Cultris 2

meow · September 29, 2013, 12:59:33 AM

yotipo91 · September 29, 2013, 01:07:10 AM

I like spins on C2. They're a bit more challenging because you have to know what you're doing before you do it since the lock delay is short. Didn't know a bunch of these spins were available. I'd have to study and practice a lot of these to make use of them, but I think it'd probably be more entertaining to watch someone else who's bothered to master it

Integration · September 29, 2013, 01:31:58 PM

Kevin did some twist videos:

Guide to Cultris 2 twists
10 more twists part 1
10 more twists part 2
10 more twists part 3
10 more twists part 4

If a rotation fails, the game tests some positions nearby to which the piece might be kicked: left, right, down, left & down, right & down, 2x left, 2x right. The kicks aren't symmetric as left is always checked before right. For example the 180 degree I teleport will only work from right to left (reversely the I piece will be kicked back to its original position). Since over a year I demand to test the down kick first because it would simplify the S and Z twists and allow some additional kicks.

I did some randomizer analysis here.

Here a table which tells you how much a combo sends (at the start of a game):

Code Select

Counter -> lines | total
    1   ->   0   |   0
    2   ->   1   |   1
    3   ->   1   |   2
    4   ->   1   |   3
    5   ->   2   |   5
    6   ->   2   |   7
    7   ->   3   |  10
    8   ->   5   |  15
    9   ->   7   |  22
   10   ->   9   |  31
   11   ->  11   |  42
   12   ->  11   |  53
   13   ->  11   |  64
   14   ->  11   |  75

Note, that combos send more and more lines as longer the game lasts. For example after roughly a minute a 5 combo will send 3 lines instead of 2. On the other hand, multi-line clears won't become more powerful. It keeps at 1 line sent per Double, 2 lines sent per Triple and 3 lines sent per Tetris.

[!--quoteo--][div class=\\\'quotetop\\\']QUOTE[/div][div class=\\\'quotemain\\\'][!--quotec--](taken from the Cultris manual) Every time you delete a line you will see that you get some time on your combo-timer. If you delete lines again before the time runs out your combo-counter will rise. If you delete more than one line you'll get more time. [...] The amount of additional time you get on your combo-timer depends on your current combo-counter and on the amount of lines you cleared with the last piece. An example: If your combo-counter is currently 0 and you clear 4 lines at once, you'll get 4 times the amount of additional time than you'd get from clearing just 1 line. The additional time you get will get considerably smaller the higher your combo-counter gets.[/quote]
Here the additional time for low combo counters:

Code Select

Counter | Single | Double | Triple | Tetris
    0   |  3.60  |  +1.2  |  +2.4  |  +3.6
    1   |  1.70  |  +0.6  |  +1.2  |  +1.8
    2   |  0.70  |  +0.3  |  +0.6  |  +0.9
    3   |  0.25  |  +0.2  |  +0.4  |  +0.6 
    4   |   0    |  +???  |  +???  |  +???

You even lose a part of your combo time if the counter is already very high. Note, that you'll also lose 0.4 seconds each time you don't clear a line with a piece (additional to the time you need to drop that piece).

Sisu · October 15, 2013, 01:39:59 PM

In modes without a constant fall speed, the gravity progression is

[attachmentid=500]

but what does that even mean?

Integration · October 15, 2013, 10:47:20 PM

That formular tells you how fast a piece falls down a row, dependent on the time past in a game. If the game starts (t=0), the piece falls down a row (grid cell) every 0.36 seconds. It's 0.01 seconds less for each 14 seconds the game progresses. After 460 seconds (about 8 minutes), the gravity is maxed. Then, a piece falls down a row every 0.03125 seconds. This would mean, a piece would touch the bottom of the matrix after 0.5625 seconds.

Note that newer versions of Cultris 2 use lock delay. So if the piece touches the ground, you still have some time to move or rotate it. Initially, it's about 0.4 seconds. It looks like it becomes a little bit smaller in the course of a game. The lock delay uses a step reset system, i.e. the lock delay is reset each time the piece falls down a row. Note, that the lock delay isn't reset if the piece is kicked down a step (like in the spins in meow's initial post). IMO lock delay should never be reset because it allows you to survive high gravity by building a pyramid in the middle.

Kitaru · October 16, 2013, 12:26:08 AM

Quote from: IntegrationIMO lock delay should never be reset because it allows you to survive high gravity by building a pyramid in the middle.

I'm not sure I see the problem there. That sounds like every good game with high gravity ever to me. Would you prefer stilted Spectrum Holobyte lockdown mechanics? (Though, if it's really ~0.4 seconds lock delay as of late, -- it used to be way shorter from what I had observed -- then entry reset should still be quite playable, just a slight constraint difference.)

caffeine · October 16, 2013, 05:53:58 AM

From a design standpoint, it adds an extra level of complication unnecessarily. The purpose of lock delay is to allow movements/rotation after landing-- to a degree. With fast autorepeat, all placements take at most two movements.

What is the purpose of step reset, then? If you want to give the player more time, then balance the entry delay accordingly. Why should a piece get double or triple the time just because it descends a couple of rows? That doesn't have to do with how much input the piece requires to place it, after all.

Straight entry delay allows a more uniform and more easily balanced difficulty progression.

Integration · October 16, 2013, 09:37:40 AM

Oops I made a misstake in my previous post. Speed is maxed after 460 seconds of course.

Quote from: KitaruI'm not sure I see the problem there. That sounds like every good game with high gravity ever to me. Would you prefer stilted Spectrum Holobyte lockdown mechanics? (Though, if it's really ~0.4 seconds lock delay as of late, -- it used to be way shorter from what I had observed -- then entry reset should still be quite playable, just a slight constraint difference.)

I can't really tell, if something is wrong with Spectrum Holobyte's lockdown mechanics. You can't rotate in both directions and rotating often fails (if piece is on top of the matrix or if you rotate and move simultaneously), which prevents me even a minute on level 9 (I assume it's supposed to be a kill screen anyway). Other than that, I have no experience with entry reset. IMO it sounds nice in theory (if lock delay isn't too long), because a half filled screen wouldn't mean instant death (as in NES or original Game Boy Tetris). The only problem I see is that it's hard to predict when the next piece spawns.

Step reset can have their advantages in multiplayer (e.g. doing T spin combos in SRS), but IMO they are a no go in Marathon (I regard a Cultris bot match against 0 opponents as Marathon). Step reset can be abused by building pyramids, move reset can be abused with infinite rotation. In my opinion, if a piece touches the ground twice, then you were too slow.

Anyway, the problem in Cultris is rather that its final speed is unchallenging. You have roughly 0.8 seconds per piece (0.5 seconds to bottom, 0.3 seconds lock delay). This corresponds to 1.25 pieces per second.

-----------------

I will post some Cultris video since this is supposed to become the default Cultris thread.

Maserati WR (40 lines in < 30 seconds)

First 14 combo

Uber L33t (last singleplayer challenge)

Team Country Tourney 2012 final

Spring Tourney 2012 winner bracket final
(Jtadore vs Blink, last few minutes missing)

ohitsstef · October 16, 2013, 09:42:53 AM

New favorite challenge mode on c2 = James Clewett's

Make as many tetrises as you can is really challenging without hold and more previews. But I've learned to stack flatter somewhat

Sisu · October 16, 2013, 05:23:46 PM

Déjà vu @Integarufeine

Could a new marathon mode be forged by merely tweaking existing C2 infrastructure?

For starters, remove the 460.25 gravity cap.

Integration · October 16, 2013, 10:48:29 PM

Quote from: SisuCould a new marathon mode be forged by merely tweaking existing C2 infrastructure?

For starters, remove the 460.25 gravity cap.

If you remove the gravity cap, game will become 20g like after a while (piece spawns at the bottom). With the current formula it would happen after 8.4 minutes. I plead in favour of a system that reduces both gravity and lock delay at the same time. BTW the gravity cap worked fine before the lock delay was introduced.

This talk is really hypothetical. I don't believe in major Cultris updates anymore.

Kitaru · October 17, 2013, 01:06:10 AM

Quote from: IntegrationI can't really tell, if something is wrong with Spectrum Holobyte's lockdown mechanics.

In the SpecHolo case, entry reset is mostly crippling due to poor interaction with how the rest of the game is tuned -- key repeat start-up is not sufficiently fast and the lock delay is too restrictively short. Without pyramiding, the game becomes pretty much straight-up unplayable at high speeds. This puts a bit of a sour taste in my mouth with regards to entry reset. However, with all other timings properly tuned, I think the distinctions between entry reset and step reset in practice will typically be fairly minor. I'm just kind of surprised to see someone explicitly advocate against step reset or cite pyramiding as a major concern.

I recall Puyo Puyo using entry reset, FWIW.

Quote from: Integration(I assume it's supposed to be a kill screen anyway).

One would have to wonder then, why would they give you the option to start right at the kill-screen? Anyway, I think it has much more to do with lack of tuning than any lofty intention.

Quote from: IntegrationOther than that, I have no experience with entry reset. IMO it sounds nice in theory (if lock delay isn't too long), because a half filled screen wouldn't mean instant death (as in NES or original Game Boy Tetris).

Brief aside, half-filled stack isn't exactly death in NES 19, but your movement options are starting to be more dangerously constrained. It all depends on stack condition, of course, but with a well-managed stack it's not necessarily even a problem at all -- you just need to start downstacking out of it, lest you risk deficits on the left side of the stack that will trigger the start of actual choking. Even with a poor stack there may still be hope, but you'll probably have to nail the input on several tight moves in order to recover.

Quote from: IntegrationThe only problem I see is that it's hard to predict when the next piece spawns.

The next piece spawns when you press down! Barring manual lock, good visual (fading, lockflash) and audio (collision and lock ticks) design also provide good feedback about when to expect the next piece.

Quote from: IntegrationStep reset can have their advantages in multiplayer (e.g. doing T spin combos in SRS), but IMO they are a no go in Marathon (I regard a Cultris bot match against 0 opponents as Marathon). ... In my opinion, if a piece touches the ground twice, then you were too slow.

Quote from: IntegrationIf you remove the gravity cap, game will become 20g like after a while (piece spawns at the bottom).

I was kind of working from a framework of 1~20G speeds being part of the picture. If there is a gravity cap in consideration, then I suppose I can see how that could influence a stronger opinion on enforcing said lockdown constraints.

Quote from: IntegrationAnyway, the problem in Cultris is rather that its final speed is unchallenging. You have roughly 0.8 seconds per piece (0.5 seconds to bottom, 0.3 seconds lock delay). This corresponds to 1.25 pieces per second.

I find it interesting that you measured the challenge solely by the minimum rate of play rather than in terms of play constraints. No further comments on that point, just an observation.

Quote from: IntegrationThis talk is really hypothetical. I don't believe in major Cultris updates anymore.

Perhaps. I think hypothetical talk can be a good thing too, though.

meow · October 17, 2013, 07:20:58 AM

[fumen]v110@DdE3kbE3kbD3lbE3kbE3gbA3ibH3hbH3hbH3hbI3gb?I3gbI3qbA4G7eCSkc6DBAAA[/fumen][fumen]v110@DdE3kbE3kbD3lbE3kbE3gbA3ibH3hbH3hbH3hbI3gb?I3gbI3qbA4G7eCCkc6DBAAA[/fumen]

This is a twist that I've always tried and failed in other games.

Integration · October 17, 2013, 10:21:00 AM

Quote from: KitaruBrief aside, half-filled stack isn't exactly death in NES 19, but your movement options are starting to be more dangerously constrained. It all depends on stack condition, of course, but with a well-managed stack it's not necessarily even a problem at all -- you just need to start downstacking out of it, lest you risk deficits on the left side of the stack that will trigger the start of actual choking. Even with a poor stack there may still be hope, but you'll probably have to nail the input on several tight moves in order to recover.

What I meant is self made garbage. If you've blocked half of your matrix, it means that you've only half of the time to move a piece in NES or Game Boy Tetris. 20g is the opposite. It doesn't matter how many lines you've blocked. In my opinion time per piece should dependent on your stack's height, but not proportionally to the number of empty lines in your playfield. A short lock delay is a good balance.

Quote from: KitaruI find it interesting that you measured the challenge solely by the minimum rate of play rather than in terms of play constraints. No further comments on that point, just an observation.

In case of Cultris your average BPM can tell you pretty well, which gravity you're able to handle and which not. Instant DAS means that you can get to every spot rather quickly. I've tried the 0 player mode on an older version (no lock delay). Maximum gravity means 105 to 110 BPM if you stay at the bottom. I was able to survive it longer than a minute but not 2 minutes. My average BPM lies at 125. I am sure, the fastest players wouldn't be able to handle a 20 % higher gravity.

Quote from: KitaruPerhaps. I think hypothetical talk can be a good thing too, though.

I don't want to talk about it in detail unless Simon shows some interest. Some more general things about a hypothetical Cultris Marathon mode:

20g (or close to it) is bad because you can choose spawning directions in Cultris. Players using SRS spawning directions would be disadvantaged, because it's unprobable that you're still able to rotate the piece if it hits the ground
I would make gravity dependent on dropped blocks resp. cleared lines. if gravity is based on elapsed time, you are forced to play as fast as possible in the first minutes. Marathon is more about relaxing and good stacking/scoring Tetrises. Pure speed is better suited for ultra-like modes (Multris 10 minutes or Palatris 5 minutes are 2 examples).
Regarding the speed curve, I'd prefer a convexe function over a linear function. For example you could use a quadratic function. If you specify the lowest point, the highest point and the slope (derivation) in the end point, a quadratic function is uniquely determined. An example (the dashed blue line is the current gravity curve):
[!--ImageUrlBegin--][a href=\\\"http://i41.tinypic.com/i5wzfr.png\\\" target=\\\"_new\\\"][!--ImageUrlEBegin--][img width=\\\"400\\\" class=\\\"attach\\\" src=\\\"http://i41.tinypic.com/i5wzfr.png\\\" border=\\\'0\\\' alt=\\\"IPB Image\\\" /][!--ImageUrlEnd--][/a][!--ImageUrlEEnd--]

Kitaru · October 18, 2013, 02:32:54 AM

Quote from: IntegrationWhat I meant is self made garbage. If you've blocked half of your matrix, it means that you've only half of the time to move a piece in NES or Game Boy Tetris. 20g is the opposite. It doesn't matter how many lines you've blocked. In my opinion time per piece should dependent on your stack's height, but not proportionally to the number of empty lines in your playfield. A short lock delay is a good balance.

See, for NES 19, it's never really a matter of "time to move the piece." (The only way it can be considered as such is if you have Thor Thumbs, but that is a whole different kind of discussion.) Rather than consider the amount of time to execute, it's much more helpful to look at it in terms of height affecting the range of a piece and the resulting valid placement options. DAS and gravity interact in a fixed, consistent fashion, -- it's not like starting to buffer earlier is going to give you more charge or faster DAS -- so you're nearly always approaching the game from a constraints standpoint.

Also of note, entry delay in NES scales with placement height, which meters the resulting BPM increase somewhat. I think 10 rows actually works out to around 70% net time per piece relative to empty rather than 50%.

Quote from: IntegrationIn case of Cultris your average BPM can tell you pretty well, which gravity you're able to handle and which not. Instant DAS means that you can get to every spot rather quickly. I've tried the 0 player mode on an older version (no lock delay). Maximum gravity means 105 to 110 BPM if you stay at the bottom. I was able to survive it longer than a minute but not 2 minutes. My average BPM lies at 125. I am sure, the fastest players wouldn't be able to handle a 20 % higher gravity.

Right OK, makes enough sense. Instant DAS would shift the focus away from being a "constraints game" a fair bit.

Quote from: Integration20g (or close to it) is bad because you can choose spawning directions in Cultris. Players using SRS spawning directions would be disadvantaged, because it's unprobable that you're still able to rotate the piece if it hits the ground

Refresh my memory -- pieces in Cultris can't kick up? That would be a concern.

Quote from: IntegrationI would make gravity dependent on dropped blocks resp. cleared lines. if gravity is based on elapsed time, you are forced to play as fast as possible in the first minutes. Marathon is more about relaxing and good stacking/scoring Tetrises. Pure speed is better suited for ultra-like modes (Multris 10 minutes or Palatris 5 minutes are 2 examples).

I don't know about relaxing, but I do agree with decoupling time and progress. Perhaps it can be justified for things like Blockbox Survivor / Nullpomino Dig Challenge, but I'm leaning against it for marathon-likes. (Of course, the primary purpose of this system in Cultris is to serve for multiplayer matches, so it's not necessarily going to be ideal for a single player context.)

Quote from: IntegrationRegarding the speed curve, I'd prefer a convexe function over a linear function. For example you could use a quadratic function. If you specify the lowest point, the highest point and the slope (derivation) in the end point, a quadratic function is uniquely determined. An example (the dashed blue line is the current gravity curve):
[image]

Also digging this.

Here is the NES speedcurve for sake of comparison (because Reasons): link. (I'd embed, but HD doesn't accept dynamic pages as image targets.)
There are a few places where it flattens out, as the minimum resolution is one frame. Other than spots where values were snapped to the nearest frame, it plots a pretty nice curve.