Submission #5095: Invariel & Spikestuff's NES Bible Buffet in 03:33.28

This is an improvement of 132 frames (00:02.20) over the previous submission.

Bible Buffet a turn-based board game with action stages and bible quizzes. It is simultaneously the best and worst game in its genre, and the only consequence for failure is having to continue playing. It has an incredibly frustrating RNG and a lot of unnecessary delay, but it also has incredibly tight controls during the action stages, and the most chipper announcer of any Bible-based game about buffets.

7:38 AM <Invariel> Can't click link, have to brute force a 10.
7:39 AM <Samsara> you know what's great? context.
7:39 AM <Samsara> last time i brute forced a 10, insert domestic abuse joke here

I am not fond of this game, I make no secret of that. However, when I saw the other submission, I was fairly certain there was a better way. This took me down a rabbit hole that I would not escape from for three days.

I attacked this initially by reading disassembly, and with p4plus2's help, was able to make sense of the RNG. However, I wasn't able to read the required memory to accurately predict the RNG, so I decided to move on and look at other parts of the game. At about the same time, Spikestuff poked me in IRC and what came out of that conversation is the collaboration we present to you. Hand-checked RNG, hundreds upon hundreds of failures, and eventually workable results give us this video.

With all of that said, I'm glad this is done, as it may encourage me to go back to that other thing I was working on, which is a lot less RNG-heavy than this was.

I knew this was improvable!

I'll mainly explain the encode.

So, what's up with the encode? Why is it clean compared to the other TAS?

This is a build version of BizHawk, where Bible Buffet is actually watchable compared to the current version, I know it's a shock.

If you want to see the encode as it is, without it being patched up and given a pretty bow, [dead link removed].

Honestly, it's fine if this gets delayed and we have to wait for the next BizHawk to come out, 'cause it wouldn't be hard to fix the input to support this... maybe.

Veggie Land is ten spaces long, with a bonus spin at space six. Because the easiest way to waste time in this game is not landing on bonus spaces, we do exactly that by spinning the smiley face and having it move us forward ten spaces. The smiley face could have also either given us a Key (useless) or a Bonus Health (marginally more useful, but it would mean not moving, which is the worst thing of all.) Some damage boosting is used to shove us to the exit and win us some frames over Meerkov.

Isn't the potato a vegetable?

The bonus spin is five spaces away from us, and they're the way to travel in style. Not requiring an action stage when you land on them, they're your fast track out of the Land and out of this game. It was possible (three ways) to spin a 5 earlier than we do, but each of them resulted in more spin time, ultimately being slower.

From the bonus space, we spin a 10 to leave the Land. There were a couple of other 10s that took the same time to start spinning, both of which were ultimately 18 frames slower. We go from bonus space to bonus space here, which is the fastest method of travel as it does not require playing a minigame to progress. The reason we don't do this from Veggie Land is because the bonus space there is nine spaces away and there's no way to move nine spaces in one spin.

Why is the icon for Salad Land a potion? There are parts of this game that make no sense!

Salad Land isn't even explored, as we automatically take the shortcut past Fast Food Land and into Bread Land.

This is where all the money is. Oh, wait, no it isn't. Wrong kind of bread.

We spin another 10 and move onto Pizza Land. As before, we found a faster-to-start-spinning 10, but it turned out to be almost 100 frames slower to use.

As Meerkov noted in his submission, it's possible to press Start and Select (Ss) at the same time to exit a stage. You have to do this three times to be allowed to advance, but it turns out to be much faster to do that than to try to beat this maze of a level. We do that, and spin another 10, moving into Bar-B-Q Land

The shortcut is taken automatically and we end up in Snack Land.

Here's where we thought we could most improve the previous submission. We were hopeful that taking the Quiz here (the question mark) would be faster than exiting out of a stage in Liquid Land and things were looking great. With no waiting at all, we spun a 4, landed on the Quiz, failed it in epic, time-saving fashion, and used its reward to send us forward another four spaces. And, if the spinner didn't take forever to spin and move the character, that would have been the end of that. Unfortunately, all of our time savings (six seconds worth!!) got flushed away by the game's inherent take-foreverness, and we found ourselves back here, looking for a fashionable 10 so that we could mimic the other submission in movement around the board.

We found a 10 pretty quickly (having to check ~100 possible inputs before getting a workable 10), but continued to search for a faster one because that's how this game works. We didn't find one, went with our fast result, picked up some more time and ended up in ...

a level. A level which ended up seven frames slower to play than to quit out of, so we did that.

The next spin yielded a 10 pretty quickly but we searched more of the space and saved three frames on our initial arrival time, catapulting us into Fruit Land and a level.

It turned out to be faster to play this level (as with the previous submission), so we did. What followed, however, was the biggest disappointment we had faced when working through this game.

Idling for four frames and pressing A, there it was! A 10! Alright! And, by the time the character stopped moving and we started the victory ceremony, our 80-frame lead had turned into a 90-frame loss. The RNG was determined to shut this run down right at the end.

Frustrated, we kept searching every possible sequence of input until we found something: idle for three frames, hold anything for three frames, release for two frames, and press A. 1658 frames spent in transit compared to Meerkov's 1706 and our initial 1880! Success, victory, we've finally managed to reach the buffet.

Samsara: Oh, thank the Heavens above, I got here before someone else could snatch up this hot submission!

Samsara: Setting this to Delayed at both authors' request, pending the next BizHawk release which will fix the graphic issues.

Samsara: File updated with a more accurate version done on BizHawk 1.11.7, which is 10 frames slower due to NESHawk accuracy improvements but no longer has graphical glitches. Also, finally accepting to Vault. Holy hell. It happened. It's over.

fsvgm777: Time to finally publish this.