#Speedrunning Tips and Tricks
216 messages ยท Page 1 of 1 (latest)
Currently i am working on migrating my calculation sheet to a google doc, so i can make it public to use
but that is TBD
and i will post a treasure 2.0 speedrun explanation after that
nice
ooh nice ive been thinking about speedrunning the game to like 1mil or 500k something like that
1mil coins is a speedrun
Though the 2.0+version makes it take forever
i always wondered what OP meant
OverPowered
yea
I thought it let us know who the Old People were
lmao
go fast ๐
dont waste time
(hehehe)
Just going fast won't cut it :S
Go faster then
so are there any categories for this? i checked speedrun.com but it doesnt have any results for hydroneer
i always thought it was Outright Pretentious
yes
its on bridgepour.com, hydoroneer's wiki/website
oh okay sweet, ive been working on some techniques for gettin to 500k and 1mil
OHOHO theres no 2.0.8 runs yet
1MIL is cursed in 2.0 compared to 1.6 and 1.7; it takes like 2-3 hours at least to even get 100K
so no one plays it
same with treasure%, until i found a new route; 2.0+ treasure % without glitches or without that specific glitch would take like 90 mins at least
smithing% is just get a 2nd person or tough luck
after the holiday i want to work on a more solo-able run/strategy
you wouldnt need a second person you can just use your feet to control the second character smh
as i said i am working on it ๐
i have done a solo smithing% run that with me controlling both characters didnt record it though just did it as a proof of concept
yeah ive done a few test runs for 1 mil and my record so far is like 10 hours i think
basically just had the second character be using the brush with the main character doing everything else
as i understand it timing is, as of 1.7,
from when you move the character
till the frame you see the thing you need. 1mil coins; smithed item; treasure chest moving
in 1.6 timing was a bit different, we might have to retime the whole board but IMO the old runs need to be purged
is there a way to see how much time you have in one save?
not really, no
btw what is the way to get out of bounds near icehelm that you mentioned in your message for the bounty
you go up to a stone and walk into the corner
and you just.. go
this one
Hello everyone!
i am finally done with the spreadsheet.
https://docs.google.com/spreadsheets/d/1G7_Aicxz_51_AAF4pV1i3jY40dDlQj3WX9sBT86M4ts/edit?usp=sharing
I need to make it a bit more readable and edit some names probably.
if you want to help me edit it, feel free to DM me and i MAY add you
prices
Name,Price,Name,Price,Name,Price,Name,Price
Agile Wooden Cart,740c,Small Furnace,86c,Dull Pickaxe,122c,My First Fishing Pole,52c
Brush,2c,Crucible,18c,Dull Shovel,22c,Wonky Fishing Pole,-
Bucket,4c,Casting Mold,42c,Common Pickaxe,742c,'ol' slippy grip' Fishing Pole,-
Centraliser Hook,462c...
Speedrunning Tips and Tricks
Hydroneer 2.2 will introduce a new tutorial area, guiding players through the basics of the game!
New speedrun for 2.2?
Speedrunner, in a few weeks:
"If you grab this pan that has both water and dirt inside, you can press it up against this wall. Save/load, it will float there. Quickly, grab your brush and spam LMB while running/crouching at the wall to clip past the conveyor tutorial, shaving off a whole 7-8 seconds! (depending on main menu time)"
Ahaha
Well I'm interested to see how people do it. We've hardcoded in things the player has to do, so glitching your way through it might be tricky
challenge accepted
tutorial%
You're on!
Speedrunning getting a automation system? preferably no glitches
not really what this forum is about,
sooo i found out you can do this in hydroneer and im wondering if you guys can experiment with things so we can find a glitch into the tresure chamber
cart glitch in the rock face
berd
#wetried
it could be barely possible with the pallet with perfect placement
oh wow i ddint think anybody would even try this
this is like 2 days ago, we just had the idea because we were discussing 2.0 treasure%
oh ok
there is a couple things that i did find using the truck but both are really rare and random
one is having a truck on the roof and somehow getting it upside down
block both doors using blocks, then enter/exit should put you on the invis roof of the cave
the other was uh
we were trying to glitch the truck through the terrain
only once it clipped into the cave but not into the treasure part
glitch used here: block in hand glitch (then drive the truck)
to do that you need to click really fast (reccommend auto clicker) and press E a couple times
i use a built in mouse button or OP autoclicker
so the old variant of the bug aint gonna work. i allso tried to make it better but the items dont behave like back then (thank god). if i break trough i will ping back. (i did come close so maby in a few days...)
๐
Is anyone doing speed runs? If so what categories are there already?there isnt anything on Speedrun.com for this game yet
Speedrun.com won't accept this game because it's "a sandbox with no clear ending goal"
So we keep track of it on the discord
Ahh I see, where could I see the runs/standings? Cause Ive been playing around with an all properties glitchless run (essentially a race to 22970 coins) and my runs have been around 1h20m
I was wondering if others have done a similar runs, or just generally how people go about getting lots of coins fast
They're on the wiki
The issue with most speed runs is that it's heavily biased by RNG
The only real ones are crafting%, and treasure%.
1 mil coins just isn't viable anymore in 2.0
Tell that to our cooked footwear
It isn't viable via mining then I'll say
Yea it heavily based in RNG but lots of other games are too
Ok so i am going to go through the current speedruns and see what kind of resources we need.
this way we can clean up backtracking and information
Categories of speedrunning
1million coins%
1.7 - current is 52minutes but 40 minutes is speculated to be possible. Need to work on a strategy calculator. TBD
2.0+ - current uses boot cooking which is now deleted; so any strategy will be record. my best guess is it will take like 2 hours. TBD
treasure%
1.7 - current is 29.7 seconds (unverified). this category is dead by all means as i dont see a major improvement
2.0+ - current is 20min but this uses a fixed bug. in 2.1 there was an even stronger bug but both were fixed in 2.2.
since you need to buy 1 or 2 plots and also mine to the relics i estimate the time of completion to be 35min with 4min digging time & movement
smithing%
1.7 - no runs ; strategy TBD
2.0+ - current is 3:48. oh boy multiplayer ruined this run. the strategy is pretty clear just try to estimate when you can afford the necessary stuff really
tutorial%
2.0+ - current is 1:58.7. the strategy is pretty solid and i've not found new skips. 1:45 is possible and maybe lower
plot%
1.7 - no runs ; strategy TBD
2.0+ - current is 51:18. it's probably possible to do in 45mins but this is optimistic
hydrosaur%
DLC - current record is 25:05, sub 20 is surely possible, 16:xx is theorised. am working to route currently
untapped potential:
Bastion
- craft a drill %
- farm a crop %
- any quest %
- all relics % *
DLC
- all relics %
- plot%
- 1m coins %
- water system %
- craft a drill %
- new glade %
- MOT's island time trial% AKA circuit%
not happening probably ever:
- kings quest % โ ๏ธ
- all achievements % โ ๏ธ โ ๏ธ
holds sprint button harder
hydrosaur %
old bone route
if you have luck for the mist bones to not spawn which is roughly 1/3 chance then the estimated best time is 17.40 - 18:00
hypotesized new bone route, but going up the volcano turned out harder than planned
still relies on 1/3 no mist bones
ngl I would like to write algorithms to find the fastest possible route (meaning all bones would have to spawn at the correct location) and the route with the fastest average time. I doubt i have time for that though
i mean if we are going to go for "perfect" rng then that is so vanishingly unlikely like on the order of 1 / 10^15 or less
as for just using the following sets we could craft a theoretical best
"skip 0" - all 65 bones so no luck
"skip 1" - skip either bottom mist or volcano center - 56% roughly
"skip 2" - skip either both in the mist or both on the volcano - 33% roughly
skip more, 3>17.8%; 4>9.8%; 5>5.4%
"36 perfect" - really handwavy @~ 8.9383853528ร10โ19 or 1 in 1.1 quintillion
since we have to start at H and end at M i adjusted a TP solver route until i got the lowest i could
https://www.lancaster.ac.uk/fas/psych/software/TSP/TSP.html
it's pretty sure on the shattered field order; that is solved for all we care about
i know it's basically impossible to have the perfect run, it just would be funny to do the algorithm.
probably something like this
but its goingto do something else as it wont consider terrain especially around the 5 on the ice block island
i thought a bit about "the beste route" and realized it is very complicated. Cause you don't have to visit empty spots, you just have to see them, to know they are empty.
oh btw i thought about skipping the one above M. It looks like including it takes quite some time.
yeah but respawn is favorable and its less out of the way than either mist one
true, unstuck is very favorable there
when going up that road you can see if its spawned, from either side
if you go from right to left you'll go a bit out of the way to spot it
left to right it's the last bone so if you counted you'll skip it or its a wash due to mist probably
anyway if you have more experience with weighted graphs and searching routes please do say it i got ideas involving only checking the 5 closest neighbors
so you wont get a "lol go from lumberton harbor to the barrens harbor"
i don't really have experience on that, but i think i know enough to write the algorithms i Was talking about.
there's certainly alot of research on it
https://www.cs.cmu.edu/afs/cs/academic/class/15210-s14/www/lectures/shortest-path.pdf
https://dev.to/jjb/part-17-finding-shortest-paths-in-graphs-using-dijkstra-s-bfs-554m
and plenty of examples
https://en.wikipedia.org/wiki/A*_search_algorithm#/media/File:Astarpathfinding.gif
https://www.google.com/maps
i think dijkstra would probably work greatly
If you are unfamiliar with graphs a previous post of mine covers the basics of them. Resour...
A* (pronounced "A-star") is a graph traversal and pathfinding algorithm, which is used in many fields of computer science due to its completeness, optimality, and optimal efficiency. Given a weighted graph, a source node and a goal node, the algorithm finds the shortest path (with respect to the given weights) from source to goal.
One major prac...
thing is i could probably code that but my programming language of choice isnt exactly the best to work with lol
well A* and dijkstra are great for finding the shortest path between two points, but idk about "visit all of those, order doesn't matter"
the code having a runtime of less then a lifetime would also be nice lol
i mean thats why the traveling salesman problem is an unsolved problem; and it's NP
well bruteforce would be relatively easy to code, but idk about the runtime...
well if i do my 5 closest neighbors
and you somehow force the start and end point, #0 and #66
that is 5^65 paths
if every path is 1 bit thats 10^32 terabytes
How about intuition. I think you could've gotten the run below 20m already if you stopped tinkering with algorithms and just played ;p
i know i should be counting 2.2 GHz but same difference it would be 10^35
i have been experimenting dont you worry
i just got the worst luck and volcano can be a PITA
Also clearly ball busting, but better than arguing over nearest points
too easy
there is probably a smarter way to go around it but considering =see above= i dont think its fruitful to do a full on maths paper on how to optimally search bones
I think the above orange path is the best you're getting if I'm being real and even then that might be a touch reset heavy as it requires missing both barrens spawns in the seeding
well I think the first thing to do would he to create a weighted graph (with time needed to go from one point to other as weight). And connections, only where it would make sense to use those connections
However, I think your category needs more rules
to get a good weighed graph that means traversing the whole map over 100 times more than likely
such as?
You start the timer after loading the dlc area, but what's to prevent someone from scoping the seeding beforehand, riding the boat back and faking tf out of the run
Might want to include creating a new save into the submission as manditory
ok new rule: video includes creating and going to volcalidus, but timer starts at volcalidus entry
Because anyone can currently just start the recording during the boat ride and make it seem like it's legit
i mean sure that's why i recorded the whole making of the save right
You did but you never specified that's how it should be handled in the future. You just stated timer starts after volcalidus load
See what I'm getting at?
Gotta define the rules or people get cheeky
Also did you setup splits for that by chance or you manually hotkeying your timer?
i guess let people scope. if the order of bones is crazy we can catch it
plus if the truck isnt where it's meant to be that's tampering
i no longer have livesplit since the wipe this was a timer added in post
Ah fair. Was curious because I was debating goofing on twitch and using a vod for submission if I managed a good run but something about manually pressing the timer key sounds like butts
you can turn the livesplit sounds off iirc
i used to use it back when and never really had issues
i was also getting cheeky like i did scout some saves to see if there is barrens bones
and i found some crazy glitch that is totally not going to fly for the speedrun but it would save time..
player wrong warping which requires the save to be new and the original "volcalidus" folder to be missing in a specific way
Yeah I know about that
See it a lot when loading peoples saves for bone hunts
Since my main is marked on the boat but if they include the player.sav it'll warp me there on vol load instead of doing the boatride
honestly as long as the save isnt "tampered with" i dont mind scouting myself
would need to think about that
missing in a specific way?
we could split the category into scouted and not scouted lol
it was weird
it worked when i renamed the volcalidus folder but not when i did it a second time..
You'd just have to delete the vol folder to get a new seed, scout it then if you found one you liked, you cut the folder and back it up elsewhere. Then you load the save without the vol folder active so it sets you back on bastion boatside then save to menu
Fools the game into believing it should load the boat next play session then you copy back in the vol folder
So when you take the boat ride, it reads your vol player.sav and drops you off there instead of the long boatride to lumberton
But yeah, it's set on the boat load to grab the player.sav and figure out where to dump you. Normally you'd be on the boat as that's the only vanilla way to transfer between so it's not a huge deal. But obviously that way of doing things isn't perfect if you play with your save files outside of the game
Hopefully that wordy explanation made sense. It marks one of the two player.savs as the active one when loading your save then uses the other when loading between areas. So you just use file shenanigans to manipulate loading behaviors
Obviously they could code it to always override the player.sav when taking the boat to a default location, but they didn't. So it's fair game for speedruns that allow it in the ruleset
Also why there were bugs with taking the boat back to Bastion dropping you into the Bastion ocean since the player.sav wasn't lined up with the boat starting point for a bit there
I "think" that got fixed during the hotfix stuff but it could honestly still be a thing
Ive seen it when deletinf the volcalidus folder but not naturally
Ill try seeing if your scouting stuff works
If i got a save without mist it cuts out that rng
allright so when i made a new save and went to volcalidus i copied that folder and that had all the good properties
then adding that to a new save allows me to go to volcalidus on the same rng
then exiting to menu without saving on the new save loads you back in allowing for easy resets
Ahh I didn't think about abusing not saving too
this is actually great, pre scouting be damned
just make the run req looking at the empty museum display from inside/on the donation box. doesn't matter if players identify where the bones are so long as they all start from the same point.
we already resolved this issue by allowing scouted runs
it is obvious if a bone is new or pre-placed
1: Augray | 00h 51m 18s | Ember Cradle
2: Augray | 00h 57m 23s | Ember Cradle
1: Steef | 00h 01m 58.7s | Ember Cradle
1: Squidy / Freeto | 00h 03m 48s | Ember Cradle
2: Squidy / Freeto | 00h 04m 05s | Ember Cradle
3: Big Mike UWU / a muffin cat | 00h 04m 20s | Ember Cradle
1: Steef | 00h 51m 20s | Ember Cradle
2: Lukas | 00h 52m 07s | Ember Cradle
3: landonairey | 00h 58m 11s | Mildew's Aquifer
1: Applebox | 00h 00m 30,6s | Ember Cradle
2: steefgozercool | 00h 00m 31,9s | Ember Cradle
3: Applebox | 00h 00m 32,3s | Ember Cradle
1: Steef | 00h 17m 24s | Ember Cradle
2: Steef | 00h 25m 05s | Ember Cradle
well, until someone else then steef even does it, it doesn't really matter
I musta missed that message, anyways you needed a clean point to start from so I figure an empty dino display would be garauntee nothing has been donated yet
im too good lol (joke okay)
it would be evident as soon as the player gets to delivery
you cant deliver without showing the museum, besides we can count
fair, I supposed the difference between preplaced and not so is the glow right?
yes
how does someone count?
no i mean count the bones that are collected
not count the bones that are delivered
again quest scroll does it for you
did you never notice that you can keep the scroll for as long as you like and just not get the tokens?
but i mostly did it with TAS
i mean sure its a storyline quest, it doesnt go away
yes he did
same as KQ and KQ2
don't even need to ask in order to know that xD
i know alot
some would say too much
i am just saying ||remote control bomb||
then just dump the scroll in the truck and start the timer at that point, could just req the truck be recalled before hand so that it's always in the exact same spot
i like starting the timing from when you load into the area
this makes it more clean instead of "stand on XYZ spot" since you can do the <soccer throw the ball> method and gain like 10 meters
eh, not my prob, I'm not doing the speedrun anyways XD
ehh, can we get back to the start of this conversation: why even change the point at which the timer is started?
i mean you coulda asked about things instead of commenting on a 3mo old comment lol
well I only found this thread today soo...yeah ๐
i have considered and lobbied with itchy to change timing methods actually but we decided against it
i think i was part of that conversation
at least i talked about that topic too at some point
2.0+ has different timing than pre-2.0 due to misunderstanding and not wanting to vet things 1x1
anywho i wanted to change all timing to 2.0+ and also remove load times
ambitious, but not doable
oh if anyone is trying to do the flag thing on mot's isle, I managed to somehow do it without doing it on my very 1st visit XD
obviously we would only count a full lap lol
I just drove through the few flags on the road from the boat to the store got out, looked at a few prices, got back in an drove round something and all of a sudden I had achievements for something I never did XD
this bug exists as long as the dlc itself. i thought they fixed it though.
don't even know how I triggered it, curious to know now that you say it's a known bug
idk the triggers. i think there where multiple.
i have no clue why the 0 second and 1 second lap thing work
the triggers are between the cones it shouldnt be able to do that
maybe it's simply driving through the finish line after just getting into the truck?
when you drive over the bridge then back over the finish it wont count
or maybe it's the angle I came in at? perhaps it only expects that angle from a done lap?
I hadn't even touched the bridge
its line or plane detection, angle shouldnt matter too much
that is what starts the timer
well I had only touched it AFTER getting the achievement just outa curiosity of what the route was supposed to be
i just checked. they "fixed" bugs with those achievements triggering 3 times.
ah yes
daily life of a dev
then maybe I managed to somehow glitch my way into starting the timer unknowingly before I could even try the route?
I did drive the roller to the boat before getting back into the truck, perhaps I managed to trigger the timer by going via an unexpected route?
I went straight over the green to the boat with it since I was not putting up with the snail pace around the u-bend
this bug hasn't be reported for a long time. i honestly doubt you could manage to trigger again (but feel free to try). The devs probably can't do anything without knowing the trigger and they probably won't look into this issue themselves, considering they are working on a new game, which takes up all their time.
only just noticed you mentioned they're working on a new game, what new game?