WebFire;1004491 wrote:But the records are determined by the play on the field. How can they not be part of the equation in determining who is best?
You could use the Philly Eagles this year. If you looked at the roster, many thought they were the best team this year? But their play on the field, and hence their record, says they are not the best. How many people would send Philly to the Super Bowl?
The records are, of course, indicative of quality. Here's an example:
You have a coin. You flip it 10 times. It lands on heads 7 times. Would you be 70% certain that if you flipped it again, it'd land on heads? Of course you wouldn't--that'd be stupid. If you flipped the coin a million times, the odds are overhwelmingly good that it would revert to its "true" nature (of 50%). If we flipped the coin a million times and it was heads 70% of the time, then we’d say that we have a weighted coin, and it’s “true heads value” is 70%.
That's what we're dealing with here with teams (but trying to figure out "best" instead of "true heads value"). We've flipped the coin 12 times. If we could flip it a million times and have every time play every other team dozens of times, we’d be able to figure out the exact weighting of the coins. But we can’t. So if the goal is to figure out which coin, when flipped, has the highest chance of coming up heads (that is, is the “best” coin), then the 12 flips are very helpful, and they give us an idea, but they’re too small a sample size to make a definitive statement.
If you flip a coin 12 times, and it comes up tails 12 times, the odds are very, very good that it isn’t a 90% heads-weighted coin. But you can’t be absolutely certain of that based on such a small sample set, either.