Experienced Forum User, Published Author, Player
(14)
🇺🇸 United States
Joined: 11/13/2022
Posts: 36
Location: 🇺🇸 United States
JosJuice wrote:
You might need to have the "Save Texture Cache to State" setting enabled for savestates to work correctly there. This setting isn't available in older Dolphin versions like 5.0.
“Save Texture Cache to State”? Never heard of that one.
Petrual Emill; got milk?
Oh yeah, specking of milk I was born on the holiday (milk day is January 11th)… so I’m a pretty rare person.
Experienced Forum User, Published Author, Player
(14)
🇺🇸 United States
Joined: 11/13/2022
Posts: 36
Location: 🇺🇸 United States
Spikestuff wrote:
You can literally just click on the "YouTube" link each Cyberscore video has.
You don't even have to look at the top run in some cases since for Snow Go, you're slower than fourth place which does have a video.
Link to video
Chain Event might be within error for instance.
Oh, they don't have runs or is only images? Use the boards on speedrun.com.
WR there by Tsov has Balistic Beach having a personal PB of 12.28 and that run getting a 12.90 which clocks in better than your 13.78.
Tsov's personal PB on Pizza Me, Mario is a 32.65 compared to your 34.41.
Flinger Painting? Tsov PB is a 12.86 which is a far cry to top run on Cyberscore, but their 14.40 clears your 15.90.
xTASz wrote:
Again, is there a video I can watch to make improvements to this TAS?
You aren't using your resources and you aren't searching properly if you're making questions such as this.
Another instance to join the Mario Party Discord which is found in that srcom link I provided.
Ask questions, get answers.
Provide what you know, get information about what to do next.
My apology, I should’ve said where I can find the video on Cyberscore because my access to Speedrun.com on my phone isn’t available. I was more likely referring to the Island Tour time attack TAS, though.
Petrual Emill; got milk?
Oh yeah, specking of milk I was born on the holiday (milk day is January 11th)… so I’m a pretty rare person.
Ah, thanks for that. I couldn’t access Speedrun.com for the records and now I needed this
Is there a video I can watch to see how I can improve the TAS?
Petrual Emill; got milk?
Oh yeah, specking of milk I was born on the holiday (milk day is January 11th)… so I’m a pretty rare person.
Experienced Forum User, Published Author, Player
(14)
🇺🇸 United States
Joined: 11/13/2022
Posts: 36
Location: 🇺🇸 United States
jmosx36 wrote:
xTASz wrote:
Flinger painting’s load state does this awkward action where the paint disappears once splatted on the whiteboard which I sometimes have to be lucky enough to get the inputs to sync to finish the Minigame in record time. Did you have this problem too?
Nope, i did not
In my own TAS it was as simple as trial and error, the shape of every shot in my TAS was consistent to the same frame, so, i did a shot to a random spot, took note of the shape and later put it right
Ah, ok. It’s just the paint erases itself whenever I load state for me.
Petrual Emill; got milk?
Oh yeah, specking of milk I was born on the holiday (milk day is January 11th)… so I’m a pretty rare person.
Experienced Forum User, Published Author, Player
(14)
🇺🇸 United States
Joined: 11/13/2022
Posts: 36
Location: 🇺🇸 United States
jmosx36 wrote:
It is nice to see a MP9 TAS, but it seems that some minigames are a bit unoptimized
In snow go it is better to take the edge of the ramp on every jump, that way you get a better angle for the turn, less air time and more speed
Jigsaw Jumble is a RNG nightmare, you depends on the correct piece and the board to land first and quick, but i think it can be manipulated
Ballistic Beach, another RNG concern, the first and last shot are the most important ones, you must make the first shot as quickly as possible and the last shot must impact as quickly as possible, that's why you depend on the first and last objective to be as close as possible to the cannon, i think the positions of the objectives can be manipulated too
Upward Mobility is unoptimized in this run
On flinger painting you must take advantage of the shape of the stains to finish as quickly as possible
And pit or platter is unoptimized, it is better to manually enter your inputs from the emulator, rather than using an external controller
Meh vote from me
Flinger painting’s load state does this awkward action where the paint disappears once splatted on the whiteboard which I sometimes have to be lucky enough to get the inputs to sync to finish the Minigame in record time. Did you have this problem too?
Petrual Emill; got milk?
Oh yeah, specking of milk I was born on the holiday (milk day is January 11th)… so I’m a pretty rare person.
Experienced Forum User, Published Author, Player
(14)
🇺🇸 United States
Joined: 11/13/2022
Posts: 36
Location: 🇺🇸 United States
nymx wrote:
So are we done with everything?
Well like i’ve said I have no idea if I’ll return to this, but Walgrey said she was gonna point out the improvements to it. And the TASers in the Mario Party speedrunning server are planning a v3 unhacked TAS, too. So… maybe some more additional time?
Petrual Emill; got milk?
Oh yeah, specking of milk I was born on the holiday (milk day is January 11th)… so I’m a pretty rare person.
Experienced Forum User, Published Author, Player
(14)
🇺🇸 United States
Joined: 11/13/2022
Posts: 36
Location: 🇺🇸 United States
nymx wrote:
I put this submission on delay at your request, maybe a month ago. it's been a while and I'm wondering if you have any improvements. If you are not making the progress that you figured by now, we can just cancel this , in lieu of a future submission.
Yeah. I’ve decided to drop this game as I’m stuck on the first board and can’t seem to get past it what so ever. So no further improvements will come.
Petrual Emill; got milk?
Oh yeah, specking of milk I was born on the holiday (milk day is January 11th)… so I’m a pretty rare person.
Experienced Forum User, Published Author, Player
(14)
🇺🇸 United States
Joined: 11/13/2022
Posts: 36
Location: 🇺🇸 United States
nymx wrote:
What is the current status on the effort to demonstrate the improvements mentioned by Walgrey?
Okay I haven’t done too much to this game yet, but after watching the hacked TAS of this I’ve gotten a few ideas of how to optimize it a bit further. Seems like a v3 will be in the works soon!
Depending on how much free time I have…
Also atm I’m gonna re-encode the unhacked TAS to include the end credits this time.
Petrual Emill; got milk?
Oh yeah, specking of milk I was born on the holiday (milk day is January 11th)… so I’m a pretty rare person.
Experienced Forum User, Published Author, Player
(14)
🇺🇸 United States
Joined: 11/13/2022
Posts: 36
Location: 🇺🇸 United States
Walgrey wrote:
I'd highly recommend joining the Mario Party speedrunning server. There are quite a lot of people who are knowledgeable in Mario Party 9, and they can probably help you with what you need: https://discord.com/invite/0hwVJZ5USYlGkZre
Ok. but I have mentioned before people in this server are figuring out how the dice rolls here work. I've also mentioned this is my second attempt meaning my previous one was 1:59:43.09. If someone manages to beat my current record here, I'll just consider this one a failure in favor of the new submission.
But anyway, a team is working on this TAS so I guess the judge can reject this one for the faster TAS.
Petrual Emill; got milk?
Oh yeah, specking of milk I was born on the holiday (milk day is January 11th)… so I’m a pretty rare person.
Experienced Forum User, Published Author, Player
(14)
🇺🇸 United States
Joined: 11/13/2022
Posts: 36
Location: 🇺🇸 United States
Walgrey wrote:
xTASz wrote:
Depending on a specific board with this kind of manipulation (Bom-omb Factory, Blooper Beach and/or Magma Mine) I think you could get them to change a direction but it’s pretty intense
I'm talking about manipulating CPU dice rolls. As far as I can tell, I don't see you using this to its fullest potential in the run.
Edit: Could you also take a look at the screenshot I posted above?
That’s because CPU dice roll manipulation in this game isn’t like how manipulation is done in the older Mario Party games which was simply to pause and pause, in this game backing out of your dice multiple times changes their dice blocks around a lot but very often I’m unable to reach the nearest neighbor of what I want them to roll for and causes me more time just for them to use their special dice or roll what’s needed.
Petrual Emill; got milk?
Oh yeah, specking of milk I was born on the holiday (milk day is January 11th)… so I’m a pretty rare person.
Experienced Forum User, Published Author, Player
(14)
🇺🇸 United States
Joined: 11/13/2022
Posts: 36
Location: 🇺🇸 United States
Walgrey wrote:
But you could still use this in other places, correct?
Depending on a specific board with this kind of manipulation (Bom-omb Factory, Blooper Beach and/or Magma Mine) I think you could get them to change a direction but it’s pretty intense
Because in this minigame you can’t easily push the COMs into the spiked balls (now I’ve actually played this minigame before switching to the Goomba Bowling because I felt like I needed one that is faster than this one).
Petrual Emill; got milk?
Oh yeah, specking of milk I was born on the holiday (milk day is January 11th)… so I’m a pretty rare person.
Experienced Forum User, Published Author, Player
(14)
🇺🇸 United States
Joined: 11/13/2022
Posts: 36
Location: 🇺🇸 United States
Walgrey wrote:
xTASz wrote:
55:06: I can’t manipulate anything around this part, as Luigi was the Captain here and Yoshi refuses to roll at least a 6 that could easily send Shy Guy towards the boss gate. Tegron also asked me the same question to this part and I briefly answered it with the nearest number I wanted not appearing on my dice. However, your point here still stands depending on how much faster you can get to the boss gate through the bottom (which of course the player will need to be Captain to do so).
No, no. It’s not about the dice rolls here, it’s the arrows that I can’t manipulate. Yeah I mentioned the “dice rolls” earlier, but they aren’t always the problem unless if it becomes so obscure that continuing it is just not worth it.
Petrual Emill; got milk?
Oh yeah, specking of milk I was born on the holiday (milk day is January 11th)… so I’m a pretty rare person.
Experienced Forum User, Published Author, Player
(14)
🇺🇸 United States
Joined: 11/13/2022
Posts: 36
Location: 🇺🇸 United States
nymx wrote:
Thanks goes out to DrD2k9 for performing a sync test on this one. It is clear that I am having issues with some of my machines. I'm still stuck in the past, using older operating systems.
That’s unfortunate. Can your system run Dolphin 5.0 [LUA Core] at all? Perhaps the sync will on the original 5.0 by some chance?
Petrual Emill; got milk?
Oh yeah, specking of milk I was born on the holiday (milk day is January 11th)… so I’m a pretty rare person.
Experienced Forum User, Published Author, Player
(14)
🇺🇸 United States
Joined: 11/13/2022
Posts: 36
Location: 🇺🇸 United States
Walgrey wrote:
I'm afraid you're not understanding my question here. I'm referring to the in-game time at which you finished the minigame, which was 11.15.
Oh, you’re saying the in-game time to the jigsaw puzzle Minigame that I played? Yeah I guess this had something to do with desync issues
Walgrey wrote:
You're essentially forcing yourself to do extra turns just to make up for that.
Well I was trying to copying what I’ve watched from the niconico 999 video (https://youtu.be/kkOLf9MssVc?si=sqhb-kZtCeDMKPGt and https://sp.nicovideo.jp/watch/sm37037523?redirected=1) in which I tried to match exactly that just by manipulating for well all ten Koopa’s or so. But I see now that I need to make an improvement towards the bowser boss and make a little bit of sense to what I could’ve done here
EDIT: Oh wait, after rewatching the 5:22 part, I realized that Toad here was actually going to hit a spiked shell instead of a bullet bill which I was preventing.
Petrual Emill; got milk?
Oh yeah, specking of milk I was born on the holiday (milk day is January 11th)… so I’m a pretty rare person.
Experienced Forum User, Published Author, Player
(14)
🇺🇸 United States
Joined: 11/13/2022
Posts: 36
Location: 🇺🇸 United States
Walgrey wrote:
As a former Mario Party 9 speedrunner of the game, I have some questions about the optimization about this TAS. Keep in mind, I skimmed through most of the board sections, and my knowledge of specifically TASing this game is limited.
5:22 - Why did you skip that Bullet Bill box and then run back to it? Wouldn't the next spawn spawn quicker if you just let Toad hit the block on the left and hit the remaining one?
15:30 - RTA runners like to get a 1-on-1 match for Blooper Beach because it significantly reduces the time spent in the Captain Event. Why was this not attempted?
41:06 - There is a long pause here. This could be to manipulate a minigame, but the minigame chosen was Goomba Bowling. Couldn't a faster minigame be chosen?
53:27 - Unsure if it's possible to manipulate the dice rolls in the Unlucky Spaces, but why roll a 6 when a 1 would be faster?
55:06 - On that note, why go on the unlucky path at all? Stopping on 3 Unlucky Spaces doesn't seem fast. The bottom route has a Shuffle Space and a Spin Space, so it's not like it's completely impossible to avoid a minigame there.
59:30 - I believe it would be faster if no one salvaged any Mini Stars.
1:04:30 - You can skip this second Captain Event by going on the top route. Not sure why this is done, considering it would save minutes.
1:07:44 - The Lucky Island is not skipped.
1:32:44 - You can go through the Superstar announcements quicker if everyone ties for first place. While it would be difficult, it would at least be nice to see some attempt being made to tie players together.
1:32:45 - On that note, winning first place on every board isn't necessary and limits your options on saving time.
1:37:51 - Wouldn't it be faster if everyone just always rolled a 6? More frequent attacks, less time spent on misses.
1:47:34 - The Cyberscore record for Jigsaw Jumble is 9.76 seconds. Why is this 1.39 seconds off?
1:49:02 - Having blocks in the air while the timer hits 0 wastes time.
1:49:03 - Having any blocks be on the Bowser face is bad because it means that Bowser is spending time shooting fireballs and because less damage is being dealt to him. Ideally, no faces should be on Bowser during this boss fight.
1:52:00 - Having the Bowser mini-boss dice bounce on the ground wastes time.
1:52:09 - I recall that in the past, I timed how long every mini-boss dice segment takes, and if I remember correctly, Bowser is the fastest. Why doesn't this face appear more often?
All in all, I think this TAS still has room for improvement. I am glad you gave it a try though, as TASing this game is definitely a challenge.
Okay. Firstly thanks for pointing all of these out. I want to mention that some of the parts in this TAS didn’t give me the correct rotation as I wanted for the later boards after Toad Road, but I’ll go and explain everything that I’ve been brutally crushed on during the making of this whole TAS:
5:22: Toad was going to hit a 1 bullet bill here, which I prevented from happening as hitting just the 1 bullet bill will slowly decease Lakitu’s health bar, and I often used Toad’s mushroom head as a boost up to hit the bullet bill boxes faster.
15:30: This probably happens when you play against two CPUs on the Toad Road map. But Blooper Beach’s 1v1 is needed for this TAS.
41:06: This part really gave me a struggle because Kamek uses his slow dice for a pointless Blue space instead of a lucky one which would’ve saved me quite a few frames here. Although when I was going through the mansion I didn’t want the COMs to go inside the mansion rooms and kept everything just straight forward. If the game had Kamek go after Peach I would likely have a big time saving in and out of the mansion.
53:27: Because of this map’s way of manipulation of dice blocks, I noticed a 1 is too far away from nearest number of the dice block but it would’ve been smoother if a 2 was given to me instead of a 6.
55:06: I can’t manipulate anything around this part, as Luigi was the Captain here and Yoshi refuses to roll at least a 6 that could easily send Shy Guy towards the boss gate. Tegron also asked me the same question to this part and I briefly answered it with the nearest number I wanted not appearing on my dice. However, your point here still stands depending on how much faster you can get to the boss gate through the bottom (which of course the player will need to be Captain to do so).
59:30: Could I make it so Shy Guy wouldn’t salvage any of the Ministars at least?
1:04:30: As mentioned above, I can’t manipulate the directions of the COMs choices. I was sorta forced to go Down instead of Up because it was what Yoshi was choosing for me to do. Which I see why this one needs to be a 1v1 board only.
1:07:44: And here once again Yoshi makes another poor decision and doesn’t use his 1-to-10 dice block which literally could save me about 57 or so seconds. Otherwise trying to get him to roll a 6 here was far impossible enough to do, I often get 3s and 1s and 5s on the part so I had to stick with it.
1:32:44: Yes, but then maybe if I tied with Mario I could definitely gain a little bit of speed.
1:32:45: While I know that I really didn’t have to win on every board, I still sorta want Koopa as a superstar on every board just to show what the TAS player can really establish by briefly winning everytime.
1:37:51: No, this is not considered faster because there’s a pattern of which ends the Bowser Jr fight immediately. 15 to 15 to 9 for 3 player and 15 to 15 to 12 for 4 player, while 2 player it is 15 to 15 to 6.
1:47:34: I tried to skip a Minigame here but fell into some major problems with getting Kamek to roll a 6 to land on the Lucky space that gives the player ministars. So I obviously knew that this Minigame was gonna be suboptimal because of the wait of one COM beating another COM but was forced to play it anyway by RNG.
1:49:02: keeping the blocks on the ground towards only Koopa’s icon was far difficult to do. I spent hours doing trial and error for this until I managed to figure it out entirely to get on to Koopa’s face.
1:49:03: It is really bad when Kamek or Shy Guy also throw them on a Bowser face, too. Sometimes I will try to hit possibly 9 Koopa’s but I want to avoid the Kamek and Shy Guy faces too.
1:52:00: Oh maybe I didn’t realize that it was bouncing in the first place.
1:52:09: Bowser’s face is the fastest? I guess I haven’t thought of that when I was struggling to keep Koopa in front of every dice block.
Wow this was a lot to go over. An improvement may be possible to make, but there are people in the Mario Party speedrunning server who are still figuring out the dice blocks in this game. I’ll probably not return to this game as of today it is now 14 years old and this is my second attempt at TASing this and Mario Party 5. I’d appreciate it if someone else would like to give my attempt a try and improve from this game’s brutal rotating madness (don’t try it if it overwhelms you too much unlike me). Although, this TAS still beats RTA by exactly under 2 hours and has 5 yeses by voting, if you think this shouldn’t be published I’d feel ashamed but glad to have tried to beaten RTA with some sub-optimal movements just like I was thrown at on my Mario Party 5 submission, causing that one to be delayed and if the judge wants to delay this one too, I’d be happy to say yes otherwise. Reworking on this TAS is certainly a focus that I can work out for.
Petrual Emill; got milk?
Oh yeah, specking of milk I was born on the holiday (milk day is January 11th)… so I’m a pretty rare person.
Experienced Forum User, Published Author, Player
(14)
🇺🇸 United States
Joined: 11/13/2022
Posts: 36
Location: 🇺🇸 United States
Didn’t watch this yet as I’m not too familiar with a animal crossing speedrun or TAS, but will have to say that the quality overall for the submission seems to be lacking a little bit. Not to be crude, but I’m pretty the animal crossing game I’ve played was that darn Mario Party ripoff on the Wii U. Meh vote.
Petrual Emill; got milk?
Oh yeah, specking of milk I was born on the holiday (milk day is January 11th)… so I’m a pretty rare person.
Experienced Forum User, Published Author, Player
(14)
🇺🇸 United States
Joined: 11/13/2022
Posts: 36
Location: 🇺🇸 United States
andypanther wrote:
So the only thing the TAS has going for it is that you get to watch the RNG manipulation unfold. This is sometimes fun, like that one time when the TAS deliberately steps on a bowser square at the very end to quickly get mini stars from the CPUs. But most of the time, even the RNG manipulation doesn't look all that interesting and it's made worse by the fact that the setup is sometimes painful to watch (very long waiting periods or going in and out of a dialogue, making the camera zoom back and forth in an irritating way).
Sorry that you didn’t enjoy watching this TAS. I know the slow and timing of this game can be a lot of pain, partly when it forces me to manipulate the dice rolls for that long period of time (while being unable to reach the correct dice number in TAS timing). Between this and Wii Party, it is clear that the developers for Mario Party 9 maybe rushed development/coding to the COM pathfinding and dice rolls (especially at the “almost there!” flag where they select “Yes” to literally more bowser spaces), but my concern for this would be how much longer it takes for me to not get a immediate Minigame box (as they slow the player down). Perhaps skipping the minigames altogether was speedy, but also saves a lot of frames and time to how (less) longer it takes for a COM to hit the correct number just the first time they’re given a turn. I plan on doing the time attack mode which plays the minigames with a single player.
Petrual Emill; got milk?
Oh yeah, specking of milk I was born on the holiday (milk day is January 11th)… so I’m a pretty rare person.
Experienced Forum User, Published Author, Player
(14)
🇺🇸 United States
Joined: 11/13/2022
Posts: 36
Location: 🇺🇸 United States
Bigbass wrote:
xTASz wrote:
Ya know, MarioSonicTASer, if you want I could teach you a little bit more of optimizing your TAS for a more interesting audience review. For example, my Mario Party 9 TAS has a audience review of 100 users. Not too big on the discussion thread, but more likely on a value that matches my submissions standards.
"audience review"? You mean, 100 views, which isn't a useful metric of quality. That's not a unique quantity of people either, just that the video was viewed/opened approximately 100 times. Your encode has received 2 likes so far, but otherwise no one has "reviewed" it.
More importantly, whether intentional or not, the way you've worded your post comes off as exceptionally rude, as if you're more interested in building yourself up than the person you're talking to. It'd be better if you could offer specific, constructive, feedback on their submitted work. If you have some knowledge that is applicable to this game and could help the author improve, then feel free to offer it up in this thread.
You’re right. suppose I should’ve mentioned that it was 100 views in the YouTube video and not the submission (I was thinking the wrong thing while writing this post, but the author bugged me to review this submission/video). I apologize for the mean wording and I’ll just say that from now on I’ll just give feedback to the author and not telling them about the false information that has anything to do with this topic. Sorry again, MarioSonicTASER.
Petrual Emill; got milk?
Oh yeah, specking of milk I was born on the holiday (milk day is January 11th)… so I’m a pretty rare person.
Experienced Forum User, Published Author, Player
(14)
🇺🇸 United States
Joined: 11/13/2022
Posts: 36
Location: 🇺🇸 United States
Bigbass wrote:
I would personally recommend writing some brute-forcing scripts if possible, and/or performing some deeper analysis of the game. I strongly suspect that the computer difficulty doesn't make this significant of a difference between easy and hard, particularly in a TASing environment. It may take a lot of work, but if hahaPhD's work is anything to go by, I think you could shave at least 10 minutes off this run.
Mini-games are incredibly slow. While I noticed that you tried to manipulate the selection to minimize the time, avoiding mini-games altogether would be far faster. This is where scripting could be very helpful, but it's not absolutely required. Potentially you could use it across multiple turns although that's a much more complex challenge. In any case, identifying key locations in memory could offer some useful advantages too, such as immediately knowing the dice rolls, or any other logic related to RNG.
Instead, I'd suggest making more use of the orb'ed spaces around the board. Hammer-bros for example eat 5 coins each time. Piranha plants eat half of their current total, making them especially good during the first couple turns. If it's a red space, then they also lose 3 additional coins. You may even pickup orbs from the vending machine that you can strategically place on the board. That'd be a fairly complex maneuver but it could save a lot of time, and offers more routing flexibility. Using the pipe spaces may also prove useful to move the koopas to a better position. Similar to the piranha, Donkey Kong also takes half their coins, but will give coins if you land on him.
While vs-spaces effectively eat 15 total coins (while all koopas are alive), three piranhas against fresh koopas will eat 30 in significantly less time than a single mini-game (30sec vs 46sec for the fire mini-game, or 55sec for a 2-round rock-paper-scissors).
Additionally, the orb vendor has a chance of giving you the paratroopa orb. Using it on yourself costs 10 coins, but allows you to take up to 25 coins from each koopa in a single turn! That can easily wipe all of them out at once.
Edit: Forgot to add this. During the bowser fight at the end of the game, there are some improvements you could make here as well. As shown in hahaPhD's TAS, it's possible to pick up the robots and throw them while jumping, thus destroying multiple robots at once. It's a challenging task to track and route this many enemies but it's worth the effort.
I think I suggested this in Discord already but, during the next round with bowser there is one instance where bowser charges at you. With some different movement, it's definitely possible to eliminate this time loss.
I'm not 100% sure about how this works, but during the final round it's possible for bowser to throw 4 orbs back-to-back (instead of only 2) before breathing any fire. I think it's something to do with getting hit by one of the first two orbs, or possibly where you're standing when they're thrown. Minimizing the distance each orb is away from the edge of the field can also help minimize the time it takes to pick up each once they're activated.
Yes. This is why I'm requesting a delay for this submission to fix these issues. Thanks for your heads up.
Petrual Emill; got milk?
Oh yeah, specking of milk I was born on the holiday (milk day is January 11th)… so I’m a pretty rare person.