Always try to beat the closing number and be ready to pull the trigger when need be!

**The General** · 04-23-10, 04:21 AM

Thanks for sharing and good job.

I think a guy named Nicky Satriano always pointed out that you need to beat the closing line

Yes sir, Nicky is very adamant about beating the closers.

**pats3peat** · 04-24-10, 04:30 PM

Not bad eh

**donjuan** · 04-25-10, 04:30 AM

Odds Justin moved/left this to/in the Think Tank

Yes +10000000000000000

**Peep** · 04-25-10, 10:41 AM

I would have an easier time beating the closing number if I knew what it was going to be.

I would even settle for knowing what way it is going to move. You must be very smart to know that.

Hell, I would even settle for the odd half a point, give a down dog a bone for Crissake....

**Justin7** · 04-25-10, 07:10 PM

Originally posted by Peep

I would have an easier time beating the closing number if I knew what it was going to be.

I would even settle for knowing what way it is going to move. You must be very smart to know that.

Hell, I would even settle for the odd half a point, give a down dog a bone for Crissake....

Sharps don't know where it will close, but they have a good guess which way it will move.

**IrishTim** · 04-25-10, 07:26 PM

If I were a good enough fortune teller that I could predict exactly where the line would close, I would just take the extra step and forsee the final score.

Peep, aren't you a modeler? You know it's all about finding instrinsic value. And when a line has value, the market should drive the value out of it (or at least most of it), although sometimes the markets seem to do irrational things in my opinion. But that can be attributed to me just hacking around with these models. I would probably save myself a lot more time just flipping coins and finish with the same result.

**Peep** · 04-25-10, 08:31 PM

Good points Tim and Justin.

I think I model differently, cause I don't find the openers and/or the line moves in my direction (necessarily).

now here is a question for you.

If a line moves against you after you bet, do you generally
A) Bet it again because it now has more value or
B) Figure you are probably wrong and just let it be?

**Thremp** · 04-25-10, 08:45 PM

Originally posted by Justin7

Sharps don't know where it will close, but they have a good guess which way it will move.

I disagree with this. Its the same age old argument of "If you can't quantify your edge, how do you know you have an edge".

**IrishTim** · 04-25-10, 09:01 PM

Originally posted by Peep

Good points Tim and Justin.

I think I model differently, cause I don't find the openers and/or the line moves in my direction (necessarily).

now here is a question for you.

If a line moves against you after you bet, do you generally
A) Bet it again because it now has more value or
B) Figure you are probably wrong and just let it be?

When creating my model, I tested its validity based more on whether the market agreed with it (as tested by the movement of the open to the close) than a W-L percentage or "units won". My whole goal really, was to create a model that could beat the market, or rather create one that put me on the side that the movers and shakers agreed with.

As you're aware, it's about expectation - what would happen if the game is played hundreds of times under these conditions. I mean, how the hell cane you predict going to "win" a one-game series? You can throw Cy Young out there against some Single A gas can and he could get roughed up for 8 runs in the first inning and lose the game. So instead of trying to "pick winners," I tried to create a model that would get me on the side the market agreed with. The winning, I hope, follows this. Obviously, it's a simple idea but not easy. The point of all this rambling is that I put a lot of stock into the markets being efficient and accurate, so when it doesn't agree with me (which is happening all too often) I'm pretty devastated.

As to your question, I don't really trust my model enough to bet it according to Kelly. I'm pretty much flat betting right now small amounts right now, trying different things out. But if I were scaling my bets according to my (perceived edge) and the market moved against me, I would try to figure out if there was an obvious reason, i.e. pitching change, injury, lineup info I missed, umpire info (for totals). If I couldn't find an obvious explanation then I would probably just assume I screwed up and let it ride, maybe try to find some book hanging a bad number that I could scalp out of the position. So answer B.

**Peep** · 04-25-10, 09:46 PM

Thanks Tim.

I don't bet more either, I just let my original bet stand.

I use a database with closing numbers to look for derivatives of the closing number, quarters, innings, stuff like that combined with other game factors. So if the number moves, I just figure the new number is more accurate, and use that to look for my plays.

**Justin7** · 04-25-10, 10:40 PM

Originally posted by Peep

If a line moves against you after you bet, do you generally
A) Bet it again because it now has more value or
B) Figure you are probably wrong and just let it be?

B. I was wrong. A good model will predict line the direction of line movement 60% of the time, and will win money in the long-term. A great model will predict it 70% of the time correctly, and will make you a fortune.

If you make a play, the line moves against you, and you don't know why, you missed something. Half of the time you can identify starting lineup changes, weather, or other factors that contributed to the move against you.

If the market does not agree with you more often than not, you're screwed. You will lose, and you will eventually lose everything you put at risk.

**Thremp** · 04-26-10, 09:35 AM

Why do you cling to this idea of only being able to discern left or right without any indication of how far left or right?

If you are unable to do this, it doesn't mean that others are as well. In addition, there are sports where the closing line is close to efficient, but people outperform expectation vs closing line (RAS totals) over fairly large samples.

You seem to be extremely narrow minded.

**sharpcat** · 04-26-10, 10:01 AM

I would estimate that roughly 75% of line movements are a result of sharp money, while the remaining line movements are a result of the public over betting teams like the Yankees or the Lakers and JM systems or RAS plays.

Therefore I personally do not hold as much value in beating the closing number sometimes beating the opening number is just as good if JM and the public are on the wrong side.

**Justin7** · 04-26-10, 10:18 AM

Originally posted by Thremp

Why do you cling to this idea of only being able to discern left or right without any indication of how far left or right?

If you are unable to do this, it doesn't mean that others are as well. In addition, there are sports where the closing line is close to efficient, but people outperform expectation vs closing line (RAS totals) over fairly large samples.

You seem to be extremely narrow minded.

I can guess where it will close. I can estimate how far it will move... but it's almost impossible to nail.

In general, the bigger the difference between the opening line and a good model, the greater the movement will be. But even good models miss fairly often.

**Thremp** · 04-26-10, 11:38 AM

Originally posted by Justin7

I can guess where it will close. I can estimate how far it will move... but it's almost impossible to nail. In general, the bigger the difference between the opening line and a good model, the greater the movement will be. But even good models miss fairly often.

http://www.handicapper.net/pick-archive.php

Stick to talking about things you know about. Between this and substituting in "power rankings" for Log5, its getting somewhat comical.

**Justin7** · 04-26-10, 12:56 PM

RAS is the nuts on NCAA totals. Do you have a point?

I'm afraid I can't find your reference to substituting power rankings for log5.

**Thremp** · 04-26-10, 01:08 PM

Bro,

Your views are wrong is my entire point.

**Maverick22** · 04-26-10, 01:20 PM

Thremp, did Justin7 sleep with your wife or something?

You take alot of shots at him... One has to wonder if you have some vendetta against him...

Just my thoughts anyways.

**Carson05** · 04-26-10, 02:45 PM

is there any special significance to 'early' or 'late' money

**Thremp** · 04-26-10, 05:53 PM

Originally posted by Maverick22

Thremp, did Justin7 sleep with your wife or something? You take alot of shots at him... One has to wonder if you have some vendetta against him... Just my thoughts anyways.

He's ignorant and arrogant. It grates on me. Using the metric of "how often does this beat the closing number" as the sole determinant of the "goodness" of models is stupid as ****. It is obvious to anyone that the ideal model would tell exactly where the line would end up and the distribution of outcomes. As we diverge from this to what really happens we get to what most people use to make money. The idea that you can't tell what kind of edge you have (First dumb point) or that you're solely looking to beat the closing number (Second dumb point) are both silly and have no basis in reality.

There are sports where you can beat the closing number and lose money. And there are models that by the Justin7 definition would be amazingly good, yet would lose money. Its just such a primitive way of analyzing handicapper success that its 100% ****ing dumb to suggest. (Flipping a coin would be an alternative method of determining the "goodness" of models.)

The Log5 thread is magically no more. How... strange. Perhaps durito and I suffer from the same affliction. (Imaginary SBR posts-itis)

**Peep** · 04-26-10, 06:12 PM

I think Justin is OK Thremp.

He tries to help posters who ask questions, and mostly sides with the players who are right in book disputes.

Of course he doesn't know everything, no one does. I disagree with him on some points, but so what? If what he does works for him in sportsbetting, I think that is great.

I would say the same for you.

**sharpcat** · 04-26-10, 06:38 PM

I once worked for a company in which I felt every single person I worked with was an asshole!

One day I said to myself "what are the odds that every single person I work with is an asshole", and it was at that point that I realized that the people around me were not the assholes but that it was me who was being the asshole.

No offense meant thremp I think you have a lot of insightful info to add to the think tank and I have zero problems with you, but a review of your posting history shows a lot of negative and insulting posts and not just to Justin7.

**IrishTim** · 04-26-10, 09:27 PM

I seem to remember KenPom endorsing someone who applied the log5 formula to his power ratings. Not for gambling but to generate win percentages for some other endeavor (bracket maybe).

**Thremp** · 04-26-10, 11:39 PM

Originally posted by IrishTim

I seem to remember KenPom endorsing someone who applied the log5 formula to his power ratings. Not for gambling but to generate win percentages for some other endeavor (bracket maybe).

Log5 is how you answer the question "How often does a .700 team (v average competition) beat a .300 team (v average competition)". Other answers are essentially variants of "flip a coin".

http://www.tangotiger.net/wiki/index.php?title=Log5

**sycoogtit** · 04-27-10, 07:52 AM

Originally posted by Carson05

is there any special significance to 'early' or 'late' money

I'm also curious about that. I found this thread in a different forum with some interesting views. http://forums.eog.com/online-sportsb...ant-37991.html

**TakeIt** · 04-27-10, 09:03 AM

Originally posted by Carson05

is there any special significance to 'early' or 'late' money

my research shows "slight" lean toward late money. for 900 nba games this year late money is + 10 units and early money is -4 units.

but that could easily reverse next year...i think

**sycoogtit** · 04-27-10, 09:12 AM

Originally posted by TakeIt

my research shows "slight" lean toward late money. for 900 nba games this year late money is + 10 units and early money is -4 units. but that could easily reverse next year...i think

That's not nearly as promising as I was hoping for. Is that for ML, spreads, totals?

**Thremp** · 04-27-10, 10:16 AM

That EOG thread is really bad. Almost everyone betting an early number is a "$100 sharp" and has the end goal of beating the closing number. They can't accurately ascertain which numbers are strong or weak without the market telling them.

**sycoogtit** · 04-27-10, 10:19 AM

Originally posted by Thremp

That EOG thread is really bad. Almost everyone betting an early number is a "$100 sharp" and has the end goal of beating the closing number. They can't accurately ascertain which numbers are strong or weak without the market telling them.

What do you think about the comment by "THE SHRINK" (http://forums.eog.com/330665-post4.html) in that thread? It sounds like you two agree.

**IrishTim** · 04-27-10, 11:08 AM

Originally posted by Thremp

Log5 is how you answer the question "How often does a .700 team (v average competition) beat a .300 team (v average competition)". Other answers are essentially variants of "flip a coin".

http://www.tangotiger.net/wiki/index.php?title=Log5

I know what it's used for, but couldn't one conceivably create a set of power rankings that, when two teams are matched up, you put their individual "power rating" through the log5 to get a projected Win %?

For instance, Duke's Pomeroy power rating of .9848 and Butler's of .9444 would give you a win percentage for Duke of 79.23% and 20.77% for Butler. That converts to a no-vig line of -381 which gives you a spread of 7/7.5 which is where the line for the NCAA championship game opened I believe.

**Thremp** · 04-27-10, 12:04 PM

Originally posted by sycoogtit

What do you think about the comment by "THE SHRINK" (http://forums.eog.com/330665-post4.html) in that thread? It sounds like you two agree.

Meh. I'm curious who these "PHONEY"s are that are betting money that moves the entire market simultaneously. I don't believe they exist.

Irish,

How you get the win % is irrelevant. I'm sure in some alternate universe you could make power rankings that are actually just expressions of team win % v league average competition. But why you would call this a "power ranking" seems absurd.

**durito** · 04-27-10, 01:42 PM

Originally posted by IrishTim

I know what it's used for, but couldn't one conceivably create a set of power rankings that, when two teams are matched up, you put their individual "power rating" through the log5 to get a projected Win %?

For instance, Duke's Pomeroy power rating of .9848 and Butler's of .9444 would give you a win percentage for Duke of 79.23% and 20.77% for Butler. That converts to a no-vig line of -381 which gives you a spread of 7/7.5 which is where the line for the NCAA championship game opened I believe.

Sure, but betting pomeroy that way is a good way to light money on fire.

Here is an interesting discussion:

SBR Forum - Sports Betting Talk - Sportsbook Review Forum

http://forum.sbrforum.com/handicapper-think-tank/122312-home-field-advantage.html

Sports betting and handicapping forum: discuss picks, odds, and predictions for upcoming games and results on latest bets.

**IrishTim** · 04-27-10, 01:58 PM

Originally posted by durito

Sure, but betting pomeroy that way is a good way to light money on fire.

Here is an interesting discussion:

http://forum.sbrforum.com/handicappe...advantage.html

I didn't mean to say that's how you should use pomeroy, but just as an example of someone who could conceivably create power ratings to which you could apply the log5. And the way my college basketball sides went after December, lighting my money on fire would have been a pleasant alternative.

I read that Ganch thread a while back, but not sure how it applies here. He derived the log5 formula from Bayes theory which is neat, but the wikipedia article Thremp posted already mentioned that.