Big Dog MLB 2011

[bombCanada]

So you are just betting within the certain price range? I'm very interested to see how this system performed in past years. Let me know if you need back testing done. I have a scraper and some programming skills to contribute.

You can look at sports betting as either a test of sports knowledge or as a math problem. I have chosen the second route. You can tell which way is more popular based on the number of reads/posts for this thread v. that of, say, paco's "My Sports" thread.

While I appreciate all the offers of help, and would encourage the members of this thread to continue to contribute and work together, it's important to me that I develop and run the official thread system as my own work. I've told only one person the complete details of what I'm doing, but a couple of you have done a pretty good job of figuring it out for yourselves. I'll try to let you know where I think you might be going off track, but I don't want to stifle innovation. I would rather foster discussion and development, and I'm excited about how this thread has done that.

Good work guys.

[bombCanada]

here is something to turn on the discussion

Good work sir, and thank you.

[bombCanada]

So you are just betting within the certain price range? I'm very interested to see how this system performed in past years. Let me know if you need back testing done. I have a scraper and some programming skills to contribute.

While it'll be interesting to see how various months compare year over year, I think the statistical validity comes from the whole season... ie the larger the sample size without artificial constraints, the more valid the data and the more successful the system.

Remember, one reason I'm at 200 units on the year after 6 weeks is because of the very large number of wagers. A person who picked all winners, but only 1 per day, would be at 45 units (give or take). That's why people who try to find filters to screen out losses probably hurt themselves, because they're probably missing wins, also. I'd rather have the wins, ESPECIALLY when when a win is always larger than a loss.

[bombCanada]

Thanks! Oh great undressed one.

[bombCanada]

@ Dollar Bin:

i have that stats too. i mean, broken into months of a certain year. but i have only data for 2004-2010, and that's being only seven seasons so i can't make any conclusions. almost every year there's one (or two) black month(s) with around or under 30% strike rate of the system - and that month is different for every season. if i'd have to choose one month and gun to head i would go with June... idk why, only 2 months under 38% and 4 months over 40%.

other thing i would like to stress out... imo bookies are in constant change and no two seasons set odds identically. there are just maybe similarities with two seasons which we can detect late in the season or when it finishes. so, bombCanada's way is the only possible here - have fun and go small. hence i haven't made complete simulation using Lab lines from day to day idk what percentage is cutoff, but be cautious with this - one month strike rate went under 26%! (July 2009)

yeah i know there were only 5 gamedays in May but the interesting thing is that strike rate increased and for these 5 days it's 45,8% - unbelievable!

Great contribution. And while it would be painful, I could play through a 26% month like July 2009. Again, knowing the power of a large sample size, you can play through that knowing that it WILL recover. And whereas as chasing would crush the chaser, lab lines will keep you in the game much longer.

[bombCanada]

Great discussion in here guys. That spread sheet is very interesting. Looks like even as this thing regresses you could still make money with it.

That's the plan! Welcome.

[bombCanada]

brko-

Thanks for the detailed response.

You see were I was going. If other years had high April's and regressed to the 38-40% range there should be an expectation that this year may do the same. BC has stated he expected that some regression of the hit rate may happen and looking at the data I would expect it to regress to that range (38-40%).

Of course this year could be a statistical outlier and the hit rate could stay above that range. I'm not planning on it and will stay conservative in my unit size. A labby would be helpful but the line could get very long with a month like July 2009.

Thanks again. Great stuff!

A long line is not bad, and in fact is a good tool for discipline, in that it provides goal orientation and guidance for wager sizing. I'd rather see a long line than a blowout from chasing, or a long slow death from flat betting.

[bombCanada]

5/7:

Indians +165 6.00u
Twins +148 6.00u

Best luck!

[jpeterb]

Great Season your having.

[figue]

Underdogs are [225-217, +67.7 Units] this season. NL dogs are 118-117, +31.8 units. AL dogs are 107-100, +35.9 units. Divisional rival dogs are only 82-101, +6 units meaning most of the units won are for betting on non-divisional rival dogs. We love betting dogs.
Road teams are 221-243, +12.4 units thus far in the season. Home teams generally do much better in the beginning of the season than they do at the end. Since teams are now just getting into the groove of playing ball, the home field advantage just becomes that much more of a better. But because we are on top of all the baseball trends and differences that occur season to season, we know how to adjust and bet.
Underdogs this season off of being shut out are 18-17, +5 units this season.
Underdogs this season coming off having scored double digits are 16-9, +9.7 units. In fact, over the past 12 season, this is a winner. +77 units. Only had two losing seasons during that time.

[nicolaitanghoj]

What do you mean about NL and AL and underdogs coming off?

[bombCanada]

5/7:

Indians +165 6.00u *win*
Twins +148 6.00u *lose*

77-74 +204u

[bombCanada]

5/8:
Nationals +154 6.00u
Twins +169 6.00u

Best luck!

[bombCanada]

5/8:
Nationals +154 6.00u *lose*
Twins +169 6.00u *lose*

Skunked!

77-76 +192u

[benrama]

Hi BC - been thinking of doing something similar to your approach, i.e. playing all dogs over +130 and using Labouchere lines. I tend to generally prefer just straight and keeping my units the same until I reach certain bankroll goals - but am still weighing that up with this system.

I have done some excel modeling on dogs with the following assumptions:

Average odds: +150
Target win %: 40%
Total number of plays: 600 (~150 days left with ~4 plays/day)

The actual win/loss pattern is determined by a random number generator. Obviously not so relevant for equal sized unit betting (though it does influence when you might increase or decrease bet sizing), but important for the Labouchere. Currently working on some macros to calculate what results the Labouchere would come up with, and running it over various runs of randomisation to get a sense of what the likely outcome is.

At +150 odds, with consistent flat-betting, the break-even % is pretty much smack on 40%. With Labouchere lines you need to be able to tolerate 9-10 loss streaks - the modeling says that these can and will happen.

Intuitively why I think there is money to be made on a dog system is that sportsbook markets are efficient with respect to generating 50/50 action, not generating 100% precise lines, and favourites will tend to have slightly over-inflated lines due to the public wanting to bet on them. All it takes then is a few consistently undervalued teams (e.g. Royals and Indians this year) and you have some nice $ to be made. I still think there are some extra filters that could be applied to avoid playing dogs is "bad spots" but it starts to complicate things and I'm with you - simple is best - if it works.

Having said all that I choose a 0-2 day to start employing this strategy, let's hope I'm not my usual wet blanket on these promising betting strategies :P

[brko]

3 additional dogs drifted over +130 at my book yesterday - so that making them 5 on Sunday 8th, but only one managed to "surprise" his opponent. GL in the future. if you bet all 5 you would be down -2.65u

[bombCanada]

5/9:

Mets +161 6.00u
White Sox +139 6.00u

Best luck!

[bombCanada]

Hi BC - been thinking of doing something similar to your approach, i.e. playing all dogs over +130 and using Labouchere lines. I tend to generally prefer just straight and keeping my units the same until I reach certain bankroll goals - but am still weighing that up with this system.

I have done some excel modeling on dogs with the following assumptions:

Average odds: +150
Target win %: 40%
Total number of plays: 600 (~150 days left with ~4 plays/day)

The actual win/loss pattern is determined by a random number generator. Obviously not so relevant for equal sized unit betting (though it does influence when you might increase or decrease bet sizing), but important for the Labouchere. Currently working on some macros to calculate what results the Labouchere would come up with, and running it over various runs of randomisation to get a sense of what the likely outcome is.

At +150 odds, with consistent flat-betting, the break-even % is pretty much smack on 40%. With Labouchere lines you need to be able to tolerate 9-10 loss streaks - the modeling says that these can and will happen.

Intuitively why I think there is money to be made on a dog system is that sportsbook markets are efficient with respect to generating 50/50 action, not generating 100% precise lines, and favourites will tend to have slightly over-inflated lines due to the public wanting to bet on them. All it takes then is a few consistently undervalued teams (e.g. Royals and Indians this year) and you have some nice $ to be made. I still think there are some extra filters that could be applied to avoid playing dogs is "bad spots" but it starts to complicate things and I'm with you - simple is best - if it works.

Having said all that I choose a 0-2 day to start employing this strategy, let's hope I'm not my usual wet blanket on these promising betting strategies :P

Good stuff. Yes, long losing streaks will happen, as will winning streaks. That's the beauty of really large sample sizes.

What would be interesting to know is do the win/loss streaks fall into a standard distribution? I suspect not because that assumes that the outcome of each game is equally probable, and I'm guessing it's not (that's the whole idea of odds, after all).

There were two days where all games lost in April (the 14th: 2 games, and the 21st: 2 games, and the 24th: 3 games) ok three days. I wouldn't get too freaked out about it. We've had a day where we won all 6 games. You know what's coming.

Good luck.

[bombCanada]

3 additional dogs drifted over +130 at my book yesterday - so that making them 5 on Sunday 8th, but only one managed to "surprise" his opponent. GL in the future. if you bet all 5 you would be down -2.65u

You have to draw the line somewhere with any system, and I draw it as the overnight lines before I go to work. I know that may be difficult for some people to follow. That's why it's important to me that the system is my own work. It has to be easy for me to follow. YMMV.

[rkelly110]

I also believe it's important to have a consistent time when to look at the odds.

I'm with BetUS and they usually don't have all the games up until 11 am, that's when
I do my print outs and base my odds off my print outs. I don't follow line movements.

I have noticed sometimes if the odds are lower than my printouts when I place my bets,
the games have a better chance of winning. IMO

GL today. (two games today, an A,B,C,D box parlay?)

[benrama]

Good stuff. Yes, long losing streaks will happen, as will winning streaks. That's the beauty of really large sample sizes.

What would be interesting to know is do the win/loss streaks fall into a standard distribution? I suspect not because that assumes that the outcome of each game is equally probable, and I'm guessing it's not (that's the whole idea of odds, after all).

There were two days where all games lost in April (the 14th: 2 games, and the 21st: 2 games, and the 24th: 3 games) ok three days. I wouldn't get too freaked out about it. We've had a day where we won all 6 games. You know what's coming.

Good luck.

Yes, my analysis so far assumes standard distribution, but no reason why it couldn't e.g. randomly generate odds between +130 and +190 and have the win % differ as the odds increase. I guess at the moment I'm just getting a feel for what needs to be achieved to make money here and keeping it simple to start with.

Regarding the Labby lines, I assume you are using a reverse Labby, and one long line. How do you adjust after a win?

Say your line is:
1 2 3 4 5 6

And you go 1-1 for the day, winning, say 0.3 UNITS. How does your line get modified?

[doesnotcompute]

With all the money this guy's making he could really live up to his namesake! (that is bomb canada)

[LtDementia]

With all the money this guy's making he could really live up to his namesake! (that is bomb canada)

He could even change it to "buy canada"--

LT

[bombCanada]

With all the money this guy's making he could really live up to his namesake! (that is bomb canada)

He could even change it to "buy canada"--

LT

About once a year I have to explain my name...

Years ago a friend and I were going to sign up for online poker, and we were talking about names we'd seen that we liked. I had always liked the clever or funny ones the best, as well as some of the more offensive ones. I was trying to think of a name that was just outrageous, and after a couple of beers and some laughs it all boiled down to what if we were to do something outrageous to our neighbor to the north with the longest unprotected border in the world la-la-la... and bombCanada was born simply because it HAD to offend a certain number of people on a regular basis.

And it did... about once a day in the little live discussion underneath the poker tables some guy would start after me about, "Hey, pal, I'm from Canada and you can go screw yourself..." or similar alcohol fueled outbursts. Well, it didn't take too long to figure out that I was much better at offending people than at playing poker, so I moved on, but kept the name.

And here I am. With no interest in bombing anything or anyone anywhere anytime. (But did you see the Mythbusters episode about Mentos and Diet Coke? )

[bombCanada]

Yes, my analysis so far assumes standard distribution, but no reason why it couldn't e.g. randomly generate odds between +130 and +190 and have the win % differ as the odds increase. I guess at the moment I'm just getting a feel for what needs to be achieved to make money here and keeping it simple to start with.

Regarding the Labby lines, I assume you are using a reverse Labby, and one long line. How do you adjust after a win?

Say your line is:
1 2 3 4 5 6

And you go 1-1 for the day, winning, say 0.3 UNITS. How does your line get modified?

My initial reaction to all this is it sounds very interesting, and very time consuming, but then again I'm not so bright, so maybe it's easier than I think it is.

The fundamental problem of statistical analysis is that of the standard distribution. For example, lots of futures traders like to run systems through Monte Carlo simulations... but the problem (as I understand it) is that standard distribution is implicit in the Monte Carlo method. And because futures markets returns over time tend to exhibit "fat tails", anything standard distribution related tends to fail at correctly modelling those markets.

So that's all a long way of saying, the analysis would be interesting, but it would be more interesting if it were valid (but then again, how would I know?). Someone somewhere must have at one time written a book or paper about the statistics of baseball.

As to how I do lab lines, it's one of those things I don't really explain. Search SBR for posts by JW Cash, who has done the best job I've ever seeing of explaining how it is done well. Every time I read what he writes it makes sense to me. Do what makes sense to you.

I've seen enough, and banked enough, over the first six weeks that I can ride out a pretty tough time ahead to prove that this system does or doesn't work, and to me that's the most valid trial there is.

[bombCanada]

5/9:

Mets +161 6.00u *lose*
white sox +139 6.00u *win*

78-77 +194

[bombCanada]

5/10:
Nationals +195 6.00u
Padres +157 6.00u
Mets +154 6.00u
Diamondbacks +183 6.00u
Blue Jays +149 6.00u

Best luck!

[VicVega91]

best of luck today bc, been following your plays for past couple of weeks. Thanks and keep up the good work !

[benrama]

Great day today BC, all 1 run wins by the dogs but you'll take that, Arizona had a chance there as well. Great win by the Nationals, love cashing those circa +200 ones!

[bombCanada]

5/10:
Nationals +195 6.00u *win*
Padres +157 6.00u *lose*
Mets +154 6.00u *win*
Diamondbacks +183 6.00u *lose*
Blue Jays +149 6.00u *win*

81-79 +212u

[bombCanada]

5/11:
Mets +164 6.00u
Nationals +195 6.00u
Diamondbacks +153 6.00u

Best luck!

[bombCanada]

best of luck today bc, been following your plays for past couple of weeks. Thanks and keep up the good work !

You're welcome, and thanks for the support.

[bombCanada]

Great day today BC, all 1 run wins by the dogs but you'll take that, Arizona had a chance there as well. Great win by the Nationals, love cashing those circa +200 ones!

Yes I'll take them all. Thanks!

[bombCanada]

5/11:
Mets +164 6.00u *ppd*
Nationals +195 6.00u *win*
Diamondbacks +153 6.00u *lose*

82-80 +218u

[bombCanada]

5/12:

Cubs +136 6.00u
Nationals +176 6.00u
Royals +176 6.00u
Mariners +144 6.00u
Mets +166 6.00u

Best luck!