Hi, unfortunately VBA-M is not accurate enough (it runs too fast), so it is not allowed here and in most GB games. Also, currently there is no mobile emulator which would be accurate enough to allow for speedrunning.

If you are interested in the numbers, please have a look here: https://www.kleemans.ch/static/gb-emulator-timing/

Thanks! I added the 5 and 10 games categories. Let's wait a bit on the issue with the full word list to see if there's any development, we can still add it later.

Hi all!

Two topics about new categories and the future of the game.

1. To really allow for competition in a Wordle-sense, I would suggest to add a "5 games" or "10 games" category (or both). There is a Wordle discord and it seems also the webgame is run in both 5 and 10 games categories.

2. The game as we play it now - the ROM by stacksmashing - only features a really small subset of what the original Wordle game offers - only 212 words can be the word searched for (see source), and determining if a word exists in the dictionary is done via a bloom filter which has its flaws.

I tried contacting the original author via email and twitter but he did not get back to me yet, it seems from my point of view that like it was more of weekend project and not something they plan to improve.

As I really see this as a problem for having the "real Wordle experience", I had a look and there are forks of the original code which are actively maintained and improved, and one that stuck out to me was this one: https://github.com/bbbbbr/gb-wordle

It was developed on top of the work of stacksmashing, but features the full Wordle word list, full dictionary for checking valid words, and showing the correct word on losing and other small improvements.

I would be very interested in what you think of this. For me personally, the main issue is the small word list we have currently and I think it would be much more fun to play on a bigger/real word list. I would suggest to add a version with the real word list somehow on the boards, beside the current version.

Also, any idea how we could merge both ideas? One idea that came to my mind was having "1 game" and "10 games" categories with a leaderboard variable "Original ROM" vs. "Full word set", which would show up as a toggle in the respective category.

Hi, just came across this thread by accident. I recently tested the 6 most popular Android emulators and all of them were pretty inaccurate, you can see the results here: https://www.kleemans.ch/static/gb-emulator-timing/.

As most of them are emulating too fast, they provide an unfair advantage and should not be allowed.

Hi, I saw that this game is listed with the platform GB too (besides GBC). I think this is wrong, there was no release for GameBoy and the GBC version is not playable on original GB platforms (as it is not a black cartridge). Could the "GB" platform be removed? Thanks.

Hi, I came by this post by accident but maybe I can provide some information. I did some accuracy testing for Nostalgia.GBC and unfortunately it is not accurate at all, so I wouldn't use it for speedrunning.

You can find the detailed results here: https://www.kleemans.ch/static/gb-emulator-timing/

Hi, thanks for your answer and details. I tried with the original emulator provided by Texas Instruments (trial version), but that one doesn't seem to handle the TI-Boy emulator and just crashes.

Thanks for the hint about CEmu, that one is running much better, and I got the TI-Boy up and running just as you described.

I also did some first tests, with the following results:

• Whitebox testing: I ran blargg's tests and they all pass, so CPU instructions and also timing is working fine.
• Blackbox testing: I ran the two most basic tests with real games, KDL () and SML (), and both are slower than expected. For SML, it was only about 0.4%, but for KDL, which is the basic loading test, it takes nearly double the time of what is expected. The emulator also states that it's running at around 52% of original speed. So loading sprites and preparing a game scene (which is what happens for the Kirby door change) is just very slow.

So although TI-Boy seems to be a pretty good emulator in terms of correct emulation, it will be slower (and for some games/scenes much, much slower) than a real GameBoy. In my opinion that makes it unsuitable for speedrunning, as runs will not be comparable to real hardware (or even other emulators).

If you still want to use it for doing a run, the question would be if it runs consistently slower than real hardware, which I would guess it does. So one could argue that it's up to you to use a slower emulator, as you're not getting an unfair advantage out of it (at least if there are no TI-Boy specific glitches available). For me personally I would be fine to accept a run which is done on a slower emulator like TI-Boy, but I don't know about @livelyraccoon.

Side note: Is there a specific reason you want to run Hook? It seems a pretty big game to test out an emulator on a calculator, you could also try a shorter game first. But that's just a thought from my side :-)

Update: I tried to get a working setup myself, but failed so far.

With OS version 5.6 and arTIfiCE, the jailbroken shell hangs when trying to start up TI-Boy CE v0.2.0b. (I also tried installing a different shell like Cesium, but no luck with that either.)

With a lower version of the OS, 5.4, no jailbreak is needed, but the TI-84 Plus CE emulator hangs when starting up the emulator from the programs ("prgmTIBOYCE").

I'm not sure TI-Boy CE (the GB emulator) is supposed to run on a TI-84 CE emulator? Maybe the GB emulator only works on a real hardware TI-84 CE device.

@greenturtle537 Please let me know if you found a setup with a working GB emulator on an emulated TI-84!

Hi there!

I spent a lot of time this summer comparing gameboy emulators, so I am very interested in what you're planning to do :) I did test every emulator I could find and run, including Smartphone or Raspberry Pi emulators. You can find an overview here: https://www.kleemans.ch/static/gb-emulator-timing/. (It does not cover any calculator emulators though.)

But back to your post. Basically the requirement for any emulator to be suitable for speedrunning is to have the same (or very similar) behaviour of original hardware. I would say the two most important points considering that are:

• Accuracy: The emulator should emulate behaviour like a real GameBoy, down to the CPU instruction level. No glitches should appear in the emulator which are not present on a real device, and ideally quirks and bugs of the original GameBoy are emulated as well.
• Speed: The emulator should run at the same speed as the original GameBoy so times are comparable.

(Some communities like Pokemon speedrunning have stronger requirements and only allow a single emulator.)

To see if your setup works out as described above, I would suggest to do some basic tests, both for accurate emulation and speed. I can provide you with test ROMs and instructions on how to do them if you want, and you can send me the recordings so I can analyze them. Alternatively you can provide me with the exact software and versions you're using, so I can do the tests myself.

If your setup turns out to be accurate enough, I'd be happy to see a run done on a calculator!

Hi all, I stumbled upon this game by searching for Game Boy games.

Just a small remark about the game metadata: "Worms Armageddon" was never released for Game Boy (GB), only for Game Boy Color (GBC), so the platforms could be updated accordingly.

Source: https://en.wikipedia.org/wiki/Worms_Armageddon

Thanks!

Hi all,

Since the top runs for the "Beginner" category are really close, rules for frame-based timing have been set in place a while ago, and this post aims to clarify how to find the key frames.

Start

Start: Timer starts after "class" is selected. Key frame: First frame that shows signs of fading.

End

End: Timer stops when Congrats screen is seen. Key frame: First frame the award ceremony is fully faded in, at which the upper part of the stage will be checkered.

Calculating time between start and end frame

Now that you have the two key frames, how to measure the time between them?

• Youtube: You can navigate with "." and "," until you find the right frames and then put them in here: https://slashinfty.github.io/yt-frame-timer/ Be sure to get the right FPS and it will calculate the time between the start and end frame.
• Video editing software: Usually some kind of timer is available which can be used to count frames or time directly. There's VSDC for example which is free and has a timer available.
• Twitch: No direct way to do this to my understanding, download the raw video and progress using the method above.

If you're not sure about your time, don't worry, as all times will be checked in the approval process.

I'm currently discussing the whole emulator timing with @Legs and am also testing all available emulators on macOS, including OpenEmu 2.3.3 (with Gambatte Core). Hope to have some results soon, especially as there is not much of an option currently if someone wants to run on macOS.

Hi all, thanks for your remarks and suggestions. Also, please stop necroing this 3-year old thread :-)

I discussed some of these ideas with others, but as I am the only active mod for now, here is my opinion on the whole issue.

• There will not be a "longest run" category (where we would convert times to a score), that is against speedrun.com's main concept of tracking fastest times, not high scores. I am aware there are a few exceptions on the site but the general stand on this is pretty clear.
• For other categories like 10m points, 50m points, 10k coins or 50k coins I don't really see the point, as the speedrun and strategies will be pretty much exactly the same as with the existing category, just the length will be a little different. What is the point of having such categories? It's not that this would add additional parts of the game to be unlocked or allow for a different gameplay.
• Obtaining all unlockables as suggested by @KeithHood actually is a completely separate idea and would allow for a really different category, with some strategy and a bit of luck behind it. If somebody wants to start running this, please go ahead! A category will be created if a first run is available.

Thanks.

@BambooShadow Thank you so much for the maps and updating the links. Could the link at "Resources" also be updated? It gives a 404 currently

From my point of view, the 10'000m category is already a quite short category, as a run can be done in 3-4 minutes. So I don't see the point of making an even shorter category.

The other point with no power ups or head starts is an interesting suggestion, but I'm not sure if that makes sense, as all runs would be pretty much the same length?

By the way, in my run, I don't use super headstart (only regular), so getting a WR without it shouldn't be too hard.

Hi, I'd like to request moderation for https://www.speedrun.com/spankys_quest_gb I have a pending run since about a month and tried to get in touch with the Mod several weeks ago. They have retired from speedrunning and said they would be glad for someone to take over the board, but nothing has happened since (last message received on 9.2.21 - I can provide the conversation per DM). Thanks!

Good idea. I added individual COM levels and removed the all-in-one category.

I only started playing into the Full Game category recently, but the suggestions sound very reasonable.

If the differences between platforms are negligible, it would lead to a much more simplified and cleaned up leaderboard. Also the difficulty variable doesn't seem to add much to the leaderboards so I think it's best to remove it.

I only have two remarks:

• To have the "Platforms obsolete each other" toggle on Off doesn't really make much sense with the new system. If the whole point is that runs in the new categories share a leaderboard and thus compete against each other, it doesn't make sense to track older, slower runs. I only think this should be considered because it might smoothen the transition so that players don't "lose" visible runs.
• I'd suggest to switch the sorting of the boards to IGT. This might reduce differences caused by loading time. IGT is the most accurate way to track time and it's a gift if the game provides it, so in my opinion it should be used.

Hi @SkeletKasket, sure, let me know when you have a run ready so I can add the categories. I'll wait until then to avoid having empty categories.

@ACPlay18 Thanks for fixing this!