Post subject: Columns III
Banned User
Joined: 12/23/2004
Posts: 1850
Columns 3 is a deviation from the normal Columns -- now you have to fight against the CPU, as opposed to entertaining youself by doing the same thing over and over. Right. You control when you attack -- clearing gems earns 'points' (3, usually, but it goes up if you get more in a row or clear several at once, etc), up to 30 total, each 10 adding one row to your opponent's field. Then there's the special gem that comes up every 20 'clears' (that smaller, white number in the playfield) that allows you to add 2 rows, clear 3 of your own, or the infamous "delete all blocks of this color" thing. Of course, each CPU is a little different in strategy (some don't attack until you do, etc)... I tried making a test movie, got through with r1f1 (round 1, fight 1 -- best of 3, some absurd number of rounds) in 0:38 in-game... ...of course, Gens decided to delete my movie for no reason (after it desynced, aparrently it was starting one frame too early) -_-
Perma-banned
Player (71)
Joined: 8/24/2004
Posts: 2562
Location: Sweden
If you decide to make another movie, I'd like to watch it. :) I used to play columns alot for Sega Genesis.
Editor, Reviewer, Experienced player (979)
Joined: 4/17/2004
Posts: 3109
Location: Sweden
>...of course, Gens decided to delete my movie for no reason Never heard of that ever happening. My initial guess is that the actual perpetrator is situated somewhere between the screen and the back of the chair. :P
Banned User
Joined: 12/23/2004
Posts: 1850
Truncated wrote:
>...of course, Gens decided to delete my movie for no reason Never heard of that ever happening. My initial guess is that the actual perpetrator is situated somewhere between the screen and the back of the chair. :P
No, I found it in some completely random folder off in space (Firefox's main app folder... right) When I loaded Gens again and tried to save another movie, it tried saving in some folder in Program FIles\Common\NT[something\####\...... it made no sense. Anyway, my old time for stage 1 was 0:38, then when I tried it in real-time it was 0:45 -- ironically, the first stages are usually much harder to beat fast, since you have to work their field up the whole way -- they place blocks too slowly.
Perma-banned
Jungon
He/Him
Player (50)
Joined: 4/2/2009
Posts: 376
Location: Porto Alegre - Brazil
Is there anyone working on this currently? And... if someone would care to give a try, it would have to beat the spider in 0:38, which means 38 seconds? I might give it a try when starting to study TASing in Gens 11a... õ_o
*Jungon*/*Johnnyz* smilie weirdo =P ^^ o.o @__@ +D
Dimon12321
He/Him
Editor, Reviewer, Experienced player (596)
Joined: 4/5/2014
Posts: 1222
Location: Romania
TASing is like making a film: only the best takes are shown in the final movie.
Dimon12321
He/Him
Editor, Reviewer, Experienced player (596)
Joined: 4/5/2014
Posts: 1222
Location: Romania
wesen has posted a new WIP where he wins the 1st battle against Spider in 21 secs which is faster than mine. http://tasvideos.org/userfiles/info/57035905288073668
TASing is like making a film: only the best takes are shown in the final movie.
Player (200)
Joined: 10/9/2016
Posts: 6
Hi, as you can see in my WIP the round finishes with three orange jewels left on the field. This makes me think that this might still not be the perfect way to finish this round (with this set of columns). Therefore I would like to try to create a script that finds the best strategy to use for each round. My idea for the fastest way to finish a round is this: 1. Uses the least number of columns (Primary condition) 2. Has the least total chain length (Secondary condition) The fastest way to get 30 points should be a chain with two 1x3 jewels and one 2x3 jewels with different colors (as used twice in the TAS WIP) which results in (3 + (3 +4) + (3 + 3 + 8 + 8)) = 32 points. My initial idea was to simply test all possible column placements and shuffles. However there are way too many possiblities. Even after filtering out a lot of possible column placements that do not need to be tested (mirrored, movable across the field or with empty field columns in between) there are still 4.209.975 possibilities to place and shuffle only 6 columns on the board (124.582.455 for 7 columns). The TAS WIP needs 14 columns to finish the round and even that is only possible because the AI stacks two columns which skips one attack. If anyone has ideas on how to test all possibilities that might be the fastest way to win please let me know. Also ways to finish Round 1-1 faster are welcome as well.