Thoughts about possible anti-teaming solutions

forget ffa. lets turn it off and add "red&green vs blue&yellow" at the main game type

Right now we can group most of anti-teaming solutions to 2 groups:

1) Tune points system.

2) Make changes to standings and/or rating calculations.

Looks like I found the 3rd way today. A theoretical solution of teaming is: Make standings (and/or points) not guaranteed ("not frozen"). I don't know yet how exactly can it be done. So, I describe just a possible direction.

Now players earn some points and the points could not be decreased during a game. If you captured a rook and then checkmated someone, you get 25 points and nothing can make this number smaller. Even if you get checkmated, you still have those 25 points.

If a player is alive, his points may: 1) increase, 2) stay the same.

If a player is "dead", his points may only stay the same.

This is good. Right?

What is bad: After 2 teamers kick 2 other players, points of teamers may increase or stay the same, but points of those unfairly kicked players can only stay the same. So, teamers do not risk at all, they both know they are the 1st and 2nd in any case.

Now imagine that points of any player may increase or decrease at any moment of the game (maybe even that player is "dead" already) under certain circumstances.

If points of a kicked player may increase or alive player's points may decrease then 2nd place may become 3rd and one of 2 teamers will lose.

I have some thoughts about "when it may happen", but I don't have a decent solution yet to show it to the community. I ask you to think about it, open your mind, free up your imagination and creativity and brainstorm this.

When points of an alive player should be decreased? (For what action of this player or other players?)

When points of a "dead" player should be increased? (For what action of other players? Obviously, "dead" players cannot act themselves.)

Some drafts:
1) Only 2 players left. One of them resigns (times out, disconnects) less than 10 (15, 20) moves later. He/she loses points. (Forces them to keep playing when only 2 players left. It is important.)
2) Only 2 players left. For every 3 (5, 10) consecutive passive moves the leader loses 3 (5) points. ("Passive" is when a player who is on the 1st place (if he/she is alive) doesn't capture a piece or move a pawn) . (Forces the leader to play more active.)
3) Only 2 players left. "B" checkmates "A", "A" loses some points.
4) Only 2 players left. "A" checkmates a stronger player "B" without help from other player(s), "B" loses  some points. What is "stronger"? A player with a bigger army.
5) Only 2 players left. "A" checks "B", "B" loses 1 point.

@BabYagun I think that losing points as an anti-teaming solution would work equally well with all other anti-teaming solutions suggested so far: would not work. Teamers have honor codes. They will not push their teammate down to the 3rd place for any reason. Mutual trust is their biggest asset. Maintaining it is top priority for them.

You cannot say they all work equally. For example, "1 winner 3 losers" does solve it.

> Teamers have honor codes. They will not push their teammate to the 3rd place for any reason.

Yes, but only if they have control. Recently someone on this forum offered a rule: "If there is a hanging piece of an opposite player, there is a 10% chance that 4PC server (not player) will make a move and capture that hanging piece." So, if Red hangs his queen, and Yellow is his trusted partner, there is still a chance that at any random moment the server will make Yellow's move and capture that queen. It is an artificial rule, converting 4PC to 4 Player Backgammon with a special dice. We will not implement it. But the idea behind this suggestion is: If teamers don't have 100% control, they cannot honor their code fully.

A funny thing is: Imagine that there is a 10% chance of auto-capture. In this case any player can capture his opposite's piece and then tell: "It was not me! It is the server!", but his "partner" cannot check it (server or player?) and will think that maybe he should not trust this teammate any more. So, this dice throwing rule ruins the trust between teamers.

If we can find some rule that makes a similar thing without throwing a dice, it can be a good anti-teaming pill.

This rule could be exploited by non-teamers, who want to damage there opposite player for some reason.

Scenario example: Blue threatens to checkmate Red in his next move. Red can easily parry the threat, but Yellow decides to sabotage Red (possibly because Red is the highest rated player) and offers him a free knight. Now there is a 10% chance that Red will auto-capture Yellow's knight, and then get mated by Blue!

Maybe I didn't explain it clear enough: I don't like this particular rule. And it is not going to be implemented, because we are not going to add dice throwing to the main 4PC variation. I show it only as an illustration of a principle. So, there is no need to spend time to discuss this rule and how can it be exploited. We should look at its principle only.

P.S.

If this rule will exist in some custom 4PC variation then checkmate protection will have the highest priority and the server won't capture that yellow trojan horse.

@BabYagun oh, come on! If you don't want to discuss this rule, then why are you trying to fix its flaws? I understand that we are talking on a very hypothetical basis here!

Such a mechanism could be more effective as an inti-teaming measure, if its inner workings were kept secret from the public. For example the random chance of an auto-move could be higher for players with a history of supporting unprotected pieces of their opposite. But only "Dashes" should know that.

> ... then why are you trying to fix its flaws

It took me less then 3 seconds to find this solution.

> if its inner workings were kept secret from the public

You are not serious. Right?

1) It would take three full days for "Dashes" to implement this fix, and I am optimistic here. And there would be still room left for more subtle exploitations.
2) I am exactly 90% serious.

Please back test the archive of games from August until now with the simple change to the value of the King.  Then please tell us in the forum what the results were for the % of opposites coming in 1st and 2nd.  

I've put in nearly fulltime hours over the last 4 years into Daily Fantasy Sports (DFS) betting.  Paying yearly fees for access to some of the best Sports analytical data scientists and Professional gamblers.  Many of the models I build from the software I pay for is always back tested.  It takes just minutes.

> Please back test the archive of games from August until now with the simple change to the value of the King. 

First of all I want to say that I appreciate your suggestions. And also I want to say that we do use the Archive database to extract some data, stats, etc.

But in this particular case (change the checkmate/king value and then "replay"/redo old games) we need a very good 4PC engine, which we don't have.

Players would change their decisions if the king is +5/+1000, not +20. For example, Red can either capture Yellow +9 queen or Blue king. Now the king is +20 and he took the king. Imagine that the king is +5. He could decide to take the queen instead. And ... from that move you cannot repeat their archived moves, because it would be a very different game.

>Players would change their decisions if the king is +5/+1000, not +20.

Exactly, that is why you need to know first if the King's value will change the finishing positions based on how the games are played now.  Then, does that change affect how some players may team.  A higher value King would not change it in my opinion.  Manually examining many games has showed that a lower value King may change it.

 Almost all of us work with our opposite in the beginning, especially top players, as per Martin0's thoughts of the game.  Currently there are a some players who see nothing wrong with teaming to the end and have rationalized why it is ok.  

Instances of real teaming may be easier to see and be objectively ruled on.  But to turn this addictive game into a casino slot machine with bells and whistles and bonus and negative points just seems like the wrong way to go to me.  I want to play chess albeit 4PC chess. 

We have Teams mode, which is 100% pure chess without a slot machine.

FFA (Solo) with points is a chess-like game, it is here for fun. So, it may have some wacky rules like "1 point queen" or "5 points for double check", or "10 points for stalemate to everyone in this casino, including those who is drunk and sleeping".

@BroncoB this is not a competition. You won't win any prize if your suggestion becomes selected for implementation!

It's always good to laugh and smile.  Thanks for the comments.

Just watched this

Game #417614 started!

Here's a good example of high rated players not teaming to the end.  Would lowering the value of a King change this game?  Thankfully no.

After the first checkmate in the real game with lower King value in parenthesis.

Yellow 11, Green 17, Red 38 (23), Blue 6

After 2nd checkmate: Yellow takes out his opposite.

Yellow 32 (17), Green 17, Red 42 (27), Blue 11

After 3rd checkmate:

Yellow 70 (40), Green 17, Red 42 (27), Blue 20

In both instances Yellow could've resigned before the last mate, which came one move after.  The final standings would still be the same under both systems even with the gifting of points had he resigned.

This game would show up on the data as opposites finishing 1-2 however, but the play seemed to suggest that had blue achieved some points along the way he may have been second.  Perhaps promoted Queens valued at 9 points for instance may have been enough? 

As you can see below, even if the king is +40, +1000 or 0 (!), result of this game is the same. Can we make a conclusion that king value does not matter at all?

King +40:

Yellow 11, Green 17, Red 38 (58), Blue 6

After 2nd checkmate: Yellow takes out his opposite.

Yellow 32 (52), Green 17, Red 42 (62), Blue 11

After 3rd checkmate:

Yellow 70 (110), Green 17, Red 42 (62), Blue 20

King +1000:

Yellow 11, Green 17, Red 38 (1018), Blue 6

After 2nd checkmate: Yellow takes out his opposite.

Yellow 32 (1012), Green 17, Red 42 (1022), Blue 11

After 3rd checkmate:

Yellow 70 (2030), Green 17, Red 42 (1022), Blue 20

King 0:

Yellow 11, Green 17, Red 38 (18), Blue 6

After 2nd checkmate: Yellow takes out his opposite.

Yellow 32 (12), Green 17, Red 42 (22), Blue 11

After 3rd checkmate:

Yellow 70 (30), Green 17, Red 42 (22), Blue 20

smh Wow, that's pretty weak. One game analysis?!?! 

I stated above "thankfully no" because had it changed a well played game then the intent would've be wrong. At least in one lone example wrong.

So my answer is still No, because in games where there was teaming to the end it matters.  As already shown in a couple of cases.  In those cases you can't get enough points to over come the King's value. But maybe they could if per se you changed the value of promoted Queens for instance and lowered the value of Kings. That would make the profitability of teaming less, correct?  I'm going All-In.  Prove me wrong and I will eat humble pie for a year.

Give skillful players a fighting chance in each and every game.  How could anybody be against that?

Bill on double points for opp: "The fear of any coordination with your opposite would make it hard to do anything."

Sounds like it destroys the trust of the teamers, just as bab suggested.