Design Thoughts

[caffeine]

I played yesterday on Nullpomino a lot of 6-Players game, and it wasn't equivalent in any way to 1v1, it was even worse than before.

The reason is simple in 1v1 the hole change  for each garbage is roughly like this :

0,0,0,100,0,0,0,100,....

In 6 players games its like :

20,20,20,20,20,20,20....

Which does not achieve the same garbage distribution at all.

I don't understand what that means. I played some ffa games in the default multiplayer mode, and the garbage was very similar to 1v1.

EDIT: perhaps we should continue this in PMs so we don't clutter up half the topic talking about this one detail.

[meow]

"There's a very good way to get the fairness of "send to all," but the pace of 1v1: a fractional system."

A 1v1 pace for a several player game is too slow. The length of guideline games is relatively long, slowing it down further only frustrates the dead players. There is no incentive to attack because the safer option of cancelling garbage and downstacking is superior for survival.

In practice, I've encountered several flaws with the garbage division system. It does not achieve the same feel as a targeting system. The garbage distribution might be similar but the timing is different and this creates a very different feel to the game. It gives rise to Tetris DS like tactics where you stack around the garbage hole in anticipation of the next garbage lines, which never appears in TOJ. In a 6 player game, it starts off like super easy Tetris DS -> normal Tetris DS -> ... -> TOJ 1v1.

A better way is to send each attack to the player who has received the least lines taking into account lines sent by each player so that a good player isnt disadvantaged. I think Wojtek suggested something like this, but here is my take on it.

Eg: There are 4 players.
Now it's B's turn to send lines.
B can send it to A, C or D.

How does B decide which player to send lines to?
Does he send to A? Compare the lines A has received, C has received (excluding lines sent from A), and D has received (excluding lines sent from A).
Does he send to C? Compare the lines C has received, A has received (excluding lines sent from C), and D has received (excluding lines sent from C).
Does he send to D? Compare the lines D has received, A has received (excluding lines sent from D), and C has received (excluding lines sent from D).
Send it to the player who deserves lines or to a random player if it is tied.

This way, sending lines does not affect you negatively since your line sends are subtracted in the comparisons. In theory, a system like this should keep the same feel as a 6p targeting game and it would be fair. I haven't thought about it in too much depth though, it may not work at all...

I still believe that >2 player games should be fast. No one wants to wait. Send to all without any divisions or reductions is the best. Waiting times are short, the games are intense, the best player wins the majority of the times, there is potential for other players to win. In short, it is the superb multiplayer that Tnet2 players all know of and love, only guidelined.

[myndzi]

From what I observed about the "divide hole change percent" setting, I didn't like it for large games either. I wound up seeing quite a lot of very long, straight, holes. An alternative might be instead to not make it random: make it change, for sure, every N attacks; possibly allow it to change randomly in less than that, but don't allow it to exceed a "tetris" in depth...

As for 1-2-3-4, a possibility for not making combos overwhelming here lies in another thing I'm playing around with right now: mutually exclusive attacks. In this system, instead of adding line clear/spin attack to combo attack to perfect clear attack, it simply takes the highest of these and uses it. In this fashion, the most you would send is 4, and your combo leadup would depend on the attacks used to get it. If they were all singles, you'd see 1, 2, 3, 4, 4... (or 5?); if they were more, you might see something like: 4, 3, 3, 4 (Tetris, triple, single, single).

[perfectclear]

From what I observed about the "divide hole change percent" setting, I didn't like it for large games either. I wound up seeing quite a lot of very long, straight, holes. An alternative might be instead to not make it random: make it change, for sure, every N attacks; possibly allow it to change randomly in less than that, but don't allow it to exceed a "tetris" in depth...

As for 1-2-3-4, a possibility for not making combos overwhelming here lies in another thing I'm playing around with right now: mutually exclusive attacks. In this system, instead of adding line clear/spin attack to combo attack to perfect clear attack, it simply takes the highest of these and uses it. In this fashion, the most you would send is 4, and your combo leadup would depend on the attacks used to get it. If they were all singles, you'd see 1, 2, 3, 4, 4... (or 5?); if they were more, you might see something like: 4, 3, 3, 4 (Tetris, triple, single, single).

A problem I see right away with the mutually exclusive attacks is that it sucks power out of three widing, preventing that from being an effective counter to four widing.

[myndzi]

I don't believe a great deal of attack comes from doubles in three wides, but you'll note that I did state I am using it in conjunction with a much modified combo table as well, so four wides aren't overwhelming. It's also being used with all-spins. Other factors may need to be balanced, but it is an option to consider is all I am saying.

Edit: I think we're getting off track a little, here, anyway. Remember that the game we are discussing has solid garbage! How does this affect things? Well, messy vs clean is worthless in this case. The only counter to attacks is to attack equally. Nerfing combos is especially important here since you can't use the garbage you receive to help counter them if it comes to that. I also think that, in general, the attacks should probably be on the mild side since, once again, you can't utilize the garbage you've received so the primary focus is on racing to output. This kind of game would emphasize spikes and efficient attacking, so I think Tetrises will get the go-by. Tetrises seem mostly used in TOJ-style play because of the clean garbage, and that won't happen here.

If you want to emphasize Tetrises, I think you really have to make T-spins and combos a lot weaker or harder to pull off. Combos are probably going to be the most effective because they have potential to generate a massive spike that will KO your opponent's ghost, particularly at the  beginning if they are stacking up still.

[caffeine]

A 1v1 pace for a several player game is too slow.

[...]

A better way is to send each attack to the player who has received the least lines taking into account lines sent by each player so that a good player isnt disadvantaged. I think Wojtek suggested something like this, but here is my take on it.

First off, any targeting system gives the same pace as a 1v1 system.

The main issue I take with the "send to who received the least lines" is that there's still an annoying luck factor. It's better than traditional targeting in that a player won't receive all garbage rows at once, but there're still problems, which make a fractional system much more desirable. The current target will cycle through players just as frequently as the old targeting system, only it won't overlap a player at the same time. So, for example, if all 6 players are near the top, and I happen to be next in line to receive an attack, and that attack happens to be a B2B TST (7 rows of garbage), I'll top out, and that was pretty unlucky considering my opponents didn't feel that B2B TST at all. So this is why I advocate fractional "divide by opponents left" as a lesser evil.

Btw, I'm totally with you in that I prefer fast paced 6-way games (aka the "attack all" of the old tnet2 games), but apparently that's too hardcore for TTC nowadays, so I was trying to anticipate that "drawback."

[myndzi]

I don't really see why 1v1 should be the same as, say, a 6 person FFA anyway. If you want 1v1, go play in a 1v1 room. A slight argument can be made in that, in a 6 person FFA, it's more like "1v1 vs whoever plays the best out of 6 people", but I don't think that's really a significant advantage. It's certainly worth being optional in something like Nullpo, but, particularly in a game played against replays, there's no reason to make 1v1 the same "kind" of game as 1v1v1v1v1v1.

[Paradox]

fractional garbage is way more fair than what tf and toj do. Who cares if it isn't like 1v1- it isn't supposed to be. It just needs to be balanced so that games don't end too quickly or last too long and receiving garbage isn't luck based. It does all of these. I noticed a problem with the change on attack but I thought it was an error that they fixed because I don't notice it anymore.

[Spirale]

fractional garbage is way more fair than what tf and toj do. Who cares if it isn't like 1v1- it isn't supposed to be. It just needs to be balanced so that games don't end too quickly or last too long and receiving garbage isn't luck based. It does all of these. I noticed a problem with the change on attack but I thought it was an error that they fixed because I don't notice it anymore.

it's fair but it changes the game in a way that its not fun anymore. Random target is more fun and is relatively fair (best player will win most of the games)

[myndzi]

Paradox: Remember the game we are talking about here. You aren't playing real people, so targeting isn't going to piss them off, and it's good in a way since it helps teach a player to pay attention to other fields. If and when things go to live, it'll be a different game anyway. Whether the mechanics should remain as close as possible is perhaps worth discussing.

Still, if we are going to talk about live, you have to take a little care not to result in a game that is just bland. Fractional garbage really smooths everything out to where spikes are pretty much useless. I like the fact that with large free-for-all games, the way you have to play to be successful changes as players get knocked out. I agree that targeting can be unfair to specific people if exploited, and I'm all for designing something that isn't exploitable.

[Alexsweden]

Thanks Alex, I actually just realized.
LPM, APM and those were stats actually I used to emphasize a lot in hangame and TOJ versions..
in the facebook version I have forgotten about it completely.

Yeah, I should start bringing back the stats because it's really fun to analyze the styles.

Ohh, nice. Would love if there would (in the future) be stats in real time(like tetris zone), or in the replay so one could se the differences over time.

[zaphod77]

Well Battle 6P is a strange game.

Here' is my idea on how to play.

Real human's get full targetting.

Replays can't truly target.

So instead, they send fractional garbage to all live humans, and the rest to their chosen target among the replays. That way, human's don't get unlucky death spikes from a group of replays that just HAPPEN to target them.

[Beastin_Shen]

sorry for double post, but this seems the more appropriate place.

eh dunno if this is a bug, but at 45+ rank it started to pair me against A.I, which constantly was 18-20 second sprints in sprint 5p. Uhhh i get 2nd everytime? pretty annoying :/

other things:

Items are very expensive, although i guess thats why the offers are there

the handicap is too strong for ranks 30-50 when playing friends, especially when the sprints dont really get any faster

6p isnt promoting me early like the other games, therefore i send 140 lines while guy in 2nd sends 1

[Paradox]

I have the max upgrades and it is still pretty slow. I honestly don't see much difference its only slightly faster. Autorepeat should be way faster + soft drop. Line clear is pretty good.

[Wojtek]

acemagic, if your dream is to play tetris with Michael Jackson, this is how you do it: http://arcade.svatopluk.com/jeil/final_tetris/ .