Experienced Forum User, Judge, Published Author, Skilled player
(1279)
Joined: 9/12/2016
Posts: 1645
Location: Italy
What a nice A2600 game! I couldn't keep myself from trying to improve this submission, and I must say that the solutions used seemed quite good! However, I found some micro-optimizations that saved some frames there and there, for a total of 21 frames, though the difference can't be noticed by naked eye. Click here to download the movie. Would you like to have me replace the submission with this new movie, and add me as second author?
Experienced Forum User, Judge, Published Author, Skilled player
(1279)
Joined: 9/12/2016
Posts: 1645
Location: Italy
Wouldn't be overall faster of you un-equip your weapon before Sans fight? Since that weapons requires three inputs and you're going to use it many times for the rest of the run, but you don't need any extra power.
Experienced Forum User, Judge, Published Author, Skilled player
(1279)
Joined: 9/12/2016
Posts: 1645
Location: Italy
Patashu wrote:
Will the even more obsolete previous run remain published in perpetuity if it is not possible to beat the new RTA runs?
That's a good question. Could an hypothetical new submission to be considered as acceptable if it forgoes making use of a glitch that could save time, but also make it trivial to replicate with human attempts?
I can think of another case of a game that got an improvement that was unacceptable due to triviality: #5946: Jigwally's GB The Adventures of Pinocchio "game end glitch" in 00:14.80. The fact is, no one has yet attempted to improve the relative current movie with a submission that forgoes using that trivial glitch, so I'm afraid that we never had chance to face the actual problem so far.
Experienced Forum User, Judge, Published Author, Skilled player
(1279)
Joined: 9/12/2016
Posts: 1645
Location: Italy
I'm afraid that there is an issue with this submission: it fails to beat RTA records. What's worse, is that it doesn't feature superhuman gameplay at all. I could consider accepting if the movie did at least beat those records under the RTA timing, but it just so happens that the speedrunning community managed to achieve 9.84 time as well.
However, The8bitbeast mentioned the hypothesis of an even faster run in RTA timing:
TASVideoAgent wrote:
Note: It’s quite possible that there’s an alternate method to trigger the glitch that fades out much faster. As I said in the brute forcing section, I found one with an almost identical setup that fades out 2 frames slower. I haven’t put heaps of effort into finding a faster fade out as that wouldn’t affect the TAS timing.
As SmashManiac mentioned, we recently had a submission accepted for having a real-time ending point faster than the previous publication, instead of beating it by using shorter input length, and with that precedent in mind I could take in consideration the RTA timing.
So this submission could still have a chance to be accepted if the author manages to find an improvement that beats the currently known RTA records, under either TAS timing or RTA timing. The8bitbeast, would you like to attempt to improve this movie further? I'd also set this submission to Delayed status if you expect to need much time.
Experienced Forum User, Judge, Published Author, Skilled player
(1279)
Joined: 9/12/2016
Posts: 1645
Location: Italy
I just want to say that I find it weird that NFL Street 1 and NFL Street 2 could be considered as similar games while on the other hand we accepted games that have been released on different platforms.
I didn't watch the movie, so take this just as a thought about this rule in general.
Experienced Forum User, Judge, Published Author, Skilled player
(1279)
Joined: 9/12/2016
Posts: 1645
Location: Italy
Latest dev build crashes right after launching to me. I get the classic "EmuHawk.exe has stopped working" notice. Here is the additional details that come with the message:
Problem Event Name: CLR20r3
Problem Signature 01: EmuHawk.exe
Problem Signature 02: 0.0.0.0
Problem Signature 03: 5e40e18f
Problem Signature 04: EmuHawk
Problem Signature 05: 0.0.0.0
Problem Signature 06: 5e40e18f
Problem Signature 07: a6e
Problem Signature 08: 0
Problem Signature 09: System.IO.FileNotFoundException
OS Version: 6.1.7601.2.1.0.768.3
Locale ID: 1033
Additional Information 1: e6e6
Additional Information 2: e6e6c208b90db2aa028f8f653cabbf23
Additional Information 3: 4c69
Additional Information 4: 4c69aaf633e14e952967009de2b2072a
Experienced Forum User, Judge, Published Author, Skilled player
(1279)
Joined: 9/12/2016
Posts: 1645
Location: Italy
The header.txt reports a cyclecount of 1116246514, but if I call the function emu.totalexecutedcycles() at the last frame of the movie I get 1116281624. Which one is correct?
Experienced Forum User, Judge, Published Author, Skilled player
(1279)
Joined: 9/12/2016
Posts: 1645
Location: Italy
Looks like the site parser identifies GBCHawk movies as GB platform, probably because the header doesn't include the IsCGBMode flag that Gambatte movies have. So we either need to implement this flag for GBCHawk or add more logic into the site code.
Experienced Forum User, Judge, Published Author, Skilled player
(1279)
Joined: 9/12/2016
Posts: 1645
Location: Italy
Aww please, I didn't do anything extraordinary... actually, I can't even follow much of the technical stuff behind this game. 😀 I'm putting a yes vote nonetheless, because the movie looks good. 🖒
Experienced Forum User, Judge, Published Author, Skilled player
(1279)
Joined: 9/12/2016
Posts: 1645
Location: Italy
Oh I completely missed that part of the submission text. But in any case, please give more details about the difficulty. As ViGadeomes mentioned, what's harder in the second loop, and why isn't the third loop even harder? And are there any known RAM addresses?
Experienced Forum User, Judge, Published Author, Skilled player
(1279)
Joined: 9/12/2016
Posts: 1645
Location: Italy
All right, thanks.
By the way, here is another thing:
Alyosha wrote:
With both the crystal run and yellow run working, the Gen 2 catch em all should be doable with realistic accuracy (although linking is not tested for accuracy itself.)
Well, not really, since in order to fill the Pokédex in Gen 2 games you need to trade with both a Gen 1 game and a Gen 2 game. So it would need to implement cart swapping in GBHawkLink in order to have a Gen 2 catch em all movie...
Well you can also use glitches in order to catch the remaining Pokémon, but at that point it would make no sense to use the trading system at all.
Experienced Forum User, Judge, Published Author, Skilled player
(1279)
Joined: 9/12/2016
Posts: 1645
Location: Italy
I just tried the BizHawk 2.3.3 and noticed the new sync setting "Timer Div Initial Time". What I find it weird is that it says "Don't change from 0" but the default value is 8, which apparently should only be used for GBC in GBA mode. So, which one is the actually suggested one? Also, if it's known to be platform-related, shouldn't it be automatically changed depending on the mode used? (GB/GBC vs GBCinGBA)
Experienced Forum User, Judge, Published Author, Skilled player
(1279)
Joined: 9/12/2016
Posts: 1645
Location: Italy
In this case, the very first thing I decided to check was to look for addresses that changes once that the player gets to 1 HP remaining. That didn't help much, as the only result was the HP address itself. Then, I looked for addresses that change each time the beep sound is played... And I found a bunch of them. Among these, only the one I've posted is directly responsible for the beeping sound. Specifically, its value changes twice for each time the beep sound is played. Forcing it to stay to zero prevents the game from being able to play that sound effect altogether.
Regarding Lua, even if it can read and edit game memory once for frame, it could allow to many neat possibilities with a bit a logic. For example, take a look at this Cam Hack script I wrote: http://www.youtube.com/watch?v=cYonwsY-VsE source code here: https://github.com/E-Sh4rk/KururinTAS/tree/master/OoB%20Viewer
Experienced Forum User, Judge, Published Author, Skilled player
(1279)
Joined: 9/12/2016
Posts: 1645
Location: Italy
Making a Lua script would be more appropriate and lead to the same result. I'll try to write one in the close future.
Edit: turns out it's even simpler than that: just freeze address 4EE8, IWRAM to zero. Done!
After some tests, I've concluded that it's TAStudio's fault. I've just opened an issue about it: https://github.com/TASVideos/BizHawk/issues/1845