Warrior General Games

Originally posted by Corndog
Warsaw is 2
Dakota is 1
MCM is 1

there are so many issues with this. 1-2 should never be greater than 1-1 it basically damns dakota for not playing music city or warsaw twice which was not in our power

Originally posted by Corndog
Using win percent instead of total wins favors out of division teams, which I'm not convinced is much more fair.

That is not always accurate (win percentage favoring out of division); where as total wins favors in division opponents:

Assuming each match is a coin flip (50/50)
4 instances has an expected value of 2
3 instances has an expected value of 1.5
2 instances has an expected value of 1
1 instance has an expected value of .5

What you are pointing out, with all due respect is gamblers fallacy; the baseline odds don't change.

The issue is inconsistency, because it depends on the # of teams tied and the divisions they are in right now. Plus the H2H is not sorting quality wins only those against other teams you are tied with.

Suppose Constellations wanted to play the Simpsons instead of Wolfpack in the opening round (I doubt they do but suppose they did) they could have thrown week 14 against a team not included in the tie without penalty.

Suppose we had beaten Ronin one of the two teams (An occurrence that has NOTHING to do with Warsaw and MCM) - how is it that without us tied MCM would be the division winner; but with us in the tie they drop?


Running a sort #1 declaring division winners, then running a second sort for wild card teams clears up this problem.

Originally posted by william78
The issue is inconsistency

I don't know why you keep using this word.

It literally has always used the same exact code and same exact tiebreakers for all situations since GLB2 launch.

Not sure how that can be "inconsistent".

Originally posted by Corndog
Why does removing divisional winners make more sense?

Should we also remove wins and losses against divisional winners?

Because they aren't in the mix. Why should wins against teams that aren't being considered for the wild card count for more than wins against teams that aren't in the wild card hunt? Again, my team had 12 wins, so my team is the #1 seed, but your calculation doesn't count wins against my team to be more important than wins against other teams that are tied.

http://glb2.warriorgeneral.com/game/league/79

Some divisions are harder than others. There's one free win in Alpha and two tough games each time you play in division.

Here's what I suggest:

1) Run the script across each division to determine the divisional leader.
Looking at Taurus, that means that the divisional leaders would be BCM, Whoville, and TT (no change from your method).

2) Remove divisional leaders and run again for the whole league

Now the only teams with ties are Simpsons and Wolfpack who could possibly be in the Wildcard hunt. Simpsons beat Wolfpack, they therefore take the H2H.

I mean, if you are going to consider "wins against top teams" to be deciding factor, the alternate step would be this one:

2b) Instead of comparing H2H against the tied teams as the third tie-breaker, compare the winning % against all tied teams and all playoff teams.

I mean, that would be the ultimate "does this team stand a chance to win as the wildcard" test.

Originally posted by USC_Trojans
there are so many issues with this. 1-2 should never be greater than 1-1 it basically damns dakota for not playing music city or warsaw twice which was not in our power

1-1 isn't greater than 1-2, it is equal. That's why they moved on to the next tie-breaker and the next tie-breaker.

Originally posted by Corndog
I don't know why you keep using this word.

It literally has always used the same exact code and same exact tiebreakers for all situations since GLB2 launch.

Not sure how that can be "inconsistent".

It is inconsistent because the variables that it is making its decisions on are inconsistent. GIGO

It is based on the # of teams tied (thus changing the head to head) even if the # of teams tied has nothing to do with playoff spot. For example taking BCM(A division winner) out of the tie means Simpsons goes in even though Simpsons and Wolfpack resolution is different than BCM versus Simpsons. Example my Testudo team beat BCM early in the season - if you remove that loss BCM is in the playoffs and the Simpsons are as well.

Inconsistent also because the head to head variable is a measure of absolute number of wins over percentage giving a greater chance for any intradivisional matchup to win a multi-way tie.

Originally posted by Time Trial
Because they aren't in the mix. Why should wins against teams that aren't being considered for the wild card count for more than wins against teams that aren't in the wild card hunt? Again, my team had 12 wins, so my team is the #1 seed, but your calculation doesn't count wins against my team to be more important than wins against other teams that are tied.

http://glb2.warriorgeneral.com/game/league/79

Some divisions are harder than others. There's one free win in Alpha and two tough games each time you play in division.

Here's what I suggest:

1) Run the script across each division to determine the divisional leader.
Looking at Taurus, that means that the divisional leaders would be BCM, Whoville, and TT (no change from your method).

2) Remove divisional leaders and run again for the whole league

Now the only teams with ties are Simpsons and Wolfpack who could possibly be in the Wildcard hunt. Simpsons beat Wolfpack, they therefore take the H2H.

I mean, if you are going to consider "wins against top teams" to be deciding factor, the alternate step would be this one:

2b) Instead of comparing H2H against the tied teams as the third tie-breaker, compare the winning % against all tied teams and all playoff teams.

I mean, that would be the ultimate "does this team stand a chance to win as the wildcard" test.



IF you are going to do 2B it should be run against all teams with 10 or more wins , because only running it against ties is a bit goofy and it still favors teams who play good teams more often.

Originally posted by Time Trial
1-1 isn't greater than 1-2, it is equal. That's why they moved on to the next tie-breaker and the next tie-breaker.

it is when you look at losses as the next tiebreaker. 1 loss is better than 2, so 1-1 is in, 1-2 is out.

it should not be a tie breaker if the 3 haven't played each other an equal amount of times.

It's like saying 2-4 is the same as 2-0

Originally posted by Aeir
it is when you look at losses as the next tiebreaker. 1 loss is better than 2, so 1-1 is in, 1-2 is out.

it should not be a tie breaker if the 3 haven't played each other an equal amount of times.

this. if we just took out division winners for the wildcard factor this wouldnt even be an issue

Originally posted by Aeir
it is when you look at losses as the next tiebreaker. 1 loss is better than 2, so 1-1 is in, 1-2 is out.

it should not be a tie breaker if the 3 haven't played each other an equal amount of times.

It's like saying 2-4 is the same as 2-0

TY well said. 2-4 is not equal to 2-0 except in this formula. Or

Far more likely
3 Teams tied in one division (6 games)
1 Team tied in a separate division

The 3-0 team has the same record for this as the 3-3 teams?

Originally posted by Corndog
Why does removing divisional winners make more sense?

Should we also remove wins and losses against divisional winners?

It makes total sense to remove all teams that are not part of that tie. So, if you have three teams with the same record and one gets 'eliminated' by making the playoffs as the division leader, you have to remove that team's stats from further tie breaking procedures. At least the NFL does that, they only use the H2H wins and losses between the tie breaking teams.

Today has caused my Head 2 Hurt

H2H needs to be changed so that it calculates based in win% instead of win total or the code needs to be changed so that division winners are first calculated and then wild card winners are calculated on a completely separate tiebreak only using H2H between the teams vying for the wild card.

No system is going to be perfect in everyones eyes but either of those 2 changes will make it a much better system than what it is now.

My teams don't lose league games so I can be considered an objective 3rd party

Just get rid of the divisions already, they are pointless and only serve to give an unfair advantage to teams that luck into having a CPU/inactive team in the division with them (and also create this crazy playoff tiebreaker mess).

Can anyone name me even one advantage to having divisions? (And don't say rivalries, we've been in 4 leagues in 4 seasons and considering there are 2 CPUs and an inactive in our current league it'll probably be 5 for 5 for us.)

Originally posted by Danthesportsman
Just get rid of the divisions already, they are pointless and only serve to give an unfair advantage to teams that luck into having a CPU/inactive team in the division with them (and also create this crazy playoff tiebreaker mess).

Can anyone name me even one advantage to having divisions? (And don't say rivalries, we've been in 4 leagues in 4 seasons and considering there are 2 CPUs and an inactive in our current league it'll probably be 5 for 5 for us.)

This is what I recommended a few pages ago if they will not change the coding for tie-breakers. It will also solve the #1 seed playing a #4 that could potentially be the 2nd best team. Common to see the #2 seed get a "bye week" because the #3 is at times the 6th+ worst team in the league.