Submission Text Full Submission Page

Emulator Version:

BizHawk V. 2.3.1 (x64)

Emulator Settings:

Default

Programming Errors Exploited:

While there are no overt programming errors that are exploited to save time, at several points in the run, I repeatedly jump in front of a barrel, pressing jump again as soon as I hit the ground so that I don't take damage (taking damage causes Pitfall Harry to stop moving for a fraction of a second). I believe this works due to some kind of asymmetry in Pitfall Harry's walk cycle. By repeatedly jumping, his walk cycle keeps being reset, allowing you to keep running without stopping when you land from a vine with a barrel right behind you, which saves a quarter of a second or so each time. An example of this can be seen at 6:53 in the encode of the run seen below.

Hits/Damage:

At 6 points throughout the run, I fall down into pits, which costs me 100 points off of my score each time. This is why I finish with a score of 113,400 points, instead of the maximum score of 114,000 points. The reason why I did this is that falling into pits is a half second faster than climbing down ladders, since it skips Pitfall Harry's animation of getting onto the ladder when he starts to climb down, and getting off the ladder when he reaches the bottom.

Goals:

The goal of this movie was to beat Pitfall! as quickly as possible. In doing so, I ignored my in-game score. To me, this makes sense, since thousands of players have gotten the best possible score in Pitfall!, but no person has ever beaten the game this fast before.

Game Description:

Pitfall! is a game released for the Atari 2600 in 1982. In the game, you play as a man named Pitfall Harry. Your goal is to explore the jungle you are in and collect all 32 treasures in the Jungle before time runs out. You are given a total of 20 minutes to accomplish this task, and you also have only 3 lives. Harry loses a life whenever he falls into one of the tar pits, when he falls into the water, when he touches a snake, when he touches a fire, or when he touches one of the scorpions. Additionally, Harry gets frozen in place for a second when he touches a stationary log, and if one of the rolling logs touches him, he is frozen in place until the log has rolled completely through him. It is worth noting that in this game Harry runs and jumps at the same speed, and also runs at the same speed as the rolling logs. As such, if a log is 15 pixels behind you at the start of the screen, if you run to the end of the screen without stopping, it will still be 15 pixels behind you at the end.
The treasures, hazards, and starting location of Pitfall Harry are always the same at the start of each game. Additionally, the game is structured as a circular maze, where if you head right long enough you will eventually wrap around the map and reach the point that you started at. Additionally, each underground exit takes you 3 screens forward in that direction (ex. if you take the underground exit to the left, you will appear 3 screens to the left of the current screen, while if took the above-ground exit, you would have appeared one screen to the left). In order to beat the game in the time limit, you must use the underground passageways.

TASing Process:

One of the more difficult parts of TASing Pitfall! was to figure out when it was faster to take the vine over an obstacle or to walk/jump over the section that the vine covered. In most cases, walking was faster, since it skipped waiting for the vine to reach Pitfall Harry, to pause, and then swing to the opposite side of the screen. However, in a few cases it was faster to take the vine if the vine was positioned such that it was already starting to swing to the opposite side as soon as Pitfall Harry got under it. In general, if a vine gets in front of Pitfall Harry for any amount of time that he walks under it, it is faster to take the vine. If the vine is always either directly above or behind Pitfall Harry, then it is faster to walk. Of course, for most screens, this choice didn't occur, since you either had to take the vine to get over the obstacle, or the vine was optional and would clearly require a long wait to use it. For parts where you had to use vines over a pit that repeatedly disappeared, the timing for when to jump off the vine was harder to figure out than usual. For any tar pit that doesn't disappear, it is always fastest to jump off the vine as soon as you reach the other side, since Harry moves very slowly at the ends of the vine. However, when the pit disappears when you are halfway under it, then it is faster to jump off slightly before this point. However, if you jump off too early, then you will lose the speedboost that Pitfall Harry gets when the vine is at the center of its swing, which will cost you time.
At some points in the TAS, I would just barely be able to make it to a vine before it swung away, or to just barely make it past one of the disappearing pits, which would cause a chain reaction of me just barely making it over the obstacles at the end of the next few screens (since most of the obstacles in the game move on a global timer that is carried over from room to room). On the flip side, sometimes a vine would be just barely out of my reach no matter how soon I jumped, which would cause me to just barely miss being able to make it over obstacles in the next few rooms, causing me to have to wait in each room. This was annoying when it happened, but luckily, there weren't too many points in the run where I had to wait a long time.
Ultimately, this was a fun game to TAS. Pitfall! is one of the earliest games to emphasize fastest completion time as a goal, and also one of the longest home video games released at the time, making it a great candidate for a TAS. To my knowledge, this is the first TAS of Pitfall! for the Atari 2600 to be submitted to TASVideos, and I think it would make a good addition to the site's vault. I hope you enjoy watching this run!


TASVideoAgent
They/Them
Moderator
Joined: 8/3/2004
Posts: 15576
Location: 127.0.0.1
This topic is for the purpose of discussing #6384: Lobsterzelda's A2600 Pitfall! "Fastest Completion" in 18:19.68
ViGadeomes
He/Him
Judge, Active player (311)
Joined: 10/16/2017
Posts: 461
Location: France
Thanks for making it ! it seems like everything is well optimised. Waiting parts could be entertainned but i understand that if it goes to vault anyways, nobody wants to make it. Meh because entertainement is as it is for a A2600 game. But TAS seems good.
DrD2k9
He/Him
Editor, Judge, Expert player (2213)
Joined: 8/21/2016
Posts: 1090
Location: US
I was wondering when someone (besides me) would get around to doing this game. I had started a TAS of it a while back and, frankly, got bored making it so I gave up. As far as entertainment....it's too repetitive to be entertaining to watch. Pitfall was always one of those games that was more fun to play than watch, so moons/stars is out of the question in my opinion. While I feel this run could potentially be matched by a human, I feel it should still be vault eligible. The game has, in fact, already been beaten with a final remaining time of 1:42 by humans (see here). These runs also include perfect scores, not sure if the actual gameplay is slower than this submission or not.
Lobsterzelda
He/Him
Skilled player (1257)
Joined: 3/17/2019
Posts: 282
DrD2k9 wrote:
I was wondering when someone (besides me) would get around to doing this game. I had started a TAS of it a while back and, frankly, got bored making it so I gave up. As far as entertainment....it's too repetitive to be entertaining to watch. Pitfall was always one of those games that was more fun to play than watch, so moons/stars is out of the question in my opinion. While I feel this run could potentially be matched by a human, I feel it should still be vault eligible. The game has, in fact, already been beaten with a final remaining time of 1:42 by humans (see here). These runs also include perfect scores, not sure if the actual gameplay is slower than this submission or not.
Is there still a video of that run? I want to time it to make sure that this run is faster. Also, it strikes me as odd that he finished with both a perfect score AND 1:42 remaining, since this means he would have lost a half second every time he went down a ladder, but still finished with the same in-game time as me. I wonder if there's a different route that he used for his score? To my knowledge, this is considered the fastest route through the game, but I'm no Pitfall! expert, so I could be wrong.
DrD2k9
He/Him
Editor, Judge, Expert player (2213)
Joined: 8/21/2016
Posts: 1090
Location: US
Lobsterzelda wrote:
Is there still a video of that run?
Twin Galaxies doesn't have the video on their site (at least that I can find).
Lobsterzelda
He/Him
Skilled player (1257)
Joined: 3/17/2019
Posts: 282
Looking through Douglas Korekach's (the person who claims to have gotten a 1:42 in Pitfall!) scores, a few odd things jump out at me. For one, in a lot of his games, he sets a world record that is 5 times more than the next closest person to him. For example, in the game BeamRider, he claims to have gotten a score of 999,999, which puts him well above the next highest score someone has gotten in the game, which is about 200,000. It's worth noting that Todd Roger's was tied in first place with him for this game until Todd Rogers was banned from Twin Galaxies. In Solar Storm, he has a high score of 12,620 points, which is more than 6 times higher than second place. This is not to say that he couldn't have finished Pitfall with a perfect score and 1:42 left, but with no video evidence and so many claims of extremely high world records, I am taking his claim with a grain of salt until I can see a video of somebody beating the game that fast.
Skilled player (1416)
Joined: 10/27/2004
Posts: 1978
Location: Making an escape
I'm surprised Twin Galaxies is still taken seriously after all their credibility has been dashed to pieces.
A hundred years from now, they will gaze upon my work and marvel at my skills but never know my name. And that will be good enough for me.
Spikestuff
They/Them
Editor, Publisher, Expert player (2642)
Joined: 10/12/2011
Posts: 6438
Location: The land down under.
Agreeing with the Ferret here. The only Pitfall runs that have actual backing proof are Peter Nadalin aka nads thanks to his runs being available on speedruncom but unfortunately he's PAL. David B Yancey also has a YouTube (D Yancey) however doesn't have a backup of the 2002 video on his account. As for the other two on table (ignoring Douglas Korekach) can't find crap. There is thankfully a video by a YouTuber known as Code Mystics that has this 114,000 score. Link to video (Video is altered to fit in the 10 minute limit YouTube had back in the day) And another by ErkTheJerk for the full 20 minutes. https://www.youtube.com/watch?v=_rjdOS42e5E
WebNations/Sabih wrote:
+fsvgm777 never censoring anything.
Disables Comments and Ratings for the YouTube account. Something better for yourself and also others.
Editor, Expert player (2073)
Joined: 6/15/2005
Posts: 3282
I got 1:46 left by using deathwarps twice: http://tasvideos.org/forum/viewtopic.php?p=329293#329293 Is this submission aiming for most in-game time remaining? Edit: In the same thread is a perfect run (no deaths, no point loss) with 1:42 remaining.
Lobsterzelda
He/Him
Skilled player (1257)
Joined: 3/17/2019
Posts: 282
Based on what I saw in that thread, I'm going to update my movie to incorporate these new strategies. I should have an updated movie finished sometime later today.
Editor, Expert player (2073)
Joined: 6/15/2005
Posts: 3282
In case anyone wants an encode of my 1:46 TAS, I made one just now: Link to video (I didn't make one before because I never expected the sudden interest in Pitfall.)
The8bitbeast
He/Him
Expert player (2633)
Joined: 11/26/2015
Posts: 183
Location: Australia
I would have voted no on this due to the potential improvements, but since you're working on those I'll refrain from voting until they're incorporated. Ladder boosting and death abuse are 2 improvements that come to mind. As for the category I would suggest either going for perfect score deathless (the easier option), or imperfect score with deaths. Doing imperfect score deathless like in this submission seems weird to me. Maybe that's just me? I'd be curious to hear others opinions. I had a WIP of this game too, which I posted in this thread: http://tasvideos.org/forum/viewtopic.php?t=13378 There are lots of other WIPs and information in that thread, so it's worth looking at. That being said, I'm very excited to see this game get attention and I look forward to the improved final product.
DrD2k9
He/Him
Editor, Judge, Expert player (2213)
Joined: 8/21/2016
Posts: 1090
Location: US
The8bitbeast wrote:
As for the category I would suggest either going for perfect score deathless (the easier option), or imperfect score with deaths. Doing imperfect score deathless like in this submission seems weird to me. Maybe that's just me? I'd be curious to hear others opinions.
For this game, deathless only make sense to me if it's faster than a run that has a death; regardless of score. This simply isn't the case. To me, this is a vault game with no real hope for moons/stars. A 'Perfect Score' branch is not the fastest method of completion. In my opinion 100% for this game is defined by simply collecting all the treasures. Deathlessness shouldn't be a required qualification for 100% simply because it improves score; especially when death abuse allows for faster completion while still collecting all the treasures. Therefore, 'Perfect Score' branch is neither the fastest any% completion or fastest 100% completion; it is thus unvaultable. Further, I don't foresee a 'Perfect Score' branch garnering high enough ratings to warrant moon publication, and as it's not vaultable, it unfortunately doesn't really have a place on our site. As the requirements to simply beat this game at any% is equivalent to 100% collection, the fastest any% and fastest 100% runs are the same thing. Thus there is really only one potentially publishable goal of this game for our site; that being a singular vault run.
Lobsterzelda
He/Him
Skilled player (1257)
Joined: 3/17/2019
Posts: 282
I finished making an updated TAS that includes dying twice to skip waiting. However, the two points I chose to die it were different than in FractalFusion's TAS, which I thought would let me save time over his TAS. However, to my surprise, when I finished the run I finished again with a 1:42, except that this one was a few frames slower than my original TAS which had no deaths! The reason for this is likely due to the fact that dying in one screen at a certain point can cause an unpredictable amount of time gain/loss later on, since its possible that a certain death could save 10 seconds at the time compared to not dying, but within 5 minutes could be 10 seconds slower due to missing several cycles on vines as a result of the death. As such, of the 30 or so possible spots where Pitfall Harry is moving to the left and has to wait for a vine or hole to dissapear before continuing, some combination of 2 of those deaths will be faster than all the other ones. To actually figure out definitively what the fastest places to die are by trial and error would be extremely tedious and error-prone, since it would involve dying on almost 30 screens, playing through most of the remaining game, and then dying in one more spot, which would take 30 squared (or 900) total attempts to find. In order to find the optimal solution to this problem without spending 8 months TASing this game, I have come up with an idea. My plan is to make a program that simulates the physics and movement of the game, and then allows for different routes to be tested to find the fastest route through the game. This would be similar to what was done for the TASes of Arkahnoid and Kuru Kuru Kuruin. The rough outline of the program that I am planning to build is as follows: 1. I am planning to code the program in C++, though I could also make it in Java or Python if other people want to contribute and prefer coding in those languages. 2. The program will have an object of type room, which stores a pixel map of the layout of the room, and the locations, types, and numbers of the objects in the room. The room object will also have variables that keep track of how many exits the room has, and if the room has a below-ground portion or not. Additionally, each room's exit will have a link to the room that it leads to. 3. There will be a Pitfall Harry object that has certain basic attributes, such as the pixels that make up his position, what screen he is on, where on the screen he is, and what state he is in (ex. jumping, walking, taking damage, falling etc.). 4. Each of the hazards will be represented by a 2D array of pixels representing their collision, and in the case of the vines, where they can be grabbed from. Each object will store its starting location/state on power on, and will move according to the properties of that type of object. 5. An array of 255 rooms will be used to simulate the game's map system. 6. A PlayGame object will be made, which will contain functions to try to find the fastest path through the game. There will be a fastestRoom() method which finds the fastest route through a room at a particular moment in time. This function will need to have certain logic programmed into it, such as jumping off of ladders being faster than jumping over them, stopping before a pit if there is no way over it instead of running into it, and running forward when it is possible to do so without hitting a hazard. 6. A PlayGame object can also have a route selected, whereby a user selects what order they would like the rooms of the game to be traversed in. 7. PlayGame will call fastestRoom() for each room once a path is selected, and will test through all 900 or so possible iterations of places where Pitfall Harry could die, and will return which 2 rooms should be died in to yield the fastest time for the route. For some of these things, I will be able to program them on my own. For other things, there are certain pieces of information that I would need to get from someone else to be able to simulate it properly. For example, I would need to know how many pixels wide each screen is, what the dimensions (in pixels) are for each object, what speed Pitfall Harry moves at (for each of the possible states he could be in), and what speed each object moves at in each point in its cycle. However, I think that doing this is the most logical way to make an optimal TAS of Pitfall!
DrD2k9
He/Him
Editor, Judge, Expert player (2213)
Joined: 8/21/2016
Posts: 1090
Location: US
That's an ambitious plan. Hopefully it will work out and yield a truly optimal run. I'd recommend cancelling this submission in the meantime and creating a new submission when you're ready to present the new TAS.
Editor, Expert player (2073)
Joined: 6/15/2005
Posts: 3282
You can use this Pitfall! disassembly for information, in case that helps.
Banned User
Joined: 3/10/2004
Posts: 7698
Location: Finland
Lobsterzelda wrote:
Looking through Douglas Korekach's (the person who claims to have gotten a 1:42 in Pitfall!) scores, a few odd things jump out at me. For one, in a lot of his games, he sets a world record that is 5 times more than the next closest person to him.
I wouldn't put any weight to any world record claims at Twin Galaxies, especially those made 20+ years ago with exactly zero evidence, given the website's horrendous track record in terms of standards of verification. (In the most infamous cases that have been the source of huge controversy in recent years, world records were claimed with the only "evidence" being the word of a TG admin who claimed to have witnessed the making of the record... while being a close friend with the speedrunner.) Twin Galaxies has got, at least in the past, a lot of false credibility and authority because the Guinness World Records used them as their official source of video game world records for many years. I get the feeling that GWR did a bit of shoddy research on the TG website and organization, and considered them more credible than they really were, perhaps because the people at GWR didn't really know better when it comes to these new-fangled fancy technologies and video game culture. (I hope they have learned better recently. Haven't checked what they currently use as their sources.)
TASVideosGrue
They/Them
Joined: 10/1/2008
Posts: 2785
Location: The dark corners of the TASVideos server
om, nom, nom... blech, salty!