Iwer Sonsch
[Partially outdated] [Theory TAS] Super Mario Fantasy: C7 route idea for 100 Star (6 stars)
updated
There are 32 different stars in this level, split evenly between the 4 colour. Each colour uses its own save file - A for red, B for blue, C for green, and D for yellow.
Alternate showcase: youtu.be/YLXJ7aqGKAE
Known issues:
- Water is only visible while the red area is active
- The vanish cap switch might not be pressed in all areas
- Dark Castle / Josh's Constant (3 dart start, Perma-Spike ending)
- Rake (shown here)
- Cornfield / No Harvest
- Another Brick (very lenient)
- Off the Coast (very lenient)
and likely other advanced and intermediate maps that I haven't tried.
Weak on:
- Multiple lanes
- Rounds 100 and 119
so I originally wanted Off the Coast to be the quality cutoff for calling a Chimps strat "viable", but Off the Coast turned out to be so lenient and easy that I decided to go for Rake instead.
I'm still not 100% sure what map I should use to showcase generic and versatile Chimps strats, but Rake seems like a solid candidate as it requires defending more than one lane.
The earlygame is a bit more lenient with a Camo village and Jungle's Bounty, but I decided to be stingy and go for the discount crosspath instead.
The lategame has a couple of options but I went for Dark Knight to ensure defense against DDTs.
Came back while katze was on his successful attempt (twitch.tv/videos/2037015200?t=00h10m26s ).
My favourite full kaizo star by a good margin. Every jump is fair and not too precise, but every jump contributes meaningfully to the difficulty of the star.
This makes it by far the densest list star that I have done, reaching tier 3 difficulty with only 6-7 minutes of jumps.
I could call the star a bit linear, with a clearly most efficient route, but there are tons of slower options as well.
The main thing that I noticed at the start of the grind was that I had to memorize the steepnesses of different slopes with the same texture, and know which I could stand on, which would give me steep jumps etc.
I didn't keep track of the grind length but I guess it was probably about 15-ish hours.
Shoutouts to MorningStorm for delivering the best of the best once again, and to Lugmillord for providing the baseline model.
Satori Maiden ~ 3rd Eye: youtube.com/watch?v=WjpVUq7S0qY
Exiting the world and reentering removes this mode. Teleporting doesn't. Apparently the hotkey to exit this mode is between del and image^.
This is the same amount of in-game time as an 11:33 on fast forward (assuming no lag). Makes the fast forward record look quite slow, that's because it is.
I later got a segmented 5:42:xx on permanent 0.25x speed, which is equivalent to a 1:25:xx on normal speed or 8:3x on fast forward. But that used a different, micro-intensive strategy where I went for multiple additional settlements for extra taxes.
A TAS might be able to beat this level in 7 minutes or less (on fast forward). But nobody will ever make one.
For scale: Building the Emperor's cathedral (which unlocks at 2000 Noblemen) takes 3 hours of in-game time, or 18 minutes on fast forward. Unless you provide resources to Ibn Al Hakim, which speeds it up but I have no idea by how much.
- Starting cash changed from 50,000 to 17,750
- Difficulty changed from Easy to Medium
- End round changed from 40 to 99
- Chinook limit changed from unlimited to 10
- Bloon Master Alchemist used. His shrinking potion doesn't pop any bloons, it only transforms them - at least according to the tooltip.
- To simulate an arbitrarily long map (e.g. "No Escape"), I lowered the ability recharge time to 5% and the bloon speed to 50%. I previously did this challenge with 50000 starting cash on End of the Road, but it was painfully slow so I cut the run after surviving round 63.
Note that I did not use Benjamin or the x4x Glue Gunner, as those can actually cause the Bloon Master Alchemist to accidentally pop bloons. I could have allowed a 032 or even a 134 Glue Gunner, but slowing the bloons by 75% would have made the level (yes, even Blons) needlessly long.
Instructions:
- Start walking forward.
- Using a pause buffer, turn right for 2 frames.
- Using jumpkicks to conserve your angle, make your way to the back corner of the doorframe.
- Turn the camera back to where it started.
- Jump.
- As you're landing, start holding right (4 frame window).
- Doublejump and pause on the 6th frame *[frame-perfect].*
- Hold forward, then unpause with A.
- Pause again on the frame before landing *[frame-perfect].*
- Hold left, then unpause and immediately press C-left and A *[frame-perfect].*
- Pause on the last frame before hitting the corner *[frame-perfect].*
- Hold slightly upleft *[roughly 5° range],* then press A (3+ frame window).
- Release the joystick and press C-right (not precise).
- At the peak of your wallkick, flick right and pause (not precise).
- Hold forward, then unpause and press C-right and A (3+ frame window).
- Just before hitting the ground, groundpound (3+ frame window, 4+ if your other inputs were excellent).
0:00 Full speed
0:18 Half speed
I felt that the map generations in T90's series might be a little too inconsistent, with e.g. the woodlines blocking off half the map and leaving the other half without wood, so I tried to build a more standard map generation, at the cost of it no longer being a random map.
While this is a 4-player map, it shouldn't be too hard to change into a 2-player map by replacing the other 2 starting TCs with woodlines, replacing their 2 Stone with 1 Gold, replacing their 2 boar with 4 deer, and deleting the rest of their starting resources.
I recommend a pop limit of 25 for 4-player games, and 50 for 2-player games on this map size.
Unfortunately, I can't upload the map because I only own Age of Conquerors. If anyone wants to rebuild the map 1:1, I can try to assist with screenshots and descriptions.
Starting from the north corner, let's go southwest and count, for each row, how many tiles southeast something extends.
First is the lake (shallow water), extending 7, 7, 7, 6, 6, 6, 6 tiles.
Then comes the shore (ice).
After that, we have 1 row of space.
Then 3, 5, 6, 6, 5, 2, 1 trees, for 28 total.
Another 3 rows of space, then 2 tiles of stone.
4 rows of space, then 1,1 tile of berries.
2 rows of space. Ice.
And finally 6, 7, 7, 7, 7, 7, 7 tiles of water.
The town center is 7 tiles away from the west pond's ice, and 4 tiles away from the edge of the map.
The gold is 2 tiles away from the TC's east corner.
Counting from west to east, the woodline on the west pond has 3, 4, 3, 4, 3, 2, 1 trees, for 20 total.
The woodline in the center has, counting from west to east, 3, 4, 5, 4, 5, 4, 3 trees, for 28 total.
The west and east ponds each have 2 shorefish, 2 marlin on the diagonals, and 2 tuna on the sides. The north and south pond each have 2 shorefish and 3 marlin.
I hope that the animal and straggler positions can be replicated from the video.
From the north corner to the center, the tiles are:
Marlin, Marlin, water, Marlin, Marlin, water, ice,
Gold, plains, hill, hill+Stone,
plains, plains, Relic, plains, plains, Tree, Tree.
I'm very confused as to how I got the Uncanny Clicker achievement.
I'm also confused as to how the click counter I got recommended desynced by about 500 clicks.
Bubzia's stream (twitch.tv/bubzia ) was playing in the background for date verification.
Unfortunately, that last point seems to have killed the framerate of Cookie Clicker. Somehow that didn't slow down the game.
And yeah. Incase you didn't hear the news, farm upgrade before cursor upgrade is faster if your CPS is no higher than like 10.5.
In any case, getting the farm upgrade before the cursor upgrade is optimal below 10c/s, the slower you are the more optimal it is. Thanks for the route and calculator @DEDERTEEE
sorry for no mic/cam, i was thinking of adding the mouse part of my input display but it didn't fit anywhere easily
also sorry for not showing that i have no mods, forgot about that part of the rules before running. anyway, heard it's a steam only rule so yeah. and i don't know any mods for like, getting better golden luck which i imagine the mods rule is there for
I'd say that this run is about a minute faster than my v1.1 WR (youtu.be/x8Ouixi5a_8 ) after taking the version differences into account, but v1.1 is harder because it has no dpad camera, slower domino platforms, and some object placements are less generous.
Previous v1.3 PB: youtu.be/jAalZK6_60g
I then went back to full runs of the category and got this run. The midgame wasn't as good as in my previous world record, but still good enough to stay on an almost even pace and cash in the timesave at the end.
I can recommend this category to anyone looking to run a highest-quality romhack that mostly focuses on forgiving, long-form movement without too many small platforms and sharp turns.
Upload failed when I got this initially and I never ended up beating it
Tala told me to try again
I didn't expect to be proud of this WR but here we go. This pace was absolutely ridiculous until C7 coins (it was on pace for an all-version PB despite v1.1 being about 4-8 minutes slower!), but the 5th Scriosadh battle turned out to be more difficult under this pressure than I thought.
This is my 4th run of this version to clear most of the game. My first run (youtu.be/-tNizqzfk-s ) had a bad C3 and B2 but was solid afterwards. My second and third run both pulled 7 minutes ahead out of C6 but lost all of it back by the time they reached B5, then fell apart for good.
Uploading a route cause there are some differences to v1.3.
Compared to my v1.3 route, I removed B1/SB reds, B4/K2 reds, and C13 reds+80c,
and replaced them with C6 bumper, C7 floating tables, C8 !-box tower, and C12 blue switch.
Compared to CB's v1.3 route, I removed C4 ultrablock, B1/SB reds, B4/K2 reds, C12 reds+80c, and C13 reds+80c,
and replaced them with C6 bumper, C7 floating tables, C8 !-box tower, C11 reverse tower, and C13 shell secrets.
I believe that this version is only about 5-10 minutes slower than v1.3, meaning that a WR difference of 40 minutes is a complete joke.
Quite a stable run throughout. Biggest mistakes were a death each in B6, C7, and C10, and falling down on C4 reds.
I'm glad that I finally understand how to do bob-omb clips more than 5% of the time.
Leaderboard: https://minesweeper.online/best-players (select "Arena", "Endurance", and "L6" in the dropdown)
Music by Sin X Nad
youtu.be/QpDHNl3NHoc Arrangement of Pure Furies ~ Whereabouts of the Heart from Touhou 15 LoLK (original track by ZUN)
youtu.be/6qKBLCZYUtg Arrangement of Locked Girl ~ The Girl's Secret Room from Touhou 6 EoSD (original track by ZUN)
Two days, two 8+ hour world records. This time on a leaderboard with 507 successful completions. To be fair, players have achieved much faster paces on the L8E and L6 endurance arenas, so this world record happened to be (and still remains) a rather feasible world record to beat.
If you are able to somewhat comfortably beat the L4 Speed arena, and if you have my level of stamina, you will have no trouble beating this world record unless someone else does it first.
You will need to have reached 500 trophies to attempt this arena, and tickets for 1 attempt are traded at the coin yield of around 400 Intermediate games. You will also need 200 honor points, and make sure to select "Elite" in the arena settings. But if you build enough skill (e.g. with Intermediate and Expert Endurance, where you can also compare your PBs to mine), then make sure to attempt this at a good daily form, one attempt should be more than enough.
I kind of had to do this since my first run had a kind of choppy recording that failed to process after uploading (maybe I closed the upload too early).
Nora's Challenge is a multi-game challenge run consisting of unlocking and beating the final level in Star Revenge 0, 1.5, 2a1, 2.5, 3.5, 4.5, 6.25, 5, 7, and 8.
These games form the current Star Revenge mainline, as they are the latest iterations of the integer Star Revenges, including both acts of Star Revenge 2.
Six players have previously completed the Scrub's Challenge, which consisted of Star Revenges 1-8, but without SR0 and with only one act of SR2.
Additionally, all except KingToad64 (youtu.be/1dX0VKuadmM ) ran on a mix of latest and original Star Revenges, rather than settling for one or the other.
SR0 Galaxy of Origins in 39'06 (+1'20)
SR1.5 Star Takeover Redone in 1°19'13 (+10'27)
SR2a1 To the Moon in 50'45 (+3'19)
SR2.5 Remnant of Doom in 1°45'32 (-0'45) PB
SR3.5 Vacation of Cursed Dreams in 46'52 (-0'46) Route WR
0:00 SR4.5 The Kedama Takeover Rewritten in 58'23 (+1'41)
9:55 SR6.25 Luigi's Adventure DX in 1°09'33 (+5'02)
1:20:50 SR5 Neo Blue Realm in 39'44 (3'46)
2:00:41 SR7 Park of Time in 54'09 (-4'21) PB
2:54:56 SR8 Scepter of Hope in 2°01'41 (+6'24)
I kind of had to do this since my first run had a kind of choppy recording that failed to process after uploading (maybe I closed the upload too early).
Six players have previously completed the Scrub's Challenge, which consisted of Star Revenges 1-8, but without SR0 and with only one act of SR2.
Additionally, all except KingToad64 (youtu.be/1dX0VKuadmM ) ran on a mix of latest and original Star Revenges, rather than settling for one or the other.
0:00 SR0 Galaxy of Origins in 39'06 (+1'20)
39:14 SR1.5 Star Takeover Redone in 1°19'13 (+10'27)
1:58:34 SR2a1 To the Moon in 50'45 (+3'19)
2:49:27 SR2.5 Remnant of Doom in 1°45'32 (-0'45) PB
4:35:07 SR3.5 Vacation of Cursed Dreams in 46'52 (-0'46) Route WR
5:23:22 SR4.5 The Kedama Takeover Rewritten in 58'23 (+1'41)
SR6.25 Luigi's Adventure DX in 1°09'33 (+5'02)
SR5 Neo Blue Realm in 39'44 (3'46)
SR7 Park of Time in 54'09 (-4'21) PB
SR8 Scepter of Hope in 2°01'41 (+6'24)
Did this run to see if OBS was going to output proper quality with my settings. Sadly it didn't. Gladly I know that now
0:00 SR4.5 in 1h04'25
1:04:35 SR6.25 in 1h15'32
2:21:27 SR5 in 38'19
3:03:59 SR7 in 58'30 [PB]
4:02:37 SR8 in 2h02'23
Nora's Challenge is a multi-game challenge run consisting of unlocking and beating the final level in Star Revenge 0, 1.5, 2a1, 2.5, 3.5, 4.5, 6.25, 5, 7, and 8.
These games form the current Star Revenge mainline, as they are the latest iterations of the integer Star Revenges, including both acts of Star Revenge 2.
Six players have previously completed the Scrub's Challenge, which consisted of Star Revenges 1-8, but without SR0 and with only one act of SR2.
Additionally, all except KingToad64 (youtu.be/1dX0VKuadmM ) ran on a mix of latest and original Star Revenges, rather than settling for one or the other.
I've been preparing for this run for about a month. It is far easier than 115 Star in Super Mario Treasure World (youtu.be/HGWChDPqJaY ), despite being longer. Since the stars are much easier than in SMTW, exhaustion takes longer to set in, and it is less punishing to endure.
This has been my longest run of SM64, and my second longest run overall (after my now obsoleted 100x Expert run in Minesweeper, https://minesweeper.online/endurance/2015994279 ).
I used a route that's closer to Mui's WR, so I made new splits. I did edit my old PB time onto the final split though
Tilted a bit in 2a1 c6 and 2.5 c10 but the rest went well.
Bad B2, otherwise alright
Without clipping through doors in the overworld, this category must obtain the first 4 stars and the key in the 5th Dream stage, and press the metal cap switch.
On the flipside, Nightmare Void is unavailable.
These two things makes the category quite different from 50 Star ISC (Intended Star Count), and a good deal slower.
not looking to shoot for a clean time on this, I just want to be able to play this at a somewhat solid level for... reasons.
Changed my route to get better golds on C4 and C7. That worked out. What didn't work out is Untriforce but I'm happy about the run overall.
Main things to improve are to command the archers more efficiently to infiltrate the gate faster, and to build the pharmacy sooner. This mission can clearly go below 26 minutes, maybe even below 25.
I'm sharing this with WR grinder @alenoremblem cause I think this build order has potential. Maybe the second conversion has to be cut out in order to reach the full potential of this route.
My splits were a mix between my previous PB (youtu.be/EhLexiTesYg ) and the run that died to persistent major lag (youtu.be/MeIzvml0F4M ).
Using PJ64 v3.0.1-N and Luna's DirectInput because they exist now.
I didn't really notice a difference except the lag feels more controllable (i.e. I had to ramp up the process priority and close my input display after a while, but after that I had zero issues anymore).
I found that buying hunt and wool is more efficient than taking a territory to gather them yourself. You can also generate a net profit from buying a bit of fish and grain.
Kestral would lower the required wares by around 17%, but I prefer Hakim because his wood discount lets me build a larger economy comfortably. Elias could get you through the early months without food buildings, but since the ultimate goal is to generate taxes, that doesn't seem like a very powerful bonus here.
- Complete the victory objective after causing it to appear on screen.
- Don't stop bandits from stealing resource carts without using at least 1 batallion of soldiers.
- Don't convert a batallion on the other side of an enemy gatehouse.
These additional rules lead to a different, slower route for missions 1, 2, 4, 5, 10, 14, and 16.
- Complete the victory objective after causing it to appear on screen.
- Don't stop bandits from stealing resource carts without using at least 1 batallion of soldiers.
- Don't convert a batallion on the other side of an enemy gatehouse.
These additional rules lead to a different, slower route for missions 1, 2, 4, 5, 10, 14, and 16.
- Complete the victory objective after causing it to appear on screen.
- Don't stop bandits from stealing resource carts without using at least 1 batallion of soldiers.
- Don't convert a batallion on the other side of an enemy gatehouse.
These additional rules lead to a different, slower route for missions 1, 2, 4, 5, 10, 14, and 16.
The world unrecord sits at 2:51 by Sir (seconds unknown).
Seemed to be the fastest time with video, although SirSirSirSirSirSirSirSir supposedly has a 2:51.
Still very improvable. Even a 2:51 would be very improvable if StarrlightSims' playthrough splice is anything to go off of.
Went for this after my 100x Expert Minesweeper run yesterday (https://minesweeper.online/endurance/2015994279 ), which showed me that I am able to play for a long time without getting exhausted as long as I don't put too much strain into it.
Extremely low-effort run. No practice, around 200 deaths, 4 stars replaced after failing each of them for like 10 minutes.
I expected this to maybe take 10-12 hours if it went badly.
Glad I finished this and can upload it in one video.
Third attempt to the pipe, first attempt to the QSLGs. A massive fluke, so that it only took me around 5 hours total including my estimated experience with the relevant parts of the stage from other stars.
I don't know whether this is easier or harder than Bramble Castle reds. I want to say slightly easier, but it's hard to tell how difficult everything outside the QSLGs and the green star platforms really is for me.
Well, it was easy enough for me to keep my nerves together and do the jumps. Hard enough for me to die a ton in practice where the stakes were low, or to die on the first 3 minutes which I should be consistent at.
this blj is impossible on 1.8 so use 1.6