crockpot's baseball picks 4/1-4/7 (incl futures)

[crockpot]

added 1u of Was +144

[crockpot]

[ QUOTE ]
i like ALL of your picks EXCEPT

tampa and baltimore future bets

3 teams would have to completely underachieve (3 teams in the same division) it just seems impossible

can you explain those bets more

[/ QUOTE ]

first of all, the blue jays do not have to underachieve to stay out of the divison race. in fact, they did a fine job of it last year despite greatly overachieving.

beyond that, each of the d-rays and orioles has roughly a 3% chance of ending up in the 92+ win range. if they get there, it will likely mean the yankees and red sox underachieved, since these things are correlated, and they are likely to take the flag.

i'm not saying these things are at all likely, but they can happen. i have no qualms about people who choose to pass on bets like this as long as they also don't play the lottery or single numbers at roulette.

[Post-Oak]

[ QUOTE ]
beyond that, each of the d-rays and orioles has roughly a 3% chance of ending up in the 92+ win range.


[/ QUOTE ]

This is not even close to being true.

I didn't read that article you referenced earlier concerning variance, because it seems you have to pay to read it, so I'm not sure where you are getting these numbers.

A true .500 team would have a 4.9% chance to win 92+ games. A true .450 team (72.9 wins) would win 92+ games 0.17% of the time!

In order to have a 3% of winning 92 games or more, you would have to have a true winrate of over 49%. Tampa Bay is not a .500 team. Your numbers are not even close to being reasonable.

[Post-Oak]

[ QUOTE ]
some 2007 props i like and have bet heavily:

White Sox under (insert number) W
Diamondbacks over 78.5 W
Dodgers under 88.5 W
Blue Jays under 87.5 W


[/ QUOTE ]

I also love the CWS and LAD unders.

The other two I wouldn't touch, although I lean over for the D-Backs.

[crockpot]

[ QUOTE ]
This is not even close to being true.

I didn't read that article you referenced earlier concerning variance, because it seems you have to pay to read it, so I'm not sure where you are getting these numbers.

A true .500 team would have a 4.9% chance to win 92+ games. A true .450 team (72.9 wins) would win 92+ games 0.17% of the time!

In order to have a 3% of winning 92 games or more, you would have to have a true winrate of over 49%. Tampa Bay is not a .500 team. Your numbers are not even close to being reasonable.

[/ QUOTE ]

okay. now go show me some projections that have the standard deviation you used, 6 games and change. go ahead, i'll wait.

[Post-Oak]

That article you read has you really confused.

It's impossible to know a team's true winrate throughout the course of a 162 game season, for many obvious reasons. We all understand this. You don't understand the ramifications of this fact though.

Hint: It doesn't mean that there is at least a 3% chance that TB will win 92 games or more. Again, that is laughable.

[New001]

[ QUOTE ]
That article you read has you really confused.

It's impossible to know a team's true winrate throughout the course of a 162 game season, for many obvious reasons. We all understand this. You don't understand the ramifications of this fact though.

Hint: It doesn't mean that there is at least a 3% chance that TB will win 92 games or more. Again, that is laughable.

[/ QUOTE ]
What was the % chance that Detroit wins 92 games one year ago?

[knicknut]

You're forgetting games are correlated, Post-Oak.

If Tampa Bay goes .600 through their first 50 games, you think they'll still be .450 to win a game after that? No, odds are they're actually better than .450.

[Post-Oak]

[ QUOTE ]
You're forgetting games are correlated, Post-Oak.

If Tampa Bay goes .600 through their first 50 games, you think they'll still be .450 to win a game after that? No, odds are they're actually better than .450.

[/ QUOTE ]

I'm not forgetting anything. I already realize what you, New and Crockpot are thinking.

Let me ask you this - what if TB is .250 after 60 games? Maybe they're not a .450 team then?

Again, it is possible that TB is the best team in the majors right now. Not at all probable, but it is possible.

That caveat doesn't change the fact that to believe that TB has a 3% or better chance to win 92+ games, you have to believe RIGHT NOW that they are an 81 win team on average. That's just stupid.

I am willing to make a prop bet on this for this season. I will take a basket of teams to NOT win 92+ games, and will pay you better odds than your crazy calculations would suggest.

[crockpot]

[ QUOTE ]
That article you read has you really confused.

It's impossible to know a team's true winrate throughout the course of a 162 game season, for many obvious reasons. We all understand this. You don't understand the ramifications of this fact though.

Hint: It doesn't mean that there is at least a 3% chance that TB will win 92 games or more. Again, that is laughable.

[/ QUOTE ]

okay, we'll try again. what are the chances tampa bay is a true 73 win team this year? how about 75? 69? 79? these are all possibilities and form a weighted probability distribution. note that since an 80 win team and a 66 win team will have a much better combined chance of a playoff appearance than two 73 win teams, this is a very good thing for them.

what about the injuries, callups, trades they'll go through this year? you can't project how those will shake out. that adds even more variance to flatten out the distribution and skew it further towards the extremes.

what about the fact that even with perfect information about a team's winning percentage, the standard deviation of their W-L record will still be around 6.3 games?

adding even more variance is that tampa's roster is full of guys without established performance levels. with their performance being more uncertain than the average team, they have more breakout potential.

when you combine all these forms of deviation around the mean, and i could add more if i wanted to be nit-picky, you get a LOT of variance. there's a thread on tangotiger's blog here discussing the effects. this is why the best projection systems will have a standard deviation of around 9 wins, much like they will never achieve an r of above .75 for individual players.

teams finish far off from their projected level all the time. last year's tigers beat their vegas O/U line by 18 wins. the twins finished 16 ahead of theirs, padres 11. the cubs fell 19 games short, the indians 14, cardinals 12, braves 11.

rewind to 2005. what were the odds of the white sox winning 99 games? how about the indians being the best team in baseball? this stuff happens every year. assuming a true standard deviation of 6.3 games, the probability of all these teams over/underperforming is extremely low; you have several 2+ SD outliers every single year. with a 9.5 game deviation, the results fall in line with expectations.

does all this mean tampa will win 92 games 3% of the time? i can't say for sure; this is just what my numbers tell me, and the numbers are going to be wrong every once in awhile. but there's no way their chances are as low as you claim, and it's ridiculous to suggest that anyone can predict a team's win total with that degree of precision.

if you want to attack the projections, that's fine, because those are a matter of opinion. the methodology is not, at least not to the extent where i'm going to be convinced your way is better until you show me some projections that don't have these kinds of outliers and consistently have a 7 game SD. and even when you do that, i still will have based my bets on projections that have demonstrated less accuracy in the past.

now, aside from that, i'm very happy the nationals won.