[Bill the cop]

I posted this information over at SFW about a month ago and thought I'd share it with you guys. My sportsbettng "fiscal year" ends after the SB and doesn't resume until preseason (if we even have one this year). I had a very good year with my teasers and other parlays (and only got booted out of two books, BetPop and Heritage). Hopefully we'll have an NFL this year!


Push ATS data can be very useful for a variety of betting strategies, but it has to be accurate to be of any value.
We've taken great pains to make sure our databases and tables reflect correct historical results. I'm going to
list two push ATS tables. The first reflects the numbers from all regular season games from 1994-2010 (SS 4232).

The second table limits the push ATS data to just games with totals of 41.5 or less (SS 2445). This differentiation is important
for push statistics as well as for teasers:

TABLE ONE

1...69-60-3...2.3%
2...75-69-5...3.4%
3...333-364-75..9.7%
4...108-95-3...1.5%
5...56-62-2...1.7%
6...98-109-8...3.7%
7...134-151-20...6.6%
8...53-54-3...2.7%
9...48-49-1...1%
10..62-55-7...5.6%
11..25-26-1...1.9%
12..11-15-2...7.1%
13..20-23-1...2.3%
14..20-24-1...2.2%
15..7-9-1..5.9%

TABLE TWO

1...46-39-3...3.4%
2...43-48-4...4.2%
3...197-213-51..11.1%
4...63-52-2...1.7%
5...31-35-1...1.5%
6...55-70-6...4.6%
7...79-80-13...7.6%
8...28-34-1...1.6%
9...24-30-1...1.8%
10..38-23-2...3.2%
11..14-10-0..0%
12..5-4-1...10%
13..9-11-1..4.8%
14..12-10-1..4.3%
15..1-5-0...0%

As can be seen by these data the SS past 10 are really too small to be significant, but that's all there is.

TEASERS

Basic strategy teasers are also influenced by two other factors, totals and whether or not the game is
between divisional opponents. I have the complete data of how the subsets did, but I'll just post the
four results from best to worse:

1 (best) BS teasers/totals 41.5 or less/between Divisional opponents...172-58-0 for 74.8%

2 (next best) BS teasers/totals 41.5 or less....381-139-1 for 73.3%

3 (next best) BS teasers/ all games and totals...610-236-3 for 72.1%

4 (worse) BS teasers/ games between non-divisional opponents/totals greater than 41.5..355-140-2 for 71.7%

For those who want to know how one of the non BS subsets I bet (RD4.5 to 6) have done under the same criteria:

1..112-28-1 for 80%

2..215-69-2 for 75.7%

3..342-120-2 for 74%

4..183-75.1 for 70.9%

[tomcowley]

12..11-15-2...7.7%

LOLOLOLOL. Not only is your methodology and understanding godawful if you think posting that the NFL 12 is worth 7.7% is useful for anything except comic relief, BUT YOU CAN'T EVEN CALCULATE *THAT* RIGHT. 28 games.. 2 pushes.. 2/28= 7.14%. You ignored the pushes as actual games and used 2/26=7.7%. A++. And you did this for all the numbers apparently. You have great insights and I would like to subscribe to your newsletter.

[donjuan]

How'd your awesome road dog 4.5-6 subset go last year?

[Pancho sanza]

Yikes.

[Bill the cop]

LOLOLOLOL. Not only is your methodology and understanding godawful if you think posting that the NFL 12 is worth 7.7% is useful for anything except comic relief, BUT YOU CAN'T EVEN CALCULATE *THAT* RIGHT. 28 games.. 2 pushes.. 2/28= 7.14%. You ignored the pushes as actual games and used 2/26=7.7%. A++. And you did this for all the numbers apparently. You have great insights and I would like to subscribe to your newsletter.

I made the slight corrections in the push percentages. I'm used to dealing with teaser cover rates and ignoring the push numbers.

If you took the time to read the post you would have seen that I said the SS from 11 on were insignificant, but I posted them anyway in case some one wanted to know.

[Bill the cop]

How'd your awesome road dog 4.5-6 subset go last year?

Last year (2010) the subset didn't do well, 16-11 for 59%. The year before (2009) went 20-3 for 87%. The last five years, 97-34 for 74%. The last 17 years, 342-120-2 for 74%. The last five years, totals 141.5 or less, 52-13 for 80%.

This subset, like many other subsets, has quite a bit of variance, do in part to only averaging about 27 candidates a year. That said, the subset is still out performing just about any other teaser candidate out there. As always, this is a bet I've been making for over 5 years now, and don't really care if anyone else thinks it's worthwhile.

[Bill the cop]

Just for the record, I don't have a newsletter (nor will I ever), but if I did it would be boring. It would include, in priority order, family/friends, working out at the gym and how many reps with 315 I could get, the equity markets and other financial and business related issues, travel/fine dining, fishing, reading good books, and then, lastly, my part time hobby of sportsbetting.

[subs]

thanks for sharing again Bill, i'm sure many appreciate the numbers because they r using the HPC. for me i can double check my numbers and see where we differ.

if u think sending them to u would be useful, then just let me know. my data is from sports database and roasthawg.

I understand the pros in here being skeptical, especially about ATS bias. assuming it will not be 50% is a dangerous prop, however so can assuming it will not continue. if something has been the case for a number of years it can be assumed that there is a reasonable chance that it will, continue or that we may see a lessening of it maybe beyond break even. if the bias does not continue then its been a great run for u, kudos for being a winning gambler.

My question really is: is there a way to normalize conservatively for this, for safety's sake. i would like to produce
EV calculations for this assuming ATS cover rates "pulled" towards 50% and compare my EV for this. it would give me more confidence in my staking plan... i would average my EV and find my stake from there.

i'm starting to use weibul. is this ok or can any1 suggest something more accurate but does not require exponentially more work.

best of luck and long may the ATS bias continue.

[FourLengthsClear]

Just a quick question on the data.
You mention that table 2 is for games with a total of 41.5 or less. Is this a closing line total or the actual number of points scored?

[Bill the cop]

Just a quick question on the data.
You mention that table 2 is for games with a total of 41.5 or less. Is this a closing line total or the actual number of points scored?

The closing lines as depicted by Don Best. The median closing line since 1994 has been 41, but the last 5 years it has creeped up to almost 42, so 41.5 is what I'm going with.

[Bill the cop]

[quote=subs;9194569]thanks for sharing again Bill, i'm sure many appreciate the numbers because they r using the HPC. for me i can double check my numbers and see where we differ.

We use the closing lines depicted by Don Best. There are many times the closing line will be -3, -130 or a little less, we've tweaked those closers to be -3.5. So databases who ignore the larger vig and just show the closer as -3 regardless of the vig are going to differ from our database. We went through this with King Yao a few years ago and reconciled any differences by this method. It's usually the 3 that has to be delt with, and occasionally the 7, the rest of the numbers they'll usually just move the line the half point. But then you've got Pinny who may stay on the -9 and screw with the vig to protect themselves from teaser exposure.

In the present format of 32 teams and 8 divisions, about 37% of the games will be between division opponents. But we had to go back to 1994 and go year by year through 2010 and configure all the division games manually to get it right.

Hats off to my data guy John, for all the work he put in!

[donjuan]

LOL at them having quite a bit of variance yet refusing to accept that it's probably that same variance over a small sample size that has led you to data mine this subset and claim it's profitable. Especially when faced with completely logical arguments against the subset having +ev going forward and failing hard in your attempted rebuttals.

[Bill the cop]

LOL at them having quite a bit of variance yet refusing to accept that it's probably that same variance over a small sample size that has led you to data mine this subset and claim it's profitable. Especially when faced with completely logical arguments against the subset having +ev going forward and failing hard in your attempted rebuttals.

What really has me LOL is you thinking that I care even a little bit what you think about my bets. I heard the same "data mining" drivel from the likes of you 5 years ago. The "sharps" were unanimous in their opinion that any teasers other than BS were doomed to failure. They said, "get back to us in 5 years when the out of sample results would show how wrong I was". Well, here it is 5 years later and guess who's LOL now!!

[donjuan]

The alt lines at Pinny laugh in your face.

[Bill the cop]

The alt lines at Pinny laugh in your face.

If those Pinny alternate lines were better than -241, and you bet every one of them over the last 5 years, and they covered at 74%, which they did (97-34), I wonder if Pinny would still be laughing. In light of the facts, your argument is pitiful, but stay with it I can always use a good laugh.

[tomcowley]

If those Pinny alternate lines were better than -241, and you bet every one of them over the last 5 years, and they covered at 74%, which they did (97-34), I wonder if Pinny would still be laughing. In light of the facts, your argument is pitiful, but stay with it I can always use a good laugh.

And if you add in the "forgotten years" of 03-04 to give your full out-of-sample record since you datamined this crap and started betting it in 03, instead of dropping off 2 losing years up front and focusing on a streak like your average scamditout, your record is 97-34 + 56-37 = 153-71 68.3%, losing money. And that's even with a winning ATS record.

[Justin7]

If those Pinny alternate lines were better than -241, and you bet every one of them over the last 5 years, and they covered at 74%, which they did (97-34), I wonder if Pinny would still be laughing. In light of the facts, your argument is pitiful, but stay with it I can always use a good laugh.

Bill,

While there are arguments both ways, you should be paranoid if there is a big market willing to oppose you. If Pinnacle is offering these same alternates (and I have seen them as low as -220), the market must be respected.

[Bill the cop]

El

So because a book puts up a line that they perceive they have an edge with, you think it shouldn't be bet? The fact is every book offers bets that they think they have an edge with. Our objective should be to find situations that overcome their perceived edge. That's exactly what I've been doing with this subset for over 5 years now.

Are you telling me that if they offered -241 0n a RD+6 (alternate line +12) in a divisional game with a total of 40 you wouldn't bet it? I've shared information relating to the power of divisional games and low totals with you, so I know you are aware of this angle.

[Bill the cop]

And if you add in the "forgotten years" of 03-04 to give your full out-of-sample record since you datamined this crap and started betting it in 03, instead of dropping off 2 losing years up front and focusing on a streak like your average scamditout, your record is 97-34 + 56-37 = 153-71 68.3%, losing money. And that's even with a winning ATS record.

You're so full of shit your eyes are turning brown. Show me one post or other bit of evidence that I posted anything about this subset in 2003.

I just went back through my records, in 2006 I got into an argument with your buddy PC and posted I would list my teaser bets involving this subset starting with preseason in 2006. And I did so and showed a profit of $1900 as of week 4 (at which time, at the request of several bettors I shut it down as they thought I was giving away too much). Best as I can calculate, 2006-2010 is 5 years.

[tomcowley]

You're so full of shit your eyes are turning brown. Show me one post or other bit of evidence that I posted anything about this subset in 2003.

LOL, learn to read. I said you bet it then. And.. shockingly.. LVA itself is down right now, but you can get these quotes from http://webcache.googleusercontent.co...www.google.com in the 10/27/08 post at 4:27 PM.

Why aren't you taking Greek's +10.5 -240?

Historical data for a RD+4 teased to +10.5 is 84-40 or 67.7% (which is why I don't tease +4 RDs). However, the RDs +4.5 to +6 have went 299-104-2 for 74.2% over the same time frame (1994-2007).

So maybe they know the record of the +4 RDs too...

First of all, you've been killed betting those +4.5-6 Away Dogs since you first noticed your "trend" back in 2003 (they've hit at a 67.5% rate since then).

Secondly, you've now going with a belief that a smaller dog (+4) will cover a fixed line (+10.5) at a lower rate than a bigger dog (+4.5).

Any alarm bells going off yet?

I bet based on historical data, I view it as a marathon not a sprint, so what's happened over the last few years isn't that important to me. Also, I don't try to cherry pick certain years in an attempt to discredit other peoples betting paradigms (that could become a full time job). When (and IF) I'm no longer making money betting certain angles, I'll just have to evolve, until then I'll just keep on keeping on (and hope I don't get booted from any more books for making these -EV bets)!

So.. yeah. /thread

[Bill the cop]

Well, I just happen to have a copy of my post from the LV Advisor, dated 7/5/06, here it is:

"Over the last year my friend John (who seldom posts here, but when he does he uses the handle "Xsquare) and I have conducted an exhausive analysis relating to NFL teaser candidates. WE USED DATA FROM GOLDSHEET, COVERING YEARS 1994-2005 (in-sample). The primary focus was to see if we could develop a teaser model that went beyond the excellent work conducted by Stanford Wong and his "basic strategy for teasers approach" in his fine book "Sharp Sports Betting". And we did."

Now even a dimwit can see that if we conducted our study for years 1994-2005 and posted that I was going to start betting them in 2006 (out-of-sample), that I couldn't have been betting them in 2003 or 2004. Of course, even if you've been proven wrong (as you have here), I doubt if you're man enough to apologize.

[donjuan]

El

So because a book puts up a line that they perceive they have an edge with, you think it shouldn't be bet? The fact is every book offers bets that they think they have an edge with. Our objective should be to find situations that overcome their perceived edge. That's exactly what I've been doing with this subset for over 5 years now.

Are you telling me that if they offered -241 0n a RD+6 (alternate line +12) in a divisional game with a total of 40 you wouldn't bet it? I've shared information relating to the power of divisional games and low totals with you, so I know you are aware of this angle.

No, it's the fact that no one else in the world is betting it, ever, that should have you concerned. Why aren't you just betting these crazy soft lines with less variance than a teaser, no need to have more than 1 leg in a week and often times at a better price than a teaser and rolling in the money? Anyway, anyone who is even remotely literate can read TomCowley's post and see you have no idea what you're doing.

[Bill the cop]

No, it's the fact that no one else in the world is betting it, ever, that should have you concerned. Why aren't you just betting these crazy soft lines with less variance than a teaser, no need to have more than 1 leg in a week and often times at a better price than a teaser and rolling in the money? Anyway, anyone who is even remotely literate can read TomCowley's post and see you have no idea what you're doing.

I really am trying to make sense of your convoluted thinking. According to you over the last 5 years I shouldn't have been betting a teaser subset that has covered 74% of the time and made me lots of money? Is that it? Well, I'll have to take that "sage advice" under advisement.

As far as not knowing what I'm doing, I wish you'd get ahold of the books that threw me out and give them a heads up, they seem to think I know something!

[tomcowley]

Well, I just happen to have a copy of my post from the LV Advisor, dated 7/5/06, here it is:

"Over the last year my friend John (who seldom posts here, but when he does he uses the handle "Xsquare) and I have conducted an exhausive analysis relating to NFL teaser candidates. WE USED DATA FROM GOLDSHEET, COVERING YEARS 1994-2005 (in-sample). The primary focus was to see if we could develop a teaser model that went beyond the excellent work conducted by Stanford Wong and his "basic strategy for teasers approach" in his fine book "Sharp Sports Betting". And we did."

Now even a dimwit can see that if we conducted our study for years 1994-2005 and posted that I was going to start betting them in 2006 (out-of-sample), that I couldn't have been betting them in 2003 or 2004. Of course, even if you've been proven wrong (as you have here), I doubt if you're man enough to apologize.

No. The logical conclusion is that you've been disingenuous the entire time you've referenced 1994-2005 as the data-mining sample. You really want me to believe that somebody with your massive ego problems made *that* response to PC calling you a losing bettor on those for those two years? I don't think so. I would wager that you lost money betting it personally before doing some great collaborative study that swept your losing years into the backtesting sample.

[tomcowley]

Hahahahahah. Hahahhahahahahahhaha. Hahhahahahahhahahhahahahha.

Not only was I right, but you quoted from the thread with the evidence.

http://www.lasvegasadvisor.com/forum...E=&STARTPAGE=1

from 7/6/06 9:16 AM post

That's one of the reasons I've been teasing -3 games down to +3.5 with 6.5 point teasers the last couple of years. So when we conducted our study, two questions I wanted answered were: Is teasing a -3 point down to +3.5 a statistically valid approach, as I've been quite successful doing this in the pass? And, is teasing a +4.5 to +5.5 dog up to over 10 also valid from a statistical perspective as I've also been doing that for a few years? Both questions were answered in the affirmative (as the data expressed earlier demonstrated).

Please just go away now. You don't have to stick around to admit to your lies.

[Bill the cop]

You really are a piece of work! You actually think I was betting that subset in 2003 and 2004 and losing lots of money? Then decided it must be a good bet if I (with my giant ego) was betting it and conducted a study ignoring any losing years and just came up with years to show it was +EV? FWIW, in 2003 and 2004 I was a member of the boxseats at SSB and was betting StevieY and Overlays picks exclusively. I do remember StevieY giving out a teaser pick of a RD+5.5 once or twice. In fact when John and I started looking at the database I mentioned it as one thing to look at as I respected StevieY's opinion. Going into a database with a specific theory of what might be a good subset pretty much kills the idea that we were just datamining.

Bottom line, from 2002-2004 I was paying quite a bit for the hosts picks at SSB, and those were the only bets I was making. I may have been losing money betting the hosts picks, but I sure as hell wasn't going off the reservation and making my own bets.

[Bill the cop]

Hahahahahah. Hahahhahahahahahhaha. Hahhahahahahhahahhahahahha.

Not only was I right, but you quoted from the thread with the evidence.

http://www.lasvegasadvisor.com/forum...E=&STARTPAGE=1

from 7/6/06 9:16 AM post



Please just go away now. You don't have to stick around to admit to your lies.

As I explained in the other post, I was betting SSB hosts picks 2002-2004 and a few times StevieY posted those plays, and I did bet them. And as I said, that was the genisis for looking at that subset for our study.

[tomcowley]

Bottom line, from 2002-2004 I was paying quite a bit for the hosts picks at SSB, and those were the only bets I was making. I may have been losing money betting the hosts picks, but I sure as hell wasn't going off the reservation and making my own bets.

is teasing a +4.5 to +5.5 dog up to over 10 also valid from a statistical perspective as I've also been doing that for a few years?

You really are a piece of work! You actually think I was betting that subset in 2003 and 2004 and losing lots of money?

Uh.. yes? LOL.

[tomcowley]

As I explained in the other post, I was betting SSB hosts picks 2002-2004 and a few times StevieY posted those plays, and I did bet them. And as I said, that was the genisis for looking at that subset for our study.

So even in this version of the "truth" you put your actual losing bets into the backtesting sample instead of the out of sample. Got it.

[JustinBieber]

Conclusion : Cops are dumb

[Justin7]

-edit

[Bill the cop]

Conclusion : Cops are dumb

I'm sure a sentiment shared by your brethren in the fast food industry.

[Bill the cop]

So even in this version of the "truth" you put your actual losing bets into the backtesting sample instead of the out of sample. Got it.

You can parse whether or not a few bets had any impact on the study, but one thing is clear:

17 years, 342-120-2 for 74%

Last 5 years, 97-34-0 for 74%

Do you see a trend here?

You're free to continue in your quest of people in other forums who may have posted a play in 2004 at -3 when it should have been -3.5 (you really are doing Gods work).

My last word on this issue I've got much more important things to do.

[bztips]

My last word on this issue I've got much more important things to do.

That's a shame; I find BTC's threads endlessly amusing

[tomcowley]

Do you see a trend here?

Terminal cluelessness.

You're free to continue in your quest of people in other forums who may have posted a play in 2004 at -3 when it should have been -3.5 (you really are doing Gods work).

Trend continued. As usual no clue what you're talking about.

My last word on this issue I've got much more important things to do.

For the love of Favre, I hope you're finally telling the truth, but after lying about it so many other times, it's hard to believe you this time. You know, upon further review, I hope you're only telling half the truth. Somebody with your problems ever having anything actually important to do is a scary thought.