de-derailing this thread:

http://www.wargear.net/forum/showthread/3821p1

What qualities would be reflected in an ideal ranking system?

Amidon37 wrote:

What qualities would be reflected in an ideal ranking system?

Which ranking system? Are you asking about a system that ranks play on a specific board? all boards? team play? etc. an aggregate? General/Colonel as an aggregate?

If I'm reading your question right, I agree that the best way to approach the problem is to look at the big picture, and though I'm not sure which 'angle' is the best, I'm inclined to say we should be looking from the top down.  I.e., what should WarGear's most coveted rank/award reward?

It's the most coveted rank/reward

It encourages play across multiple boards

It's 'fair' in the sense that It reflects the difficulty of playing both the board and the players that play it

It's relatively easy to move around on the rank list (i.e: I don't have to play for a month just to get a point)

It doesn't necessarily lock a player into a position (the top spot can be overtaken, and a player will see movement on the ranking if they play more)

It's easy to understand and to explain (to know how to get points)

It reflects an intuitive approach to getting points (i.e: the user should be able to think about the system and say "if I do x, I can get points" - where "x" is not something unrelated or bizarre

It should be easy to pin the current badge system onto

It is current - "if you haven't played in a while you will fall off the list" (this is debatable I think)

It can be stored, tracked, reported, and manipulated easily by the system or the system admin

It can be visually represented

What am I missing?

Ah, and:

it takes into account a variety of play styles, approaches and habits - i.e: it rewards those that like lightning games, and those that play alot of team games, and includes tourneys and... etc.

Great list ratsy -

As a corollary to "It encourages play across multiple boards" and "it takes into account a variety of play styles, approaches and habits" one big detriment for what we do now that I see is that the GR/CP system favors success in games with larger numbers of players.  

Amidon37 wrote:

Great list ratsy -

As a corollary to "It encourages play across multiple boards" and "it takes into account a variety of play styles, approaches and habits" one big detriment for what we do now that I see is that the GR/CP system favors success in games with larger numbers of players.  

And why do you think that is? I think it's because of the inevitable dropouts, and the dynamics of a large game tends to favor patiently hanging back while others fight, further decreasing the number who actually have a chance of winning. So in a 16-person game, typically it's 8 players fighting for the points of those 16. 

It's cause you get x points by beating a player. You get 2x by beating 2, and 3x by beating three. So you get alot more points for winning bigger games.  (but your right, barring bad placements or early targeting, it's better to wait to compete on the really big ones)

Amidon37 wrote:

Great list ratsy -

As a corollary to "It encourages play across multiple boards" and "it takes into account a variety of play styles, approaches and habits" one big detriment for what we do now that I see is that the GR/CP system favors success in games with larger numbers of players.  

I don't have a ready answer to your question, but as a compliment to the list of what qualities we want in an ideal ranking system, it might be useful to think about what we don't want (e.g. a system that favors games of particular size).  It's congruent to thinking about what we want, but it might help to look at it from the opposite side too.

So maybe a perfect system makes allowances for number of players.  

Similar to what you get now, but averaged. 1x for one player, 2x/2 for two, 3x/3 for 3.

You'd still get more points for better players, but there would be no advantage to winning a bigger game... 

ratsy wrote:

It's cause you get x points by beating a player. You get 2x by beating 2, and 3x by beating three. So you get alot more points for winning bigger games.  (but your right, barring bad placements or early targeting, it's better to wait to compete on the really big ones)

Yeah, but, the points is proportional to the players you beat. So the smaller game is supposed to be easier to win, ideally, in proportion to how small it is.

I get that this isn't true, but it's not for size, it's for how size changes the game.

True. Smaller board, more marked the difference in skill level between players - the littles skill difference will show in the victory on a head to head board. 

Larger board, you need a much bigger skill differential to predict the victor. 

So right now we allocate points at the player level, what I just proposed allocated them at the game level, but maybe we should really be trying to allocate them at the level of skill. Points based on ranks... or possibly on H rating... or something less obvious... 

Great summary Ratsy. This is all in another thread somewhere right? Anyway, maybe another way to frame big map (16 player) wins is to think about it as fighting 15 1v1's (And, it takes about as long).  The fact that a shrewd general can win half of those by just by not dying is part of the deal. 

Oh, and read but don't post if you agree to change CP!

Amidon37 wrote:

..one big detriment for what we do now that I see is that the GR/CP system favors success in games with larger numbers of players.  

I don't see how larger games are inherently easier to win. Any given player is will win 1 out of every 5 games in a 5-player games, so it makes sense that he gets twice as many points as he would for winning a 3 player game.

Sure, patience is a virtue in larger games, but conversely, more aggressive play earlier wins smaller games.  A large game full of experts won't go to the most conservative/patient player. Rather, it will be won by the player who correctly senses when it's the right time to play aggressively.  Better players may enjoy a advantage in larger games but it is only because they better understand and take advantage of this dynamic. My logic may be flawed, but right now I think weighting points in a regressive fashion would be a mistake.

M57 wrote:

Amidon37 wrote:

..one big detriment for what we do now that I see is that the GR/CP system favors success in games with larger numbers of players.  

I don't see how larger games are inherently easier to win. Any given player is will win 1 out of every 5 games in a 5-player games, so it makes sense that he gets twice as many points as he would for winning a 3 player game.

Sure, patience is a virtue in larger games, but conversely, more aggressive play earlier wins smaller games.  A large game full of experts won't go to the most conservative/patient player. Rather, it will be won by the player who correctly senses when it's the right time to play aggressively.  Better players may enjoy a advantage in larger games but it is only because they better understand and take advantage of this dynamic. My logic may be flawed, but right now I think weighting points in a regressive fashion would be a mistake.

You got to know when to hold'em, know when to fold'em, know when to walk away...

Some stats on my 2-player game discussion.

http://www.wargear.net/forum/showthread/3823p1

ratsy wrote:

So right now we allocate points at the player level, what I just proposed allocated them at the game level, but maybe we should really be trying to allocate them at the level of skill. Points based on ranks... or possibly on H rating... or something less obvious... 

My understanding/hope was that a trueskill type of ranking would accomplish this.

me too. I'd like to see it tried if possible. (that's right, i'm volunteering hugh to replicate the work of a giant corporation...)

me too. I'd like to see it tried if possible. (that's right, i'm volunteering hugh to replicate the work of a giant corporation...)

To combat the tendency for For large multi-player to have more point value, there could be a game option to choose a standard (winner take all) or elimination (points awarded based on who is eliminated by who) reward style.  Then a more aggressive player wouldn't be playing them self out of the points so long as they can get an elimination or two before they themselves are eliminated.  Plus, patience and waiting for everyone else to fighting it out so as to pick up the scraps would be less viable for maximizing point getting in such games.  So, overall, such a reward option would encourage a more aggressive play style throughout. and be more equitable at rewarding points as it would be more merit based.  For those that prefer the patient strategy they could still join and create standard, winner take all style games.