An ABLE System: toward league fairness....

[mattgslater]

See, I see farming noobs as part of the game. If nothing else, freakin' teach the noobs to farm!

I mean, how are you going to learn how to game your blocks and scores, or set up a comp rotation, unless you watch someone do it to you?

I think the instinct to tailor SOS to success is a good one: check out my fixed-league thread again, and you'll see that each team plays two extra games against teams at similar W/L record. But if bottom feeders always play bottom feeders, they'll develop each others' bad habits.

The players aren't the only ones who build skills!

[Digger Goreman]

Agreed to a certain extent.... I'm counting on a posted "Ladder" encouraging the "bottom feeders" (your words, not mine) to break out and take a swipe at fame....

In the past three leagues, it's been a restart every time (while supposedly maintaining "Legacy" teams for additional play) so teams are somewhat retired each season.... In fact, I keep all my old teams in a file for just such occassions.... On the one hand, I don't want some team sitting there, at #1, forever... unassailable... taking up a slot.... I'd prefer to throw a little confetti into the air and tell the coach what a great strategist (s)he is, make a note of greatness at the bottom of the chart, and move a new dog to the top of the pile....

[mattgslater]

Agreed there. I think you need to reset each season, though you can certainly include a mechanism to allow some players to be carried over.

[Fat_Emrys]

I think I follow the logic, but what happens if teams are level (e.g. two starting teams)?

[Edit: removed a lot of rubbish which was based on a misunderstanding of the "rule of thirds". My bad.]

[Digger Goreman]

I think I follow the logic, but what happens if teams are level (e.g. two starting teams)?

[Edit: removed a lot of rubbish which was based on a misunderstanding of the "rule of thirds". My bad.]

You would have two "0"s divided by each other... which is undefined and opens up a wormhole, swallowing all teams except chaos....

or...

I should be starting the teams at a Rating of 100(?) and make a note that, in the above case, no points are gained/lost.... A truly equal outing (except maybe for gains in spp)....

[Digger Goreman]

Just playing with the numbers in the formulas:

A 1,000,000 team takes on a 1,200,000 team.... The following possibilities arise...

Underdog wins and gains 1.2 points in rating; overdog loses same

Underdog ties and gains 0.8 points; overdog loses same

Underdog loses and drops 0.8 points; overdog gains same

A maximum swing of 2.5 points on a 200,000 tv difference doesn't seem like much....

[Lictor]

See, I see farming noobs as part of the game. If nothing else, freakin' teach the noobs to farm!

Wow, we have a very different attitude here. You could almost apply some kind of 'growing the seed' metaphor to how we treat noobs.

What could be 5-0 games are often 4-2 or 3-1 games because we show them how to play.

This is 2 out of 3 leagues that I play anyway. I am also in one that is argueably a 'Premier' league but where I recently coached a very new player to my own detriment. What whould of been a comfortable 3-0 win to me turned into 1-1 because it was obvious he needed the help and I was silly enough to give it to him (and I also thought with rolls like "5+, 4+, 3+ dodge here, 2 GFI's and a 2 die against blitz is the best option you have if you don't want me to score next turn" were safe )

This attitude is because of my mentor and the other two leagues that I am in. I think if I play in your league all your noobs would love me lol.

[Fat_Emrys]

Underdog wins and gains 1.2 points in rating; overdog loses same

Underdog ties and gains 0.8 points; overdog loses same

Underdog loses and drops 0.8 points; overdog gains same

How about replacing the tie result with:

Underdog ties and gains 0.6 points (half of what a win would have been worth), overdog gains 0.4 points (half of what a win would have been worth) so the tie is more valuable to the underdog but the overdog doesn't lose points unless they win (which seems harsh to me when the difference between overdog and underdog could be as little as 10,000gp).

[Digger Goreman]

... so the tie is more valuable to the underdog but the overdog doesn't lose points unless they win (which seems harsh to me when the difference between overdog and underdog could be as little as 10,000gp).

The above was from a specific example... let's see what a +10,000 game (1,000,000 vs 1,010,000) looks like:

Underdog wins and gains 1.01 pts; overdog loses same.
Underdog ties and gains 0.68 pt.; overdog loses 0.66pt (rounding errors)
Overdog wins and gains 0.99 pt; underdog loses same.

Maximum swing of 2.02 points... after all, they ARE nearly equal and the risk is minimum (in a theoretical sense).... This might encourage a little more risk taking by the underdogs to try and bite off a little more for a bigger reward....

Remember that the lower Ranked teams issue the challenges and the only way to "challenge up" to a lower TV team is for the lower TV team to have a higher Rank.... This has the potential to account for teams that are better at certain TV levels and NOT protect teams from challenges by intentionally dumping TV....

Thanks for the questions... they are definitely helping me "talk out" the concept and scrutinize it better....

ps: another good thing about the lower ranks issuing the challenges... higher ranked teams don't get to prey on stunties and noobs!

[mattgslater]

Wow, we have a very different attitude here.

Frankly, I don't think it's all that different. We teach our newbies the game too, and often have closer games for the same reason. But that's what preseason is for. I'd never let somebody into the regular season who hadn't played at least half a dozen games (for oh so many reasons), and by that point they should have a basic grasp. But in my book you're a newbie until you've won a playoff game, and there are a lot of "noobs" who have a lot to learn but know the basic game. This is the time when winning or losing on your own merits is good for the soul. Until then, it's all about just learning the game, and really green coaches shouldn't be allowed anywhere near games that count.

My objective is to build coaches who can play me. Part of that is explaining exactly how what doesn't work doesn't work, and playing it out is the best way to do that. I'm totally cool with helping a new coach with tactical decisions ("no, don't blitz yet; move that guy first so you get two dice and have a TZ in case you skull out," that sort of thing), but if you want to get good at this game, you can't do it without taking your lumps on your way up, so you have to at some point stand or fall on your own.

[Hache]

It's still a bit strange for me to mix 2 ratings (in your actual system) :
- ladder level determines which is the upper/underdog
- TV determines the gain/loss of ladder level

I would not mix the 2 ratings :
- I consider that inducements are good enough to get 2 rosters to be equal. TV should not interfere with ladder level.
- the real 'Team Strength' difference between 2 teams are the coaches. Therefore, the ladder level only should determine the gain/loss of ladder level.

[Carnis]

Also, in this system, there is only the perpetual ladder and no "play-offs", though being open to a bowl tourney is always a possibility....

Then this is your problem.

The top teams are never encouraged to play each other, only farm noobs. A seasonal play-off every 3-6 months, winner of which is the league winner would mean a newbie farmer would not automatically win every time.

[MrBister]

ELO ranking is supported in OBBLM and it's similar to what you want to achieve, Digger. Though it's not based on TV diff but on your rank in the league. It's (kind of) easy to adjust starting rank and how much your rank goes up or down after a game.

I had some thoughts on how to weigh in TV difference in the ELO rank but none of the maths worked out fine so far. Mostly has to do with the problem that one race may benefit more from having for example 100k inducements than a different race. (By getting that good star player who helps your team a lot.)

[Darkson]

There is no fairness in a game based on dice and progression - there is only the illusion of fairness.

Want a semi-fair game? Play a game where there is a fixed points value, using the same list.

Want a fair game? Play Chess.

[Digger Goreman]

It's still a bit strange for me to mix 2 ratings (in your actual system) :
- ladder level determines which is the upper/underdog
- TV determines the gain/loss of ladder level

Rank determines who you can challenge... that's all....

TV determines upper/underdog (same as rules)

[quote="MrBister]ELO ranking is supported in OBBLM and it's similar to what you want to achieve, Digger....[/quote]

My only exposure to ELO is the wiki article.... Is there a site with an easier explanation/applet, etc., that is usable?

*

*snipped for non-constructive waste of band-width*

Carrying on....