The reason I revisited Circus Caper again was due to an excellent discussion over the acceptance of the currently published run.
While judging, Feos demonstrated a way to improve movement over a few frames in one section of the game. He suggested that my submission exhibited less-than-optimal play for not using this improvement.
At the time I pointed out that even a minor improvement in one part of the game may lead to a longer overall run due to RNG changes (and conversely taking one portion slower can result in a faster overall run due to RNG). I suggested that until the demonstrated 'improvement' was carried through to the end of the game showing overall improvement, it shouldn't be the primary reason for a run to be rejected.
While Feos agreed that RNG changes may result in a longer overall run, he disagreed with my perspective (because my submission was lacking the perceived improvement) that my submission should be accepted as 'current best.' Thankfully Feos agreed to render judgement as 'Delayed' while I checked into the potential improvement.
After completing the level in question using the anticipated improvement, RNG changes resulted in the final time for that level being longer than what it was in the submitted run. Accordingly Feos updated the judgment to 'Accepted for Vault' as the fastest known completion of the entire game.
So what?
I felt vindicated to a degree, as I had 'proved' my perspective that partial improvement does not always result in overall improvement.
That said, I was curious if my own argument of taking one section slower to result in a faster overall run was also valid. Thus I finished the run with the anticipated 'improvement' that unfortunately resulted in a slower level and carried it out through the end of the game. In this case, my concept of slower portion = faster complete run was accurate.
So to make a long story short...
Feos found an anticipated improvement in one of my runs that didn't result in time improvement over the course of the level where it takes place, but it did result in a shorter run of the complete game. This submission.
How I look at this whole situation...
Strange as it seems, both Feos and I were correct in our perspectives (variety is one of the things that makes TASing and this website so awesome). Feos was correct that the short term improvement would theoretically make a shorter run overall. While I was correct that a short term improvement made something else longer (the level), and I was also correct that something longer in one part (the level) resulted in a shorter overall run.
To whomever judges this submission: This run does use the movement pattern suggested by Feos. So if it is more appropriate to include his name as part collaborator, please add it.
Joined: 4/17/2010
Posts: 11492
Location: Lake Chargoggagoggmanchauggagoggchaubunagungamaugg
Oh man, please don't make me debug this freaking RNG :D
Warning: When making decisions, I try to collect as much data as possible before actually deciding. I try to abstract away and see the principles behind real world events and people's opinions. I try to generalize them and turn into something clear and reusable. I hate depending on unpredictable and having to make lottery guesses. Any problem can be solved by systems thinking and acting.
Oh man, please don't make me debug this freaking RNG :D
Nothing says you HAVE to.
But...if you're so inclined...I'll help however I can.
As I mentioned in the description...it was more of a curiosity to revisit it in the first place.
Joined: 4/17/2010
Posts: 11492
Location: Lake Chargoggagoggmanchauggagoggchaubunagungamaugg
You can try yourself:
http://tasvideos.org/ReverseEngineering.html
It'd be hard first time, but it's worth it, since you'll be able to fully control the game. Relying on trial and error isn't an effective way of tasing anymore.
Warning: When making decisions, I try to collect as much data as possible before actually deciding. I try to abstract away and see the principles behind real world events and people's opinions. I try to generalize them and turn into something clear and reusable. I hate depending on unpredictable and having to make lottery guesses. Any problem can be solved by systems thinking and acting.
Only 8 frames improvement? Seriously?
This movie is not so greatly optimised as one might think from reading such description. (Even the densest amount of references to feos per string of submission text can not deceive me.)
Previous run was quite sloppy... and this one is a still sloppy.
Every one can spot ugly place in level 1 (the most obvious place). And guess what? Is still can be noticeably improved - you can easily save yet ~half of second at this corner.
Here non-optimised movie (just a buttons mush) which is 29 frames faster:
http://dehacked.2y.net/microstorage.php/info/171978829/Circus%20Caper%20-%20Any%20%25%20-%20DrD2k9---Ar.fm2
It not require advanced Reverse Engineering skills. It is more likely Basic Tools skills:
http://tasvideos.org/TasingGuide/BasicTools.html
I show you how deep the rabbit hole goes.
Current projects: NES: Tetris "fastest 999999" (improvement, with r57shell)
Genesis: Adventures of Batman & Robin (with Truncated); Pocahontas; Comix Zone (improvement); Mickey Mania (improvement); RoboCop versus The Terminator (improvement); Gargoyles (with feos)
Archanfel, TheRealThingy:
Did either of you read any of the discussion for the current published run? Did you completely read the comments for this submission?
Your arguments are exactly what this submission's comments are all about.
Both of you have claimed that the current submission is sub-optimal/sloppy in the same way that Feos did when I submitted the currently published movie. You've made an improvement over a few frames, but neither of you have shown anything that guarantees your 'improvement' will not negatively affect RNG and that the rest of the run will play out at the same pace or even faster resulting in a faster completion.
The whole point of this submission is to show that doing one part of a game slower can still result in a faster overall run.
And yes, I realize 8 frames is a minuscule improvement, but it is still better. It's also why I indicated the intended sarcasm regarding those 8 frames in the submission comments.
Furthermore, Archanfel, I may still be a relative newcomer to TASing, but I'm not so naive to think that name dropping will somehow improve my chances of a submission being accepted. Feos had a legitimate part in why this submission even happened, it's only proper to give credit where credit is due.Feos:feos wrote:
You can try yourself:
I'll look more into learning how to reverse engineer, but I know it is currently beyond my understanding. Hopefully that will change soon.
feos wrote:
Relying on trial and error isn't an effective way of tasing anymore.
While trial and error may not be the most effective or efficient method of actually TASing, it is still a method that those of us learning can use as a foundation to build knowledge of new methods upon.
We shouldn't expect all new people to the TASing community to use the most up-to date or advanced methods from the moment they join. If I had perceived that as a requirement, I would likely have never started because I wouldn't have felt that I had the necessary knowledge or skills to participate.
Joined: 4/17/2010
Posts: 11492
Location: Lake Chargoggagoggmanchauggagoggchaubunagungamaugg
That's why it's not a requirement. It's a personal taste: when I don't know how something I dislike works, and how to make it do what I like, I go and learn how to control the thing. My intent is to help you learn newer skills, especially when this run is a perfect example of things we need to understand and we don't.
If you have any questions about reversing, I can guide you.
Warning: When making decisions, I try to collect as much data as possible before actually deciding. I try to abstract away and see the principles behind real world events and people's opinions. I try to generalize them and turn into something clear and reusable. I hate depending on unpredictable and having to make lottery guesses. Any problem can be solved by systems thinking and acting.
My intent is to help you learn newer skills, especially when this run is a perfect example of things we need to understand and we don't.
If you have any questions about reversing, I can guide you.
Thank you. One of the things I have appreciated about TASvideos.org is how many people in the community have been willing to help and teach.
the current submission is sub-optimal/sloppy in the same way that Feos did when I submitted the currently published movie. You've made an improvement over a few frames, but neither of you have shown anything that guarantees your 'improvement' will not negatively affect RNG and that the rest of the run will play out at the same pace or even faster resulting in a faster completion.
Do not confuse soft with warm. There is movements optimisation and there is luck manipulation. I just pointed that movements optimisation in this run is still far from perfect.
I show you how deep the rabbit hole goes.
Current projects: NES: Tetris "fastest 999999" (improvement, with r57shell)
Genesis: Adventures of Batman & Robin (with Truncated); Pocahontas; Comix Zone (improvement); Mickey Mania (improvement); RoboCop versus The Terminator (improvement); Gargoyles (with feos)
This movie has been published.
The posts before this message apply to the submission, and posts after this message apply to the published movie.
----
[3393] NES Circus Caper by DrD2k9 in 07:03.75
