MLB 4-Play Monday (Fade Me Now)

20Four7 · 06-02-08, 09:18 AM

I did notice the BPS is in serious fade mode. When you tracked it do you recall what the worse downslide was? While the weekend wasn't good a 1 and 9 record isn't the end of the world (yet).

strictlywinners · 06-02-08, 09:48 AM

LT don't start fading yet it is about to turn around. I love the DBACKS and Mariners plays today, they will both be included in my systems plays. I'm just waiting to post cause I feel like I will be able to get a better line later.

buztah · 06-02-08, 10:06 AM

Been in a slump too. Also on the dbacks. Guess that means the Brewers are max wager material. LOL! Either that or TLS and Mr. Big Shot both breakout today and the Dbacks win 18-2. Hopin' for the latter. GL, pal!

fifawcs · 06-02-08, 10:59 AM

I like the Mariners, but I can't agree with you on the other two. Davis was pretty bad in his last outing and the Dbacks cannot hit on the road. Also, Suppan has been dominant at home, posting a .93 ERA. The Brewers offense has also been heating up. And I wouldn't put too much faith in Gorzelanny. He was good last season, but this season he is just a gas can.

compaqDikk · 06-02-08, 11:02 AM

i started playing LT's plays 4 days ago i kid you not after seeing a thread by JJ on how fukkin on fire he was . fortunately richkiss bailed me out with the red wrings

LT you gotta love the abrupt ending to that toronto game last night. hit batsmen then a bloop

BuddyBear · 06-02-08, 01:43 PM

LT, I've been pretty intrigued by this bullpen ERA system and have done some exploratory data analysis of my own to see whether the BPS stands up to empirical scrutiny or whether it is simply just random luck that has run its course.

Before I discuss the BPS though, I think the more pressing concern is to address the rate in which home teams are winning in MLB this year. It is simply unprecedented the success that MLB home teams are having so far. As of June 1st, MLB road teams are winning at a paultry 41.92% (358-496). Considering that year in and year out, MLB road teams have won at a rate of approximately of 44%-45% you have to believe that things will change. Or we could be dealing with the 2005 NFL season all over again. It's difficult to explain why this is happening but the reality is that it is happening and bettors should be very aware. Either way, I think what is important to consider in light of all this is that home teams are going to be overpriced from here on out (at least for an extended period of time) if they haven't been already. I can already see it in some of today's lines.

Now, it is worth noting though that road teams aren't always underdogs. In fact, in 236 games this year they've been favored but in 616 games they've been the underdog. But generally speaking, road teams are underdogs and if road teams aren't winning then that is a big problem for people who like to back dogs.

Just doing some basic analysis, we can look back at the past 9 years to see the rate at which home teams can be expected to be profitable. That is, if you played the home team in every game, what sort of rate do they win at? I'll present the number of homes teams that were profitable in each year. Here is the data:

1999: 9/30 home teams were profitable
2000: 15/30 home teams were profitable
2001: 13/30 home teams were profitable
2002: 14/30 home teams were profitable
2003: 13/30 home teams were profitable
2004: 12/30 home teams were profitable
2005: 12/30 home teams were profitable
2006: 16/30 home teams were profitable
2007: 11/30 home teams were profitable

2008: 20/30 home teams are currently profitable bets

On average, we can expect 14.111 home teams to be profitable in any given year based on this data. Clearly, the current 20 teams showing a profit at home is an anomaly and should correct for itself as the season goes on. Like I said, the main thing to note is that almost all home teams will be overvalued anywhere between 5cents to the standard 20-30 cents we see daily with the Boston's, NYY, NYM, and CHC of the world.

I am sure you are aware of this, but it is a very difficult betting environment for me and others like yourself if road teams are not coming through. I can't bring myself to lay -140, -150, etc...on teams.

On to the bullpen system......

Deke101 · 06-02-08, 02:11 PM

I'm against all four plays in one form or another. Sorry your gonna get rocked tonight LT. Haha just kidding...we'll see.

BuddyBear · 06-02-08, 02:20 PM

Now, to the bullpen system (BPS)....

Okay, whether the bullpen ERA is a strong predictor of betting success is an empirical question and should be subject to some empirical scrutiny. Fortunately, in baseball there is tons of data out there that can be easily accessed and retrived for analysis. Using a variety of baseball statistic websites, including Covers and Statfox, I compiled team data for the 2007 season. Data were collected on just about every meaningful statistic I could think of for a total of some 30+ variable: ERA, bullpen ERA, Batting average, OBP, OPS, Walks, Defensive efficeincy, team money profit, etc.......

1) So our first question then is there a relationship between bullpen ERA and the money a team earned for its backers?

That is we have 2 variables (Bullpen ERA and Money Earned for Backers) under investigation here. To answer this question, you could simply do a standard pearson-product correlation coefficient. Doing so gives you a statistically significant result of r = -.442, p= .014 with an R2= 19.54%.

In other words, the variance in money earned can be accounted for by a team's bullpen ERA at almost a 20% rate which is pretty decent. The main thing though is there a significant association between the two variables which means we can proceeded with more rigerous analysis.

Also, one thing I did, was I ranked all the bullpen ERAs (1-30) as well as ranked the profitability of each team (1-30) and did another correlation but this was a spearman's correlation which is what you want to use when you have two variables that are ordinal in nature (as opposed to the above pearson product correlatin which deals only with two continuous variables). Anyway, here we get a statistically significant result as well, rho= .503, p < .01. Another thing worth noting is that when do do this same exact procedure with just team ERA ranked (1-30) and the money ranks of each team (1-30), we no long have a significant relationshop rho= .284, p= .128. That means, when you throw in starting pitchers ERA we don't have a significant association anymore with ERA and money ranks.

Correlation is a nice statistic, but its limited b/c we can't really eliminate possible third variables that migh explain the relationship. True, we could do a partial correlation or ever a semi-partial correlation but we can also use OLS regression for our analysis.

2) So our 2nd question should be, is bullpen ERA still a significant predictor of money earned even after we control for other possible variables that might explain the relationship?

In this instance, we treat money earned for backers as our single dependent variable and our primary variable of interest is bullpen ERA. Now, let's throw in some other variables into our regression model that might be good (or even better) predictors for money earned. Let's include team batting average, team OPS (onbase pct + slugging), at bats, average walks per game, strikeouts per game, stolen bases per game, and GIDP (gronded into DP per 9 innings).

Even when we throw in all these variables into the model Bullpen ERA still is a significant predictor (p= .02) and is it is the only significant predictor of the 8 variables in the model. So none of the main offensive statistics are significant predictors but BP ERA is. That is a real good sign.

Now, you might be asking well why didn't you include variables like team ERA, and walks allowed, hits allowed, defensive efficiency ratings, etc...

Unfortunately, bullpen ERA is a defensive statistic and is highly correlated with other defensive statistic varaibles (i.e. team ERA, defensive efficiency, etc....) which means the model would suffer from high multicollinearity. So I had to remove those variables from the model, so it is possible the model could violate our main assumption of specification error (i.e. not including all relevant variables).

Also worth noting is that I only looked at 2007 data but according to LT the BPS has been profitable the past 3 years so it stands to reason the same type of relationships should be found from previos years and perhaps even stronger.

In any event, I am relatively optimistic that the BPS should turn it around and show a profit (soon). I suspect what's preventing the BPS from being successful is the lack of road team success as of now. If road team winning percent goes back to the traditional 44-45% I am sure the BPS should fall in line as well.

Good luck......

LT Profits · 06-02-08, 02:58 PM

BB,

Great work! The only thing I'd like to add is that the Orioles have qualified as a home BPS play quite a bit lately with negative results, and the Marlins basically carried the system as home BPS plays in early to mid-May. So while the lack of road success is certainly a big factor, it may not be the only reason the system has failed thus far.

Doc JS · 06-02-08, 04:47 PM

Originally posted by BuddyBear

In any event, I am relatively optimistic that the BPS should turn it around and show a profit (soon). I suspect what's preventing the BPS from being successful is the lack of road team success as of now. If road team winning percent goes back to the traditional 44-45% I am sure the BPS should fall in line as well.

BB,
Agree with LT, great stuff!!!

So, are you saying that for right now, at least until the road team winning % reverts to more historical norms, that the way to play the BPS is to play only home BPS plays?

Or should we just fade the Orioles as BPS plays because goodness knows I've been getting killed with those guys? LOL

Doc

ChuteBoxe · 06-02-08, 04:49 PM

Don't worry LT, my 60+ unit (past 10 seasons) system is losing like crazy as well. Let's turn this around this week, it's a new month. BTW, on the Bucs myself. Best of luck tonight.

BuddyBear · 06-02-08, 05:57 PM

Originally posted by Doc JS

BB,
Agree with LT, great stuff!!!

So, are you saying that for right now, at least until the road team winning % reverts to more historical norms, that the way to play the BPS is to play only home BPS plays?

Or should we just fade the Orioles as BPS plays because goodness knows I've been getting killed with those guys? LOL

Doc

Sorry fellows, i had to go renew my driver's license. What a fricken rip off....$34.00????

Anyway, should you fade BAL? According to LT, the answer would be a resounding NO!!! So close your eyes and play it like I did. Remember, if BAL is that bad they will eventually fall out of the top 10 in BP ERA and you won't have to worry about betting them any longer. Because it is a "system" you have to play a team regardless of how unappealing they appear. That's just the way systems are. No system should take longer than 5 minutes to identify which plays are eligible for it. Tonight, we have ARIZ and BAL.

Yes, I believe the lack of road team success is largely attributable to why the BPS has yet to take off. Because in the BPS, the majority of teams you will be backing are going to be road teams who are getting + odds, you'll need road teams to hit a higher rates. < 42% for MLB road team wins is not going to get it done. We'll need to see the road teams hit 44%-45% for this thing to be successful. We can't rely on -120 favorites to get us through...we need the +130s, _140s, +150s, and > to come through for us on a more consisten basis. It is within the realm of statistical probability that road teams continue to flounder at record lows this year or we could see a regression to the mean for MLB road team winning pct. We'll have to wait and see.

In addition, now that we are a 1/3 of the way through the season, there is less variance in the bullpen ERAs. We don't have too many teams moving in and out of the top 10 anymore. Take a look at the top 10 teams now and maybe the next 4 or 5 on that list and those are pretty much the teams you willl have to back on a regular basis in this system. As of today, of the top 10 teams on the BP ERA list, a total of 8 of them are currently profitable teams (exceptions: LAD & Atlanta).

I also said this before, but the bullpen ERA makes theoretical sense. One of the challenges dealing with analyzing sports data is that there is little theory to guide us. Sabermetrics is about it and it is not very well developed IMO. But intuitively bullpens play a major role in the outcome of games. Because starters are going less innings than ever before in MLB history (< 6 innings), bullpens are becoming increasingly important in the outcome of games. In my analysis, I looked at all the major offensive statistics and none of them were statistically significant. It makes strong theoretical sense to focus on bullpens then since 1) they will either have to win/lose the game in the event of a close game or 2) they will have to secure a lead in the late stages of the game. If it turns out the starter did a crappy job then there is little the bullpen can do other than keep it close but if the starter is servicable then having the better bullpen puts us in better position.

Think about it this way, would you rather have ARIZ or MIL's bullpen tonight if the game is tied say 3-3 in the 8th inning? Or put another way, would you rather being holding a ticket that says +133 on it or -143 if the game is tied 3-3 in the 8th inning and both starters are out of the game?

That's a pretty easy answer IMO

Good luck fellows.....

gizmo2431 · 06-03-08, 12:11 AM

Originally posted by BuddyBear

Think about it this way, would you rather have ARIZ or MIL's bullpen tonight if the game is tied say 3-3 in the 8th inning? Or put another way, would you rather being holding a ticket that says +133 on it or -143 if the game is tied 3-3 in the 8th inning and both starters are out of the game?

That's a pretty easy answer IMO

Good luck fellows.....

crap! i thought the answer was AZ, but apparently it was MIL in tonite's game.

BuddyBear · 06-03-08, 01:03 AM

Originally posted by gizmo2431

crap! i thought the answer was AZ, but apparently it was MIL in tonite's game.

It wasn't the bullpens fault I can tell you that much. Their defense was atrocious in that game and lack of clutch hitting as well. The 2 runs in the 7th were unearned (and charged to Davis, not the bullpen). In any event, that was a very difficult game to watch....ARIZ was clearly the correct call and their defensive miscues gave the game away.
Kind of funny though that I guessed the correct score in the late stages of the game....too bad you don't get paid for getting that right!

Anyway.....tomorrow is a make or break day for the BPS IMO. As of now, I see 9 games that are eligible. We'll see what happens with the line movement could make it 8 or 7 but it will be a very busy day!

BuddyBear · 06-03-08, 01:06 AM

Another miserable day for the road teams: 4-8 tonight. When will the madness end! And of the 4 road teams that cashed, 1 was a large fave and 2 were mid size favorites. Very difficult in this type of betting environment.....