March Madness - Round 2

[Yellow_Dog]

The curse of Don Mattingly has struck. I'm seeded in the same division as Panzer Ace for the 2nd round. My locker room will have Rudy, Hoosiers, and the Bad News Bears playing on a loop. I guess a top 6 performance is all I need to aim for to advance to the third round.

[roofinghorse]

do we not need 8 out of each bracket to fill 24 teams for the 3rd round?

[Jablowmi]

crossing, it's the 4 playoff teams in each bracket (16) plus 5 additional teams from each bracket (20). Bottom 3 in each are out.

[IM Bacchus]

Have the Round 2 Leagues been set up yet?
Just looking to get an idea of when we need to have our teams ready.
IM Bacchus

[Jablowmi]

IMB, they're being set up now. Hope to start either a week from Monday or, at the latest, two weeks from Monday.

[toobig47]

I started up bracket 3
sent password to

YOUNTFAN

UGLY JERRY

HECHOJAZZ

[Semper Gumby]

[quote:0af49bb6a0]You can easily argue the rankings of the managers because of SOM stupid ranking system but the top seeds should have the easiest route to make it to the next round.[/quote:0af49bb6a0]

Agreed.

And, even more reason the initial ranking system has its issues.
Otherwise, I recall the NCAA tourney seeding also accounts for a top ranked Big East's teams playing schedule versus a top ranked team from the Atlantic 10 vs. the number of games you played :D

Nonetheless, [b:0af49bb6a0]I'm ranked on the bottom so I neither win or lose from the adjusted seeding [/b:0af49bb6a0]but if I had a Top 12 ranked team - I'd surely care about how SOM's rankings are far from being the end to all questions.

If anything, EVERY NCAA Tourney comes with some idiot sports writer belly aching over why Platzburgh Community College ain't ranked higher!

:D

[TheGoodDoc]

I actually had the same concerns, too, about the SOM manager rankings thinking that it had to be highly related to the number of teams one had. Obviously, TSN is going to positively reinforce its return customers with higher ratings. It's just good business. In psych terms, rat presses lever, rat gets food. Repeat. Rat presses lever, rat gets electrocuted. rat moves on.


I decided to do some statistical analyses on this question. Using the top 100 manager's rating data, the correlation between the # of teams played and TSM rating was correlated .82, which is a very strong correlation and was to be expected based on the equation that calculates ratings (i.e. you have to pay to play and you have to play to win).

However, the most important predictor of ratings is # of playoff appearance (correlates .98) . Given the fact that the # of teams played and # of playoff appearances are correlated .79, i ran a partial correlation between #of teams and Ratings, controlling for the number of playoff appearances that were made. When this is done, the correlation b/w teams and rating drops from .81 to .38. That is a big drop, but the pure number of teams owned seems to still contribute to rating in a significant manner. Long story short, winning isn't everything, but it means alot more than just playing. yet, the mere fact that you sign up for a team gives you points towards your rating.


That said, maybe another way to look at manager's performance or effectiveness is to compare the # of championships (or playoff appearances) won to the rate of winning a championship by pure chance. Assuming all things equal, a manager has 1 in 12 chance to win the league he/she is in. Thus, to determine a manager's "efficacy", one might be able to use this formula:

Manager Efficacy = # of championships /(.083 * # of teams played).

You could also create a performance score based on playoff appearances, substituting .33 (% chances to make the playoffs at random) in place of .08 in the above eqn. These ratio would give a give you an estimate how well a manager has performed controlling for # of teams played compared to dumb luck.

OR, you could create a "Total performance" index like this:

total performance= Champisonship performance (efficacy) + Playoff performance (efficacy) + Regular Season Performance (efficacy), which would equal {.50 *(162 * # of teams) i.e. winning over 82 games in a season}.

This should give you a complete magnitude measure of performance against chance that would be sensitive to performance in each leauge one plays in.

Interestingly, neither of these performance ratios (based on championships or playoffs) significantly correlated with TSN ratings. hmmm...


One disadvantage is that it doesn't take into consideration competition level, but then again, neither does the present system. If say, one of the SOM 80's gurus played with all newbies, expertise can go a long way and i would bet that they would stand a good chance at winning the playoffs. Some type of competition level weighting would be helpful. another disadvantage is that this scale may only be valid after a certain number of leauges played, maybe 10 or 12 games. For instance, If User X gets lucky his first year playing, wins 100 games and a championship, his strength index would be extremely inflated. Is he/she better at SOM than Penngray, for instance? probably not. Just like waiting for that 6th Start or the 100th AB to determine whether the card is worthy :-)

so, who's the best SOM manager in the top 100 according to these ratios? Let's just say I wouldn't want to share a division with Panzer Ace . he's one wicked Strat manager :-) However, DSM925 (ranked #82) has the highest strength rating - 5 championships in 8 trys-- pretty damn good.


Ok, i REALLY need to work. later

[roofinghorse]

teams in so far are,

bracket 1----0 teams
bracket 2----1 teams
bracket 3----3 teams

and doc,that was a lot of work and time and i for 1 appreciate it for i just don,t have the time to look up figures and calculate every variable such as you did....and the same goes for the jablowmis and yount fansand all of the other managers whom put so much effort into sharing their knowledge of this strat game.thanks fella,s.i appreciate it.
slug

[UGLYJERRY]

So let me get this straight. The top [b:b49acab35e]8[/b:b49acab35e] teams from each bracket advance to the 3rd round, correct. 3x8=24, 12 teams per league, I thought that I read 9 per bracket for this round. Just making sure.

Ugly