TT lower bound

General discussion about the games and the scene.

Moderator: Moporators

Post Reply
User avatar
AndrY
38mins club
Posts: 639
Joined: 14 Dec 2015, 17:41
Team: .lev
Location: Perm, Russia
Contact:

TT lower bound

Post by AndrY » 24 Jun 2020, 15:53

Its just interesting imo what is the optimal TT for ints without unallowed bugs.
Upper bound is like some 33.xx (saveload)
But at least, which is the lower bound?

Is it provable that 29.xx is impsy? Or 19.xx at least?
How to prove it? By mathematical+logic calculaions, with AI, with some other programming?
Image
My levs and other Elma things: www.elma-andry.wmsite.ru

User avatar
ArZeNiK
38mins club
Posts: 704
Joined: 30 Jul 2016, 09:18
Team: Ferrari

Re: TT lower bound

Post by ArZeNiK » 24 Jun 2020, 16:50

this has already been discussed about 23849256892413592845 times and the answer was always "it is impossible to determine"
hi im arzenik :>
september 23th 2017 never forget
Image

User avatar
AndrY
38mins club
Posts: 639
Joined: 14 Dec 2015, 17:41
Team: .lev
Location: Perm, Russia
Contact:

Re: TT lower bound

Post by AndrY » 24 Jun 2020, 17:14

ArZeNiK wrote:
24 Jun 2020, 16:50
this has already been discussed about 23849256892413592845 times and the answer was always "it is impossible to determine"
oke dont remember then:/ where is it?

but must be pos imo. mb determine max pos speed in each part of lev?

if to consider one route, bike can take speed when gas or when push...

hmm its a math problem imo)
There is a part: starignt way from one point to another, like in int02. And there is NO MATTER what kind of corners or polys there. there must be limit of time to pass this part, and it must be pos to determine it somewhere about, knowing the initial speed.

-----------------------------------------------------------------
Btw, imagine: there is a start in one point and flower in another, you have to draw polys between them (smth like tight ===== is not allowed) to minimize time in this lev. It depends on location of start and finish, but spikes/rails seems the best variants in diff cases:/
ahh also the initial speed of bike could be >0 and head direction could be not up.
Image
My levs and other Elma things: www.elma-andry.wmsite.ru

User avatar
ArZeNiK
38mins club
Posts: 704
Joined: 30 Jul 2016, 09:18
Team: Ferrari

Re: TT lower bound

Post by ArZeNiK » 24 Jun 2020, 19:03

hi im arzenik :>
september 23th 2017 never forget
Image

User avatar
pawq
38mins club
Posts: 6447
Joined: 24 Aug 2008, 19:56
Team: TR
Location: Southampton, UK

Re: TT lower bound

Post by pawq » 24 Jun 2020, 21:14

viewtopic.php?p=7771#p7771


ArZeNiK wrote:
24 Jun 2020, 16:50
this has already been discussed about 23849256892413592845 times and the answer was always "it is impossible to determine"
More like 6234509169089865143695072306951094356912139068745 times. Pretty funny that after 20 years and "theoretical limits" being broken countless times, some people still think it's possible to predict the future improvements.

User avatar
AndrY
38mins club
Posts: 639
Joined: 14 Dec 2015, 17:41
Team: .lev
Location: Perm, Russia
Contact:

Re: TT lower bound

Post by AndrY » 25 Jun 2020, 06:38

oke oke.
these topics are (imo) about finding the best pos TT. i mean finding at least some 100% impos TT.

you can not improve TT to 0:00.00 ye?
for example 0:00.54 TT is impos, bcz of distances between start and flower in ints. pos to raise it?:/

(oke mb not the most important question, but just interesting:/)
Image
My levs and other Elma things: www.elma-andry.wmsite.ru

User avatar
kuchitsu
Kuski
Posts: 1388
Joined: 13 Aug 2010, 20:31

Re: TT lower bound

Post by kuchitsu » 25 Jun 2020, 08:31

Yeah, you guys are misunderstanding Andry's question a bit. He isn't looking for the perfect TT but for a way to set some boundaries for it. Can we prove that the perfect TT would be over 5 minutes? 10? 20? How high a number can we name with certainty and what methods could be used for that aside from pure guesswork? Like with many things in mathematics, it can seem pointless to prove obvious statements like "perfect TT is over 5 minutes", but that would be a start that could eventually lead to further, more interesting developments.

User avatar
AndrY
38mins club
Posts: 639
Joined: 14 Dec 2015, 17:41
Team: .lev
Location: Perm, Russia
Contact:

Re: TT lower bound

Post by AndrY » 25 Jun 2020, 09:49

ye i this mean:)

most simple imo:
Divide a lev to parts, find the best time in each (minimize by polys used between start and finish points). And then minimize total time in lev by choosing the points in break lines (between 1-2 and 2-3 in pic), bike angle in brake lines and its speed there.
Image

ofc need to check all possible routes, not only this
Image
My levs and other Elma things: www.elma-andry.wmsite.ru

User avatar
ribot
Not banned
Posts: 2377
Joined: 19 May 2002, 16:20
Location: Miranda: the true state
Contact:

Re: TT lower bound

Post by ribot » 25 Jun 2020, 10:54

There should be ways to calculate this, but it's not that simple.

First you have to find all possible routes, but that's not trivial in any way. Since frame rate is not absolute when you play (or is it absolute when you used fixed frame rate?) there are endless possibilities for even short levels. A small difference can mean a new allowed bug in pretty much any place of the level.

To find acceleration of any level you have to know the theoretical max speed in all allowed bugs, such as pops, bounces or bugs used in animal farm/tricks abound.

Theoretically you can do an allowed bug with acceleration much faster than any style that you can practically reproduce. Then you have to find a new style for the following part.

Take this start of rec as an example: https://elma.online/r/xjkrltec7h
If you find a bug like this it 1) accelerates faster than actual physics acceleration, and 2) makes wr style useless. In theory this rec could probably jump from wall to flower and make much faster time.

So the theoretical tt would just be pop after pop after bounce after bounce... which is not really that interesting. In practice however it's a bit different, since the bugs used are reproducible.
-
"leader status in the Elma against-the-system underground" - Abula
-
IncrElastoMania - Elma Simulation - Browser Game 2020
Elma Imager - Command Line Tool 2020

User avatar
AndrY
38mins club
Posts: 639
Joined: 14 Dec 2015, 17:41
Team: .lev
Location: Perm, Russia
Contact:

Re: TT lower bound

Post by AndrY » 25 Jun 2020, 11:07

ribot wrote:
25 Jun 2020, 10:54
To find acceleration of any level you have to know the theoretical max speed in all allowed bugs, such as pops, bounces or bugs used in animal farm/tricks abound.

Take this start of rec as an example: https://elma.online/r/xjkrltec7h
hmm i think pexi will not take this rec to wr table... although :? if even to allow any such pops (but with fixed fps >= 30), there must be a bound in each part.
ribot wrote:
25 Jun 2020, 10:54
So the theoretical tt would just be pop after pop after bounce after bounce ...
yes, but even in this case what is the lower bound for each part? 1, 2, and 3 in pic (3 small down pics). How would you draw polys between start and finish to achieve the best time even with such wheelpops? (ofc there is no reason to use tight wheel-pipes and such, bcz there are no such things in int04)

i want just some very rough estimate at first, as kuchi said
probably would be pos in future to generalize it to the case of game in real time somehow:/
Image
My levs and other Elma things: www.elma-andry.wmsite.ru

User avatar
Tm
35mins club
Posts: 585
Joined: 13 Feb 2005, 15:08
Team: EF

Re: TT lower bound

Post by Tm » 25 Jun 2020, 11:33

the lower bound is ofc 32:43,01 :lmao:
Image

Post Reply