Editor, Experienced Forum User, Judge, Published Author, Expert player
(2081)
Joined: 8/21/2016
Posts: 1017
Location: US
On a note completely unrelated to TASing (as cheating is not allowed)...
After examining some of the game files...The "holy shit" cheat can also be performed with any of the following text inputs (Warning: potentially offensive Language):
holy asshole
holy cock sucker
holy cocksucker
holy dick head
holy dickhead
holy fart
holy fuck
holy motherfucker
holy piss
holy shit
holy shit head
holy shithead
Editor, Experienced Forum User, Judge, Published Author, Expert player
(2081)
Joined: 8/21/2016
Posts: 1017
Location: US
Radiant wrote:
DrD2k9 wrote:
I however can't get the returned value to calculate in Excel; perhaps because the seed^(seed>>8) portion can yield a result too great for excel to calculate.
Also, ^ means xor, not power.
Well hey, that makes writing the formulas a bit more difficult, but actually yields a result!
Editor, Experienced Forum User, Judge, Published Author, Expert player
(2081)
Joined: 8/21/2016
Posts: 1017
Location: US
Radiant wrote:
Suppose it's 3. Then the next value is ((7C4D * 3) + 1) % 65536 = 29928. The third value is ((7C4D * 29928) + 1) % 65536 = 35273. Then 52598. And so on. (edit) wait no, I forgot about the >> in the return formula. Anyway you get the general idea.
The return command portion of the function doesn't change the seed itself, correct? So wouldn't the above arithmetic be correct for the agi_rand_seed?
Example sequence for agi_rand_seed:
3 > 29928 > 35273 > 52598 > 62591 > 3636 > 30117 > 20130 > 7867 > 53824 > 15681 > 59534 > 47799 > 52492 > 31901 > etc?
I just used excel to find this sequence quickly.
I however can't get the returned value to calculate in Excel; perhaps because the seed^(seed>>8) portion can yield a result too great for excel to calculate.
Editor, Experienced Forum User, Judge, Published Author, Expert player
(2081)
Joined: 8/21/2016
Posts: 1017
Location: US
c-square wrote:
I'm not a judge, but I'd say if it's not going to take you long then just leave it and replace the file.
Regardless of whether the current one is judged/cancelled; I think a separate submission may be more appropriate as we have no idea how the new graphics settings will affect things.
A future submission will hopefully have at least one time improvement, assuming the slot machine RNG can be better overcome. Though I am curious if the longer startup will negate any improvement that can be obtained at the slots.
Editor, Experienced Forum User, Judge, Published Author, Expert player
(2081)
Joined: 8/21/2016
Posts: 1017
Location: US
fsvgm777 wrote:
No garbled text!
However, whether it syncs at all with this TAS is a different story, considering it seems to take longer to boot with the aforementioned VGA BIOS..
I acquired the new (legacy) BIOS and tried to load the movie file using the new settings. The movie ran as if I was using the stock VGABIOS.
Now that we've found a usable image that correctly displays the text, I am happy to redo this run.
At this point there seems to be mostly positive votes (though few) on this run. I can either leave it for a judge to accept/reject knowing a new version will be submitted in the relatively near future, or I can simply cancel the current submission.
Thoughts?
Editor, Experienced Forum User, Judge, Published Author, Expert player
(2081)
Joined: 8/21/2016
Posts: 1017
Location: US
Radiant wrote:
Entering the slot screen calls random() three times, to pick the initial position of the wheels.
Leaving (entering the bar again) calls random() once for the band, and once for the barkeeper's delay.
Playing the slots calls random() only once; turns out the spinning wheels are just for show, the result is picked immediately by generating a random number from 1 to 100 inclusive.
1-2: 000 kills you
4: 222 pays 20x
3,5: 333 pays 10x
6-9: 111 pays 5x
10-16: 11X pays 3x, and calls random(0,3) until it gets a value for X other than 1
17-32: 1XY pays 1x, and calls random(0,3) until it gets a value for X other than 1, then calls it precisely once more for Y
33-100: XYZ pays nothing, and calls random(0,3) until it gets a value for X other than 1, then precisely once more for Y, than calls random(0,3) for Z until is different from X or Y.
It is important to note that randomness does not increment unless the random() function is called, meaning that idling is not going to help you.
Finally, the random function works like this:
int random (int min, int max) {
agi_rand_seed = 0x7C4D * agi_rand_seed + 1;
return ((agi_rand_seed ^ (agi_rand_seed>>8)) % (max - min + 1)) + min;
}
The initial seed is the amount of milliseconds since midnight. That should give you enough for a pattern analysis :)
So if I'm understanding this correctly: If I can somehow manipulate the random variable (perhaps by leaving/reentering the slots)to equal 4, each 'spin' will result in the 3-Diamond results and yield 20x payout?
Editor, Experienced Forum User, Judge, Published Author, Expert player
(2081)
Joined: 8/21/2016
Posts: 1017
Location: US
Radiant wrote:
I'm wondering though. There are easily dozens of different versions of the AGI interpreter, and there are three versions of Space Quest (1.0x, 1.1a, 2.2) that each use a different one. Does this font issue occur in all AGI version? This should be easy to check.
I don't have other versions of SQ1 to test at this time. But in my reading, it seems to be the same problem with nearly all the AGI based games.
Is there a way to compare which graphics cards are emulated by DosBox and JPC-rr?
Editor, Experienced Forum User, Judge, Published Author, Expert player
(2081)
Joined: 8/21/2016
Posts: 1017
Location: US
Radiant wrote:
Oh and yeah, "holy shit" is a cheat code. If you allow that, you might as well press Alt-D on the first screen to teleport to the endgame.
Which is why I resisted the temptation. The only reason I can see it being even considered is that it only shortens a tedious repetitive action without skipping any game material. Whereas, Alt-D teleport skips the entire game.
The best solution though would be to better understand the coding behind the slot sequences to manipulate luck and win just as fast as using the 'holy shit' code. I don't foresee any way to do this without analyzing the game code somehow (which I have no idea how to do at this point in life).
Editor, Experienced Forum User, Judge, Published Author, Expert player
(2081)
Joined: 8/21/2016
Posts: 1017
Location: US
c-square wrote:
In that case, it's not an emulation bug. If the text for the game is garbled under real "console" conditions, then this run should be acceptable IMO, even with garbled text.
Thoughts?
Well, I'm biased, but I hope it can still be accepted.
To further discussion though...I've been searching everywhere I can think of for info on this problem. It seems to be a conflict of some sort between the game's AGI interpreter and the (in this case emulated) graphics cards. It even occurs on many real systems trying to play the game.
The only solutions I have found in my searches, is to play the game using third party interpreters like NAGI, Sarien, etc.
This leads me to believe that DosBox and JPC-rr are likely emulating different style graphics cards. So the problem with the submitted run is not a 'emulation error' per se. I am thus assuming that JPC-rr is emulating a newer style graphics card resulting in the same display bug that many actual systems would also display with these old AGI based games.
Sorry I can't explain any further, I don't understand the software/technology to that depth.
Editor, Experienced Forum User, Judge, Published Author, Expert player
(2081)
Joined: 8/21/2016
Posts: 1017
Location: US
So I've TASed Space Quest 1 on JPC-RR. But as can be seen in this image, the in-game font doesn't display correctly.
Using the same game files on dosbox, the font displays appropriately.
Any suggestions on how to fix this?
Editor, Experienced Forum User, Judge, Published Author, Expert player
(2081)
Joined: 8/21/2016
Posts: 1017
Location: US
A comment on Score Based TAS after reading FatRatKnight's categorizations and preliminary rule guidelines
It seems to me that determining in which scoring category a particular game fits, will be the most the most difficult aspect of implementing any score focused system.
Editor, Experienced Forum User, Judge, Published Author, Expert player
(2081)
Joined: 8/21/2016
Posts: 1017
Location: US
Higher/lower rating has been addressed regarding confirmation bias based on whihch tier the video gets published in. I don't know if the following falls under that or if it is a different concept on rating bias:
Has anyone considered that a significant percentage of members voting pre-publication aren't going back and rating the videos post-publication?
Using myself as an example: I feel comfortable voting pre-publication on whether or not I find a run entertaining; but I don't feel I have the requisite degree of knowledge/skill to determine a published movie's technical rating. Thus I'll likely never rate a movie post-publication.
Further, the only way I know to see how many different members have even rated any published movies is to go through each publication individually and catalog who rated.
Regardless, I'm assuming that the group of members actually doing ratings is a significantly smaller percentage of members than those voting pre-publication.
If this assumption is found to be true, shouldn't the pre-publication votes be more heavily considered when determining entertainment tier? Isn't that the point of the tiers in the first place; to point any interested viewer in the direction of entertaining/high quality videos (whether they know anything about making a TAS or not)? Stars are special runs that most people will find entertaining and contain something extra that sets them apart from all others. Moons are runs in which many viewers will find entertaining, but don't necessarily contain anything to set them apart for Stars. Entertainment value is not even considered for Vault.
Just because a run isn't technically spectacular, doesn't make it any less entertaining to the average viewer.
A second assumption I'm going to make is that those who are rating post-publication videos are a group of members with higher levels of experience.
I fear this may unintentionally skew entertainment (and likely technical) ratings downward as it inherently takes a more impressive run to satisfy players of higher experience.
I'll be the first to admit that I don't fully explore the forums (even if I probably should to a greater degree) unless there's a particular topic I'm seeking.
So I did (mostly) read through the linked topic and I want to know if I understand correctly:
When it comes to tier-shifting a movie, the site considers post-publication ratings as more important than pre-publication voting. Correct?
Mothrayas wrote:
...publications can be moved across tiers at all times whenever deemed appropriate.
When this kind of tier shifting occurs, should a note be made in the submission notes to prevent others from being confused the way I was?
Editor, Experienced Forum User, Judge, Published Author, Expert player
(2081)
Joined: 8/21/2016
Posts: 1017
Location: US
For my own understanding/enlightenment, I'm going to throw a wrench into the engine and see if it causes problems...
Mothrayas wrote:
What makes Virtual Pinball max score and Desert Bus bad are repetitiveness to enormous ends, and the fact that practically nobody is seriously interested in watching playthroughs of their nature in their entirety. But those are still subjective qualifiers in the end, and creating a speedrun archive without subjective qualifiers was the whole point of the Vault.
^emphasis mine.
I recognize that no one (or very very few people) will be even remotely interested in watching one of these runs in its entirety.
However, for Desert Bus, using score as a 'point of completion' was not the reason for rejection. Similarly, length of the run was also not the stated reason for rejection. It was rejected from Vault publication because it wasn't distinguishable from a human speedrun.
Isn't determination of whether or not a run appears to be different than a human speedrun a subjective qualifier?
If we use this subjective reason for rejecting the Desert Bus submission, perhaps all the other Vault-tier publications (as well as any submissions rejected for this same reason) should be reassessed, since "...creating a speedrun archive without subjective qualifiers was the whole point of the Vault"
As an example (and at risk of having one of my own publications un-accepted), NES Castelian is perhaps technically distinguishable from a human run (menu management). But, regarding game-play itself, the Castelian TAS is hardly distinguishable from what a human speed-run would look like if the human player knew the enemy spawn positions/movement speeds and the path to take to the goal. These are things many human speedrunners do know about the games they run. Should this Castelian run then have been rejected for the same stated reason as Desert Bus?
If so, I am in a position to potentially lose a published movie. And maybe others are too, as there may be other current vault publications whose game-play isn't necessarily distinguishable from human speedruns other than the technical management of arbitrary things like menus. If Castelian was acceptable without distinct evidence indicating it wasn't a human speedrun (other than arbitrary stuff like menu management), then acceptance/rejection of Desert Bus unfortunately needs reassessed as well as any other submissions rejected for the same reason.
Please understand, I'm NOT trying to advocate for Desert Bus to be accepted. In fact, the rule for judges regarding triviality may be an even better rejection reason for both Desert Bus and Virtual Pinball.
The run must not be seen as trivial.
If a run consists of doing a trivial strategy, made only nontrivial by having to do it over the course of several rounds, it will still be judged as trivial. For example, bowling games where the player gets a strike with ten pins every time.
Then again, one could argue that 'triviality' could be considered a subjective assessment also.
I guess what I question about something being 'distinguishable from a human speedrun' centers on whether or not arbitrary technical actions (pre-game menu management, frame perfect clearing of level intro-screens, frame perfect skipping of cutscenes which are intended to be skipable, selection of action icons in an adventure game, etc.) are enough to deem a TAS run distinct from a human run when all other aspects of game-play aren't distinguishable; especially when considering that many human runs don't even consider pre-game menu management as part of the timed events, and options are usually selected before time starts--which is usually upon pressing start or sometimes at the first control opportunity after an initial cutscene when starting the game. Human timing also often ends on certain in-game events (i.e. boss death animation, credits roll, "the end"), not on button presses that initiate that event, so this could also be considered a technical distinction not a game-play distinction.
As a side note: A2600 Dragster is a game currently published in the Vault tier even though the judgement was for Moons. This should probably be fixed.
Anyway, Dragster is an example where the in-game time is considered over total game time--this is fine for vault purposes--but the best in-game time of 5.57 achieved by the TAS has been matched in real-time by a human. Should this disqualify the TAS from vault tier publication (had it not been accepted as a moons quality)? Please don't make the argument that it wasn't accepted as a Vault run to disregard this hypothetical question.
I guess my perspective is that the fastest/best TAS is exactly that, the fastest/best TAS. It should be acceptable as long as it isn't sub-optimal or trivial, regardless of whether or not it can be distinguished from a human speedrun. Example: A2600 Adventure is a mere .4 seconds faster than the current real-time record. The only distinguishing characteristics of the TAS is the wobbly movement. But if you showed both videos to someone who didn't know one was done by TAS, they wouldn't necessarily assume it was. Or they would assume both were.
A TAS shouldn't necessarily be rejected just because a human is able to match the game-play aspects of the run once technical arbitrary details are disregarded (i.e. Dragster, Adventure). If that were the case, then TAS runs that only appear to showcase faster speed than a human due to non-gameplay technical aspects should be rejected and/or removed if already published (i.e. Castelian).
These thoughts aren't meant to be restricting what runs get accepted, nor are they meant to have current publications removed. If anything, it's advocating for a slightly broader definition of runs deemed Vault worthy by softening/eliminating the 'distinguishable from human speed run' requirement so that more games can be showcased.
Side note #2: If this post would be better suited to a different area of the forum, please let me know and I'll copy it to that area for discussion.
Editor, Experienced Forum User, Judge, Published Author, Expert player
(2081)
Joined: 8/21/2016
Posts: 1017
Location: US
Thank you very much for the clarifications and perspective.
My own comments weren't meant as advocating for the inclusion of Score-based TAS videos on the current site, just voicing thoughts/concerns that I (and likely others have) had regarding the concept.
Your arguments regarding the current approach to such submissions are exactly why I suggested that a completely separate website would likely be the best option for score based TASing instead of speed based TASing.
Mothrayas wrote:
Sorry, but I am going to instantly reject any proposal that would force us to publish a 74-hour pinball run, or a 792-hour long Desert Bus run.
LOL! I had thought about Desert Bus myself before I posted the list of game examples above, but forgot to include it. I fully expected it to be used as a key argument against ungodly long score based TAS videos.
Thanks again!
Editor, Experienced Forum User, Judge, Published Author, Expert player
(2081)
Joined: 8/21/2016
Posts: 1017
Location: US
FatRatKnight wrote:
We commonly use time as a criteria for obsoletion. Is there a reason we can't use a score value as a criteria for obsoletion as well?
Does the following hypothetical example explain what you mean?
Current publication of game X = 11270 frames with final score of 148765
New submission of game = 11270 frames with final score of 153972
Should the new submission with better score obsolete the old publication even though they are equal on time?
FatRatKnight wrote:
...but that run was accepted in another tier. It therefore did not have to deal with Vault definitions.
Though I'm sure it would be a pain to implement, perhaps a completely new tier or vault sub-tier needs to be considered for runs of maximum known achievable score (assuming the score can't be run to infinity).
This would require modification to the current rules. Specifically the following:* The maximum score is easily defined and absolute - it must not be possible to gain a higher score, even theoretically. It must be definable without being dependent on precise time, speed, or similar requirements.
This would need modified to allow games where a maximum score isn't as easily defined. Also the requirement to not have a theoretical improvement to score would need eliminated.
Honestly, this site accepts speed runs as the best known that have theoretical (yet unproven) speed improvements (many submissions even detail these potential improvement possibilities in the submission text!). If unproven theoretical speed improvements don't restrict a know submission from being accepted as fastest, why should an unproven theoretical score improvement restrict a run of highest known achieved score from being accepted? (Assuming there's a tier/sub-tier for it).
* The maximum score is limited by not being able to gain any more points, not by hitting a score cap or overflow. If it is possible to score points infinitely, score cannot be used to define full completion.
The restrictions on cap/overflow could be changed to accept reaching the cap/overflow point as the maximum attainable score.
Also this rule suggests a question on when a game is considered completed. Currently for endless games, a point of no-new-content is considered completion for TAS purposes. If a maximum score is attainable but long after the game has reached a point of no-new-content, which of these is considered the actual completion point? This question also supports a reason to have a separate tier/sub-tier from speed based runs.
Again, I anticipate that implementation of these ideas would likely be a significant amount of work for the operators of the website. Even if the general community accepted these ideas, It would ultimately fall on the site admin to determine whether or not new tier/sub-tier system would be created and implemented.
If these "maximum score" ideas could somehow be implemented as simply new branches, implementation would be easier. But I think that would be a harder sell to the community.
Perhaps the best solution would be a completely separate website (which could be linked as a sister site) which contains only TAS runs of known highest scores. This would again be more work to implement, but would eliminate the congestion of current tiers and movie lists. "TAS" would then gain yet another evolution to it's definition as "Tool-Assisted Score-run."
____________________________________________________________________
I recently read Mothrayas's judging comments of Submission 5560 which addresses arbitrary goals. Perhaps it would be good to get his perspective on these "maximum score" ideas.
Editor, Experienced Forum User, Judge, Published Author, Expert player
(2081)
Joined: 8/21/2016
Posts: 1017
Location: US
Some further thoughts regarding this concept.
1) If "max score" becomes a valid publishable category, timed sports games could theoretically become a valid game choice with a goal of maximum score.
Examples: Baseball, Basketball, Soccer, Football, etc. for any system.
Though there is no maximum score in many of these sports, the video game mechanics would determine a maximum potential score in the set playtime.
I can imagine there's a subset of TASers who would be interested in some of these type of games.
2) There are games with theoretical maximum scores that would end in longer runs and showcase more game (even if it's somewhat repetitive).
Examples:
Many athletic and non-timed performance based games. (Track & Field type games, Mario Party type games, gambling games)
3) There are some games where the endgame/"game over" animations vary depending on final score (whether it is a maximum or not).
Examples : Gameboy Tetris, NES Track & Filed II
4) There are some games where the endgame credits are only shown if a certain score level is achieved even if it's not the theoretical maximum.
Example: NES Open Golf
5) If this becomes a publishable category, I feel sorry for the judges who have to sit through these potentially very long, potentially very repetitive videos necessary to achieve some of these score based goals.
6) To my understanding, TAS originally meant "Tool-Assisted Speedrun;" with the original intent of TASing games being strictly speed-running. TAS has since evolved to also mean "Tool-Assisted Superplay" as in the title of this website. This evolution allows for speed to not be the only consideration of what makes a great TAS video. The simple fact that we have additional categories beyond the absolute fastest "any%" run demonstrates this idea.
It seems to me that effective definition of "Superplay" for the purposes of this site is the following: A run that accomplishes X goal(s) as fast as possible while also attempting to provide some degree of entertainment value. Some speed/entertainment trade-offs are acceptable.
If that is an acceptable definition, "maximum score" runs should qualify. It also, unfortunately, means that goals such as "fastest death" or "fastest 'Game Over'" screen should also be acceptable.
Which I guess brings us back to the crux of this discussion...is "maximum score" a category we want to accept for publication on this site?...And more generally, which goals are worthy of publication and which ones aren't?
Editor, Experienced Forum User, Judge, Published Author, Expert player
(2081)
Joined: 8/21/2016
Posts: 1017
Location: US
FatRatKnight wrote:
...I see the goal in a similar way as 100% runs, as you are trying to max something out on the way to clearing the game...
I agree that achieving a maximum score (on games with a definable maximum) seems comparable to "100%" or "100+%" run.
In those "100%" type runs, the goal is collecting all possible items and/or completing all dungeons/levels/etc. in the fastest time. However, that still acknowledges that time is sacrificed when compared to absolute fastest "any %" endgame runs.
Setting a goal of "maximum score" is similarly acknowledging that you're (most likely) going to sacrifice time to attain the highest score. In that way, I can see "maximum score" being feasibly considered a publishable category like "100%," "any %," etc.
That being said, the primary difference between most "any %" runs and "100%" runs is that "100%" runs show more of the game than the "any %" runs. Unfortunately, setting a goal of "maximum score" for a run doesn't guarantee it's going to show any more of the game than any other (lower score) run. The only assumption that can be immediately made is that the run will be longer because extra time will be spent to achieve a higher score.
In my opinion, a higher score run based simply on score value that doesn't also show some new game content is unnecessary for this site. Watching such a TAS would essentially be watching a longer version of a shorter TAS with the only satisfaction for watching the extra time being a difference in a displayed digit sprite.
If new worthwhile game content is also showcased while achieving the (theoretical) maximum score, I'm for acceptance as a separate category.
Editor, Experienced Forum User, Judge, Published Author, Expert player
(2081)
Joined: 8/21/2016
Posts: 1017
Location: US
Ready Steady Yeti wrote:
...Though I would have liked a bit of messing around on the wait times, the run is still quite impressive technically. Unfortunately, the run itself was not really very entertaining for me due to an extreme amount of repetitiveness. I voted meh, as the run still satisfied me in a general sense.
Unfortunately many C64 games are probably doomed to fall under the 'meh' category due to repetition; games of that era were often simply about beating the previous high-score, not completing a story line to an endgame. As such, many of them are endless (infinite repeating levels), and the entertainment value from a TAS perspective is limited.
Where modern gaming often relies on the visual quality and detail of games to supplement the entertainment value of the game-play itself, older games during the C64 era were often more concerned with the act of playing rather than the visual presentation for entertainment purposes.
All that being said, I may have been able to consider messing around during wait times a bit more with this run. Though even that would most likely have been limited to horizontal movements on turtles/logs as any vertical movements (and possibly some horizontal) would introduce delayed arrival at the goal. In my opinion those entertainment improvements would not be enough to validate increasing game length.
It's likely that many (if any) C64 games submitted in the future will have their entertainment value judged heavily on mess-around quality; especially when they are of repetitive nature and score based instead of having a definite endgame.
Editor, Experienced Forum User, Judge, Published Author, Expert player
(2081)
Joined: 8/21/2016
Posts: 1017
Location: US
feos wrote:
Bump of interest.
I was surprised there was no submission for this game, then I found this thread...
I may attempt this game after thoroughly reading through everything.
Editor, Experienced Forum User, Judge, Published Author, Expert player
(2081)
Joined: 8/21/2016
Posts: 1017
Location: US
feos wrote:
You should be the first and the main person to judge your runs as you make them, and be hungry for anything that looks suboptimal. It's not a basic skill, but if enough effort is invested, it's always better than less effort :D
Effort isn't the problem; experience is. I have no problem putting in the time. I'm just nowhere near as good at recognizing sub-optimal situations as you more experienced players.
Sure, it's a bit frustrating when you work hard on something and think you've made it the best it can be, only to have someone else point out something that can be improved.
But what is even more discouraging is the self-realization that a much more experienced eye can see improvement possibilities that to them are obvious, while I would never see or even contemplate trying it.
feos wrote:
Improvements like this one...stand out as obvious to try.
This is an example...I would never have considered trying that method. I understand that it's due to my inexperience. But while it may have been obvious to you, it wasn't to me.
feos wrote:
Are you planning to redo this run again?
I don't know. I'm not really excited about it at the moment, but that doesn't mean I won't look at it again in the future.
No garbled text! However, whether it syncs at all with this TAS is a different story, considering it seems to take longer to boot with the aforementioned VGA BIOS..
