Outside The Box

[Thremp]

Thremp, I was happy to see you here. I was hoping you were gonna answer my teaser math question??!! lol

I didn't even spot it. This is "high season" for me atm and I'm pretty busy. Nor do I have access to longterm historical databases atm (Its amazing how long freelance bum programmers take to do anything ).

Though I like this thread and will be following it when I have time.

[HedgeHog]

I don't have the data, but I would imagine 7 extra points would improve you to about 70% in football, and somewhat better in hoops as Durito pointed out.

[katstale]

Thremp, nothing wrong with being a "bum"--when people ask me what I do for a living--I been saying that for years.

I appreciate the math Dorito and if I had guessed--I would have said that was abt right. Just glad to see some concrete thrown in with speculation/theory.

For sure there is a difference between the two sports because a point is much more valuable in football than basketball. BUT, it is not as much as people might imagine.

More tomorrow.

[Data]

a point is much more valuable in football than basketball

Actually, it is the other way around.

[durito]

Data is correct, pts are worth more in basketball.

[katstale]

Ouch!!! My bad!! Because of the number of points scored in basketball relative to the points scored in football the tendency is to to think the point is more valuable in football.

Like if i was to give you a point in hockey, football or basketball--whch sport would you take the point in?

Well, hockey of course.

Anyway, thanks for catching that.

[katstale]

Now we get into some interesting mathematical questions. If we are all in general agreement that teasing the line should make us approx 70% winners ATS--what if we were allowed to tease it 5 points. What would be the win % of that much movement in our favor?

Soon as I get help with this number I will follow up tomorrow.

[durito]

In College Basketball you'd get approx 68% winners at 5pts. (75% at 7pts). While it varies based on the spread, generally 1/2pt is worth approx 1.9%

NBA numbers are similar. I can do football too if you like.

[HedgeHog]

5 points in football would be around 63-65%, depending on the key #s you're teasing through. Tomorrow we'll be talking about a smaller advantage I'm sure. I think Kat is working his way down to that 54-56% percentage he mentioned earlier. In that case, using Durito's data, you need only a 1-2 point advantage vs the actual line. Is that where we're going?*

[katstale]

Now its coming together for you. Durito, I would love to see those football #'s because it is instructive for each of you building your own particular model. We are gonna be all over this "tease the line" philosophy and how to achieve it. Well, I should say I will keep pointing you in this direction.

[metaldome]

I have been gambling for a while and actually just got back from vegas, had a great time, won money on my Giants, was at UFC 81 and won some money on that, and my birthday was also last week so it was fun, I have been on some other forums for a while but just found this one, I think the idea of a Handicapper Think Tank is great, lots of other sites have peoples picks but I love to hear and share theories that actually help everyone cap better so thanks for the thread Katstale, interested in seeing where it goes...

[Munson15]

How about taking the opener and going contrary to any line moves of 2< points?

[katstale]

So its football season (college and pro) and you get up Monday morning and you get the "right" paper and you get a cup of coffee and a pair of scissors. Settle in and enjoy the sports section. Before you throw it away you go to that section that says "morning line" and you will see all kinds of games that will take place that saturday and sunday. You cut this section out. Now, you have your baseline for the weekend.

As the week moves forward the market is gonna decide which games you are gonna get to tease. All, you have to determine is how many points you want? On Friday night or early saturday morning, you log into the reduced juice book of choice (Pinny if available) and you make all the selections that fit your model.

Similar process for basketball. Except you have to work a little harder because its 15 minutes every day instead of just 15 minutes on the weekend.

As was mentioned in the previous post, I would tend to think 2 is the minimum tease and gets you to a degree of confidence and statistical significance.

You still got plenty of NBA and college left to run a short simulation. I am not the stat guy, but as you approach 1,000 games you should be at the .05 confidence level.

I would say "good luck", but I hope we are now thinking of ways of moving away from the luck model.

[TLD]

Why are you using the opening lines as the “Holy Grail” rather than the present lines? Wouldn’t one expect that there will have been a line movement in precisely the games where the opening line was least accurate? If the idea is to trust the market more than oneself, and thereby to let the “sharpest” folks do your handicapping for you, then wouldn’t the present lines be a better indicator of the market’s opinion on a game?

In any case, rather than speculate, surely the numbers either support or oppose this strategy. If a person were to bet, say, the morning of the game on any game where he could get at least two points off the opening line, over the last however many years how would he have fared at the various sports?

[Munson15]

So its football season (college and pro) and you get up Monday morning and you get the "right" paper and you get a cup of coffee and a pair of scissors. Settle in and enjoy the sports section. Before you throw it away you go to that section that says "morning line" and you will see all kinds of games that will take place that saturday and sunday. You cut this section out. Now, you have your baseline for the weekend.

As the week moves forward the market is gonna decide which games you are gonna get to tease. All, you have to determine is how many points you want? On Friday night or early saturday morning, you log into the reduced juice book of choice (Pinny if available) and you make all the selections that fit your model.

Similar process for basketball. Except you have to work a little harder because its 15 minutes every day instead of just 15 minutes on the weekend.

As was mentioned in the previous post, I would tend to think 2 is the minimum tease and gets you to a degree of confidence and statistical significance.

You still got plenty of NBA and college left to run a short simulation. I am not the stat guy, but as you approach 1,000 games you should be at the .05 confidence level.

I would say "good luck", but I hope we are now thinking of ways of moving away from the luck model.

During football, I heard a similar method referred to as "Crazy Eights" and it applied only to football. The idea was to look for line moves of 2 or more from the opener (Monday morning, not some outlaw line) and fade any favorites or overs. Haven't read any reference to baskets, and the method was only applicable to overs and faves. It did pretty well.

[DukeJohn]

So its football season (college and pro) and you get up Monday morning and you get the "right" paper and you get a cup of coffee and a pair of scissors. Settle in and enjoy the sports section. Before you throw it away you go to that section that says "morning line" and you will see all kinds of games that will take place that saturday and sunday. You cut this section out. Now, you have your baseline for the weekend.

As the week moves forward the market is gonna decide which games you are gonna get to tease. All, you have to determine is how many points you want? On Friday night or early saturday morning, you log into the reduced juice book of choice (Pinny if available) and you make all the selections that fit your model.

Similar process for basketball. Except you have to work a little harder because its 15 minutes every day instead of just 15 minutes on the weekend.

As was mentioned in the previous post, I would tend to think 2 is the minimum tease and gets you to a degree of confidence and statistical significance.

You still got plenty of NBA and college left to run a short simulation. I am not the stat guy, but as you approach 1,000 games you should be at the .05 confidence level.

I would say "good luck", but I hope we are now thinking of ways of moving away from the luck model.

Well, I took the opening lines at Pinnacle vs Closing Lines for this past football season. Picking only the lines that moved 3 or more pnts for the o/u or ats and going with the better side. For instance in the total opened at 51 and moved to 48 I would take the over. The ats I took whichever way the lined moved for my favor. If a +(Dog) went up I picked that one, if the -(Fav) went down I picked that one. Anyway an interesting thing did happen. These were my results:

ATS: 5-9
O/U: 7-14

So if you faded these or just went with the line movements at 3 or more pnts you would have done pretty good this season.

ATS: 9-5 64%
O/U: 14-7 67%

Not nearly a bid enough sample, but perhaps someone might go back years and test this out...

[Munson15]

Well, I took the opening lines at Pinnacle vs Closing Lines for this past football season. Picking only the lines that moved 3 or more pnts for the o/u or ats and going with the better side. For instance in the total opened at 51 and moved to 48 I would take the over. The ats I took whichever way the lined moved for my favor. If a +(Dog) went up I picked that one, if the -(Fav) went down I picked that one. Anyway an interesting thing did happen. These were my results:

ATS: 5-9
O/U: 7-14

So if you faded these or just went with the line movements at 3 or more pnts you would have done pretty good this season.

ATS: 9-5 64%
O/U: 14-7 67%

Not nearly a bid enough sample, but perhaps someone might go back years and test this out...

I appreciate the info. If you wouldnt' mind, I'd love to see the results using 2< point moves in the same way.

[katstale]

I love where you guys are going with all of this. Some of you are gonna make some money. BUT, stick with the baseline from the morning paper example (hint:it is not USA Today!! lol) and don't use Pinny as your base. You CAN use Pinny as the base, but the example given will give you many more games to wager on and will also yield many more basketball games which will meet the formula on any given game day.

In fact, using that early "syndicated line" out of Vegas will yield a few 5 pt teasers in football. Seeing is believing.

[Munson15]

Are we talking about "America's Line", Kat?

[katstale]

Far as i can go on this (and everything can be found on the internet) is the line in the morning at the Chicago Tribune will work for you.

[Munson15]

Far as i can go on this (and everything can be found on the internet) is the line in the morning at the Chicago Tribune will work for you.

Thanks, I'm pretty sure they are running the same line I see in my local paper. So we're on the right track.

[HedgeHog]

I think you lost me .* Accepting the opening line as better than a later line is clearly a mistake.* It ignores injuries and all later info.* I get the Trib too, so you're using Monday's newspaper as your Bible for the upcomimg weekend's games (as far as FB is concerned).

[katstale]

For the most part, during the week and even any given day--the line will overcorrect on news. We are looking to exploit over corrections. I don't want to completely rule out, as I said earlier, "local knowledge", but I sure want to minimize it.

I know the stats will bear out the opening line (baseline) we are using is well set. Deviations away from this are opportunities.

[Ganchrow]

For the most part, during the week and even any given day--the line will overcorrect on news.

I know the stats will bear out the opening line (baseline) we are using is well set. Deviations away from this are opportunities.

Is this speculation or do you have hard evidence to back up these claims?

[katstale]

Well, I could tell you I have over 3000 games to support this "theory", but I would rather people check it out for themselves and make their own decisions. Collect the data each week--doesn't take alot of time--then decide.

I think Fezzik printed an article last year where he told people "this has been around for some time" and he wasn't sure who came up with it. His method was similar to this but not exactly the same. I believe/know it was discovered in 1991. It couldn't fully be exploited until a few years ago.

[Munson15]

Well, I could tell you I have over 3000 games to support this "theory", but I would rather people check it out for themselves and make their own decisions. Collect the data each week--doesn't take alot of time--then decide.

I think Fezzik printed an article last year where he told people "this has been around for some time" and he wasn't sure who came up with it. His method was similar to this but not exactly the same. I believe/know it was discovered in 1991. It couldn't fully be exploited until a few years ago.

Kat, I am intrigued by this line of logic, as it dovetails nicely with some theories of my own. I'm still interested in nailing down precisely where you're at though, because I am strictly an amateur.
My read on this is that any line movement of 2 or more would trigger a situation, and that would be irrespective of whether or not the move favors the dog or the fave. I googled the Chicago Tribune to be sure but could not confirm if the line I believe you're referring to is "America's Line."
Any further nudge in the right direction would be greatly appreciated.

[donjuan]

It's hilarious that people think opening lines are more efficient than closing lines in major sports like the NFL and NBA.

[katstale]

Amazing anybody would think that? Fact is, if you read above--the closing line is often inefficient due to herd movement. In any case, the thread is aptly titled thinking outside the box. Don, if you can beat 54%--you are doing good. You don't need your thinking adjusted.

Munson, you are on it. Now you have to decide how much teasing of the line you want to do. Make adjustments based on your own work and analysis. Can't help you more than that. If you see someone else on here that is following along-you might be able to partner up with them. Bounce ideas off of them.

[20Four7]

Don,

The closing line isn't more efficient, and I do agree with kat that sometimes it gets "oversold" . I know that has led to some time where I try and middle a game. If I like the over at 136 for college hoops and the line has moved to 141 or so, I will then take the other side if I believe it has moved to far. Now I am looking for a 5 point middle. I still belive with books like pinny, the market is far more efficient than Joe the guy in the bar that books bets. You don't stay in business with a -104 model if you gamble and are wrong.

[Ganchrow]

You don't stay in business with a -104 model if you gamble and are wrong.

Very true ... you stay in business with a -104 model by adjusting your lines in response to sharp action.

I'd be very interested in seeing compelling evidence in support of the claim (made by some contributors to this thread) that markets on average become less efficient as game time approaches.

[HedgeHog]

How about lines that change due to late injury news?* Would these still be playable?

[Munson15]

How about lines that change due to late injury news?* Would these still be playable?

I believe that lines are overcooked due to injuries and I've read studies regarding football which state that the first game after the injury the team compensates for the missing player with extra effort. Don't know if this translates into baskets, would like to see what Kat has to say. Good ?.

[Munson15]

BTW, last night there was big movement against Xavier from the opener and Xavier won. This would fit the model we are discussing here, I think.

[katstale]

It's not that the totality of the line before game time becomes less efficient, it's that certain games do.

Let me give you a simple example--haven't you noticed how much a "TV" game moves in relation to other games? The degenerate wants to bet what he can watch!! This creates value for the businessman. An opportunity exists.

Hog, I see value in injuries. Unless the injury was extremely late, it wouldn't affect my thinking at all. I would hope people were running to one side.

[Munson15]

Munson, you are on it. Now you have to decide how much teasing of the line you want to do. Make adjustments based on your own work and analysis. Can't help you more than that. If you see someone else on here that is following along-you might be able to partner up with them. Bounce ideas off of them.

Thanks for the encouragement. Can anyone chime in on how 2-point or 3-point teasers would improve a straight up record? I'm hoping you'll say approximately 54% on a 2-point move/tease and maybe 56% on a 3-point move/tease.