire the Black Hole Bomb once, since doing so would result in a net real-time loss of over a second.
Favoring in-game time
Real-time: 9.38 seconds
In-game time: 2.09 seconds
Which one looks more like a tool-assisted superplay?
I think this was enough to swing my vote back over to the in-game time side.
If the in-game time w/ pauses removed wasn't so jarring, it wouldn't be so bad. I still favor real time, mainly cause I know that the lowest possible in-game time run would be this after 1 or 2 levels:
half-second, pause, half-second, pause, half-second, pause...
Realistically, you could probably get a lower real time run (for the first video if you paused once.
I don't mind pausing. I just hate excessive pausing.
If the in-game time w/ pauses removed wasn't so jarring, it wouldn't be so bad. I still favor real time, mainly cause I know that the lowest possible in-game time run would be this after 1 or 2 levels:
half-second, pause, half-second, pause, half-second, pause...
Realistically, you could probably get a lower real time run (for the first video if you paused once.
I don't mind pausing. I just hate excessive pausing.
^what he said
you can't base the vote on 1 boss fight, the levels are going to be incredibly annoying, as it already has been proved
Realistically, you could probably get a lower real time run (for the first video if you paused once.
I don't mind pausing. I just hate excessive pausing.
A weapon cancel (refreshing from the pause menu) takes 74 frames. Changing weapons takes more depending on how far away it is. Naturally, this means for any time saving tactic doesn't save at least 74 frames, a real-time route will remain on the current weapon (usually the buster).
To me, it seems I would I would have to limit the capabilities of a TAS in order to conform to the real-time goal. 74 frames means I'd have to exclude weapon canceling, zipping, little tricks with different weapons, possibly even coiling up ladders and t-blowing up some ledges, and who knows what else. I can't see how this would make a more entertaining TAS when it reduces to the game play to a near buster-only run right for justice.
The hostility towards excessive pausing is acknowledged, but I'm able to watch these pausing segments without being bothered by the breaks. For now, I'm still working on MM10 Hard Bass, but when I move on to this, someone else might need to fill the real-time void.
Well I'm still working on MM10 (Bass/Hard), so as far as I know no one is actively working on this. You're more than welcome to pick up where you left off.
Alright, so after a long drawn out battle inside my head, I've made up my mind. Since TASvideos was built on the foundation of minimizing real-time input, I feel this game should get the same treatment.
A complete run minimizing in-game would become unwatchable once the C. Shot is obtained. Nearly every ledge and enemy would have to be zipped, and the enemies that aren't zipped would be used for C. Shot refills, which also require switching weapons.
Maybe someday this game will receive a quick weapon switching hack. When that happens, I'll gladly destroy the in-game timer. But for now, real time shall be the goal. I'll start this sometime next year. ;)
Yep great decision! Maybe there could be a game time category, just to show whats possible. But it's really great that e are going to get a watchable real time run first. :)
Very dissapointed in this decision, considering the perfect use of weapons is what makes MM9 one of the most interesting runs to watch, including p auses. But looking forward to WIPs anyways.
I suppose Tornado Man won't be first since you won't be abusing T. Blow nearly as much. Will it be back to Splash to start, or something weird like Magma or Concrete?
I share your sentiments, especially having been a fan of MD9 as he dominated the leaderboards with impressive weapon tactics. The pauses don't really bother me, but I've seen how it can easily get out of hand in a TAS. For now, his in-game record (19:18 I believe?) will have to be the limit with weapon switching. His run still has enough actual play time to be entertaining.
I haven't worked out the route yet, but the fastest RTA produced so far is PrivateEye's recently finished 32:18 with a Concrete Man start.
Link to video
Alright, time for some notes on this, in case anyone is interested in hearing about progress.
I'm still comparing the possible starting routes to find the most efficient one. At the moment I'm deciding between two routes:
Tornado -> Splash -> Weakness order
This route is used in the current in-game time unassisted record of 19:18 by MegaDestructor9. However, since I'm going for real-time, I'm not sure yet if this will still be the best route.
I did a test comparing my TAS to the record, and I was able to save several seconds in Tornado Man's stage, but all that extra time was lost when I got here:
TAS is on the left, agwawaf's run is on the right.
I have to jump on that red/blue platform that moves along that track. In agwawaf's run, he arrives at the perfect time to jump on it and continue. In the TAS, I arrived too early, and have to wait for it to come back to me. This ends up putting the clear time for this level at the exact same as the current unassisted record by agwawaf. It can't be manipulated, as the position is determined at the start of the level.
Because of this, I'm also testing an alternate route:
Concrete -> weakness order
This is the route used in the current real-time unassisted record of 32:01 by TheTrueSigma aka privateye. This run looks promising at the moment, because I will be able to keep the time saved in Tornado Man. I've already completed the first stage of this route, and you can view it here:
Link to video
Finally, earlier in this thread, I was torn between running this TAS for real-time vs in-game time. There's a trick unique to this game called the concrete shot zip, that allows for small amounts of frames to gained with every zip. But it's only useful when minimizing the in-game clock, as it takes too many real-time frames to perform. Well, I think I found the solution.
The full game will be ran for real-time, as I've mentioned. But in-game time can be a separate TAS that aims to complete all of the time trials. Simply starting from the top of the list with Magma Man, and working all the way down the list to finish Wily 4 (and possibly the special stage DLC).
There are a few key benefits to this run. First, I start with all the weapons, so I'll be able to maximize zips in every stage. Second, the in-game clock is actually displayed during time trials, so the frequent pausing will at least have a clear rationalization for the viewer.
A TAS that complete the full list of time trials aiming for in-game time has a specific goal that will set it apart from the full game real time run. In addition, it outlines a clear objective that can be improved upon, as an improvement to any single stage will improve the overall time. This makes me think it might have a shot at getting accepted alongside the full game.
I've complete the first stage of this run as well. It starts with Magma Man, and although I know of some improvements that can be made (better energy manipulation and eliminate corner clipping), it and can be viewed here:
Link to video
I will continue to work on both runs simultaneously.
What a shame that MM9 doesn't have quick weapon-switching. Still, I think your plan sounds like a good one, and the two levels you've completed both look good too!
Pyrel - an open-source rewrite of the Angband roguelike game in Python.
TAS is on the left, agwawaf's run is on the right.
I have to jump on that red/blue platform that moves along that track. In agwawaf's run, he arrives at the perfect time to jump on it and continue. In the TAS, I arrived too early, and have to wait for it to come back to me. This ends up putting the clear time for this level at the exact same as the current unassisted record by agwawaf. It can't be manipulated, as the position is determined at the start of the level.
Because of this, I'm also testing an alternate route:
Concrete -> weakness order
This is the route used in the current real-time unassisted record of 32:01 by TheTrueSigma aka privateye. This run looks promising at the moment, because I will be able to keep the time saved in Tornado Man. I've already completed the first stage of this route, and you can view it here:
Link to video
Finally, earlier in this thread, I was torn between running this TAS for real-time vs in-game time. There's a trick unique to this game called the concrete shot zip, that allows for small amounts of frames to gained with every zip. But it's only useful when minimizing the in-game clock, as it takes too many real-time frames to perform. Well, I think I found the solution.
The full game will be ran for real-time, as I've mentioned. But in-game time can be a separate TAS that aims to complete all of the time trials. Simply starting from the top of the list with Magma Man, and working all the way down the list to finish Wily 4 (and possibly the special stage DLC).
There are a few key benefits to this run. First, I start with all the weapons, so I'll be able to maximize zips in every stage. Second, the in-game clock is actually displayed during time trials, so the frequent pausing will at least have a clear rationalization for the viewer.
A TAS that complete the full list of time trials aiming for in-game time has a specific goal that will set it apart from the full game real time run. In addition, it outlines a clear objective that can be improved upon, as an improvement to any single stage will improve the overall time. This makes me think it might have a shot at getting accepted alongside the full game.
I've complete the first stage of this run as well. It starts with Magma Man, and although I know of some improvements that can be made (better energy manipulation and eliminate corner clipping), it and can be viewed here:
Link to video
I will continue to work on both runs simultaneously.
