If we analyze the playoff seeding vs the true seeding (true seeding being pure W/L record ignoring divisions) to see what true seed is possible for each playoff seed.

1st playoff seed - must have the best record
2nd playoff seed - can be 2 or 3 in win % *see note below
3rd playoff seed - can be 3,4, or 5 in win %
4th playoff seed - can be 4,5, or 6 in win %
5th playoff seed - can be 2,3,4, or 5 in win %
6th playoff seed - can be 4,5, or 6 in win % *see note below

There is a chance that if the three best records in the league come from the same division that a 6th seed has the 3rd best winning % and the 2 seed has the 4th best winning %. I checked and I have never been in a league this happened , I'm sure it has happened a few times but it's a very rare occurance so it is being removed.

Based on the possiblities of the first opponents that the 1st seed and the 2nd seed could possibly play in second round of the playoffs we get this.

#1 playoff seed could play a true 4,5,6,2,3,4,5 seed. Average true seed 3.42
#2 playoff seed could play a true 3,4,5,4,5,6 seed. Average true seed 4.5

So on average the #1 seed will play a much tougher opponent than the #2 seed in the second round series. I'd actually like to see WifS true data on this, but for now this will have to do.

your argument makes sense, but the math is sloppy. like, really, really sloppy.

and by really sloppy, i mean horrendously incomplete and inconclusive. probably the worst statistical analysis i've seen in my life. i would just delete the numbers and try to make your argument in words.

How can the argument "make sense" if the math is "horrendously incomplete and inconclusive?"

I'm actually having a tough time following you too jester. Elaborate a little bit.

while your argument is not lacking hyperbole it is lacking any depth of analysis deanrod.

Actually, bad math aside, there are two things that really make a difference.

1. Match-ups. Who you play is much more important than their seeding. Some teams don't match-up well.

2. I'm pretty sure it's been proven several times that the #1 seed wins more WS than the #2 seed. If you take that into account, I think the #1 seed is what you want.

Historically in HBD #1 seeds win more world series than any other seeds. That is according to two different studies that I have done in the past.

I don't think you need to throw out the possibility of the 6th seed being the third best time by win percentage. Since the schedule is balanced (10 games against every opponent) it's very possible to have the best three winning percentages in one division. The math isn't perfect, but it's a good supplement to the argument.

The idea here, I believe, is that the fourth seeded team (worst division winner) can only be one of the bottom three teams to qualify for the playoffs. Advantage to the one seed, if the worst division winner wins. However, the 5th seeded team (best wild card) can range from 2nd best to fifth best team to make the playoffs, despite being the fifth seed.

That means that the 2nd seeded team can play, best case scenario, the worst playoff team (second wild card) and at worst the third best playoff team (third division winner). The 1st seeded team can also play the worst playoff team (if the worst division champ is the worst playoff team). On the other end, they can play the 2nd best playoff team, if that second best team was in the same division.

The logical conclusion would be to do what the NFL and NFL do and re-seed the teams after the wild card round. Not perfect, but it keeps the option of playing the 6 seed in the second round open for the number 1 team in each league.

Quote: Originally posted by jimmystick on 12/05/2009How can the argument "make sense" if the math is "horrendously incomplete and inconclusive?"

I'm actually having a tough time following you too jester. Elaborate a little bit.

conceptually, his argument is semi-feasible. but the numbers are so unscientific that it's pointless to even include them.

Quote: Originally posted by MikeT23 on 12/05/2009Actually, bad math aside, there are two things that really make a difference.1.  Match-ups.   Who you play is much more important than their seeding.  Some teams don't match-up well.  2.  I'm pretty sure it's been proven several times that the #1 seed wins more WS than the #2 seed.   If you take that into account, I think the #1 seed is what you want.

this is probably cause 1 seeds are usually the better team, and also i don't think that there's any significant difference in the matchup quality by seeding.

although one (among many) things that jester neglected to account for is that if a 77-85 team wins their division and a series, the 1 seed gets to face them in round 2.

Where it really makes a difference is when the two best teams are in the same division. The team that finishes second is the first wild card and gets the worst division leader. The two best teams then end up playing in the second round.

A lot of it depends on how evenly the strong teams are spread out among the divisions.

Quote: Originally posted by WiredTiger on 12/05/2009Where it really makes a difference is when the two best teams are in the same division. The team that finishes second is the first wild card and gets the worst division leader. The two best teams then end up playing in the second round.

A lot of it depends on how evenly the strong teams are spread out among the divisions.

That's actually the key. The 5th is the true "wild card" - as they can be the second best team in the league. The #2 seed at best never has to play the 2nd best record in the league.

Think of it this way. If you got to choose which seeding you got and your choices were

Team A could play a team seeded 2,3,4, or 5
Team B could play a team seeded 3,4,5, or 6

I would choose team B's schedule every time. The best seed you could play would be a 3 as opposed to a 2 and you have a chance of playing the worst seed.

have you considered that there are different probability distributions for each possible outcome?

if you want to actually come up with a meaningful conclusion, find the average win total of the DCS matchups for each seed. otherwise, you have nothing more than a valid but untested hypothesis.

Quote: Originally posted by deanod on 12/05/2009have you considered that there are different probability distributions for each possible outcome?

if you want to actually come up with a meaningful conclusion, find the average win total of the DCS matchups for each seed. otherwise, you have nothing more than a valid but untested hypothesis.

Well the chances of the 2nd best record in the league being in the the same division as the best record is 1 in 4. The chances of the 3rd best record in the league being in the same division as the 1st or 2nd seed is 1 in 2. In either of those cases the #1 seed has a chance of playing the tougher team.

So the #1 playoff seed has a 1/4 chance of having the true #2 seed on it's side of the bracket while the #2 playoff seed has 0% chance of that happening.