Working on your Sports Model and betting with an Edge

[coolguy73739]

I have been working on my Basketball Model for years and hoping to bet with AN EDGE this season.

Some of the Great Brains on this forum had mentioned that Modelling is Dead and nothing works better than so called Adjustments to injury, toughness of schedule, beating the closing lines, Wisdom of the crowd and all other FANCY terms.

Is it true that you can never know your true EDGE before placing your bets, thoughts?

[VeggieDog]

Modelling is NOT dead. Sportsbooks use models to set their opening lines. Keep doing what you're doing and make adjustments as you see fit. Also, remember to look for the unexpected. Don't get tunnel vision. Good luck to you!

[d2bets]

It's not dead, but you have to think how you're going to make a consistently better model than the sportsbooks which have bright minds, lots of money and fancy computers. I don't expect you to spill it here, but you better know what makes your model better than them.

[KVB]

It's not dead, but you have to think how you're going to make a consistently better model than the sportsbooks which have bright minds, lots of money and fancy computers. I don't expect you to spill it here, but you better know what makes your model better than them.

When you talk about sportsbooks and their models wht do you mean?

You mean the opening line? Closing line? Line movement? Where do you think the sportsbook is expressing their "model"?

Because we routinely get inside the line and lines routinely move, sometimes back and forth, sometimes just one direction. Are we beating a model?

What's the goal of the sportsbook's model?

Are they trying to beat the boo...oh wait, they are the books.

Some questions to ponder there for sure.

[Waterstpub87]

Good luck coolguy.

Modeling is not dead.

BTW, College or NBA or europe?

[KVB]

Agreed, with the introduction of sports betting just growing daily to the general public, this could be another little golden age.

We welcome the new money and the public's penchant for the Favorite and Over is timeless and continues to provide a positive field from which to potentially choose.

...Some of the Great Brains on this forum had mentioned that Modelling is Dead and nothing works better than so called Adjustments to injury, toughness of schedule, beating the closing lines...

Even when you talk about adjustments to injury, toughness of schedule, beating the closing line, all of those things you mention as some kind of alternative to modeling, quite frankly, revolve around modeling.

[slapshot]

Is it true that you can never know your true EDGE before placing your bets, thoughts?

yes it's true you can never know your true edge before placing a bet because human behavior can't be quantified.

example....how can game deciding events like
a pick 6 in football
a red card send off in soccer
a two minute 5 on 3 in hockey
an early inning pitching injury in baseball
be incorporated into an edge calculation?

you can estimate the chance of your team overcoming these events or that they won't happen.....but you can never know beforehand.

[Waterstpub87]

yes it's true you can never know your true edge before placing a bet because human behavior can't be quantified.

example....how can game deciding events like
a pick 6 in football
a red card send off in soccer
a two minute 5 on 3 in hockey
an early inning pitching injury in baseball
be incorporated into an edge calculation?

you can estimate the chance of your team overcoming these events or that they won't happen.....but you can never know beforehand.

You're right, but it goes far beyond that. A good way to think of it is " if the game was played 5000000 times, what would the results be" with a large enough sample, and the correct probabilites, you would have the true odds. This would incorporate all the scenarios above, and all the other ones that are possible.

Most will simplify with the following assumption:

Vig free Closing line probability = actual probability

I don't believe this to be exact, but it is better than any other estimation. If you run the data, in most sports, taking the opener on lines that moved significantly, you can see how much value you get. I would encourage everyone to do this, it will dramatically increase your understanding of the math behind sports wagering.

In short, you cannot know actual edge when you place the bet, but you can get estimated edge at tip off.

[coolguy73739]

Thank you for your valuable insight Gents.
I was specifically referring to Basketball since I always enjoyed working on its analytics.
I have done back testing of last season NBA, NCAAB and some of the international leagues and here are some of my findings based on my Model:
A) For a large number of games, I would say about 75%, the numbers are random (point spread and total points) and line movement, wisdom of the crowd, closing lines dictates the proceedings. This is because Linemakers themselves are not very confident of their numbers and hence don't mind actions on both sides and living off the vig or commission for those games.
B) For the remaining games, they definitely have plans and are confident of their numbers and are ready to Gamble on those games..(one side action).
For these select 20-25% games ( percent could be less or more on day to day basis), closing numbers, line movement have no meaning imo since books are firm on where they want to keep their numbers ( PS and totals) till the very end.
C) For average bettor ( me included but hoping to change after Modeling these games), its hard to differentiate these two category games and we remain perplexed and lost in finding out our edges.

A good Model is the one which can give you clear insight of games A and B above and then you can bet confidently knowing your EDGE based on the numbers.
An EDGE is a relative term.. I believe it has to be compared with something. For example, Specific Gravity or Relative Density is a relative term and its always compared with that of water which has been given a value of 1000 kg /m3 and then all other fluids are compared to that number, heavier or lighter.
In a similar fashion, If we assign a number 100 to Bookies, We have to compare with that number to check how close or far are we from that number, since they know their games inside out and finally we are trying to Outsmart those brainy guys in the business of betting.
Not sure if I am making a sense here, but that's my definition of EDGE in betting terms.

[coolguy73739]

Good luck coolguy.

Modeling is not dead.

BTW, College or NBA or europe?

Thanks Mate.. All leagues included..

[Waterstpub87]

Thanks Mate.. All leagues included..

I like this idea in theory. Not my approach, but that doesn't mean anything.

I would assume there would be more of the type B games in NBA, less in NCAA, and less internationally. It might be interesting to compare this based on book. Can you find games that should be B in europe, but are treated like A in the US, would a Bet365 or whatever they have over there have more insight.

Seems a little bit like reverse line movement type stuff. Getting the accurate information is the tough part, you could look on SBRodds and another site, and they would give you opposite numbers.

The phenomenon you are describing has been referred to as "a Pinny lean" from Pinnacle.

[coolguy73739]

Need to dig deeper but I believe all games are treated with same Category (A or B) universally else there will be conflict of interest amongst Bookmakers..

[TommieGunshot]

Is it true that you can never know your true EDGE before placing your bets, thoughts?

You can always estimate your edge and put a range on it. How confident you are in this estimate is completely up to you.

[TommieGunshot]

example....how can game deciding events like
a pick 6 in football
a red card send off in soccer
a two minute 5 on 3 in hockey
an early inning pitching injury in baseball
be incorporated into an edge calculation?

They can be incorporated fairly easily. Just determine the probability of each event. If the probability is high, then those events should be incorporated significantly. If the probability is low, then those events should be not be incorporated much at all.

[coolguy73739]

You can always estimate your edge and put a range on it. How confident you are in this estimate is completely up to you.

Range? Can you pls elaborate..

[TommieGunshot]

You can always estimate your edge and put a range on it. How confident you are in this estimate is completely up to you.

Range? Can you pls elaborate..

An example would be for me to determine that my edge is between 2 and 3%.

[ChuckyTheGoat]

I have been working on my Basketball Model for years and hoping to bet with AN EDGE this season.

Some of the Great Brains on this forum had mentioned that Modelling is Dead and nothing works better than so called Adjustments to injury, toughness of schedule, beating the closing lines, Wisdom of the crowd and all other FANCY terms.

Is it true that you can never know your true EDGE before placing your bets, thoughts?

Cool, good post. I wish you the best.

I watched a video from a great investor. He made one great point:

*If you can't identify your edge (within a very small margin)...you are likely not playing with an edge.

Bottom line, you don't bet UNLESS you can identify your edge:
*What is your angle?
*Have you proven your angle PROSPECTIVELY?
*Is there any reason to think this angle might STOP working?

Truth is that people are flawed. No reason to think u can't find an angle that will clear a profit. Have patience and try to clear a profit.

[ChuckyTheGoat]

Range? Can you pls elaborate..

Cool...u want one CollBask angle to ponder? Consider this:

*The Duke/North Carolina is USUALLY very sharp.

Too much $$ on this game. House is motivated to put up a sharp #.

Compare that to the Woff/Samford game. 100x as much risked on Duke/UNC. Focus on small College games where you're more likely to identify a line error.

[coolguy73739]

Cool...u want one CollBask angle to ponder? Consider this:

*The Duke/North Carolina is USUALLY very sharp.

Too much $$ on this game. House is motivated to put up a sharp #.

Compare that to the Woff/Samford game. 100x as much risked on Duke/UNC. Focus on small College games where you're more likely to identify a line error.

This is exactly I want to find out, how to differentiate between a SHARP nd ERROR LINE.. Can you do it without Modeling? How to quantify your edge without numbers?

[coolguy73739]

They can be incorporated fairly easily. Just determine the probability of each event. If the probability is high, then those events should be incorporated significantly. If the probability is low, then those events should be not be incorporated much at all.

Seriously, can you calculate the probability of a red card send off in Soccer?

[TommieGunshot]

They can be incorporated fairly easily. Just determine the probability of each event. If the probability is high, then those events should be incorporated significantly. If the probability is low, then those events should be not be incorporated much at all.

Seriously, can you calculate the probability of a red card send off in Soccer?

Fairly easily. Just look at red card betting odds. Or use booking points odds.

[ChuckyTheGoat]

How to quantify your edge without numbers?
Cool, I think you should remove this idea from your list. It's a Numbers game. If somebody tells you they can beat the game with rationalization, we want to book that person.

I'm saying that you need to develop an unbiased method of projecting the final for today's game. It's best if u do it BLINDLY b4 u see the posted #. Then compare YOUR projection vs posted # over time. If you're to the +, you're good.

Also, winning bettors are constantly putting +EV tickets in their pocket. BTCL (Beat The Closing Line) Value is correlated to the winning bettor. Look at the winner's ticket stack as games are kicking off. Invariably, he will have a lot of tickets that look like:

+8.5, when closing line was +7
+1.5, when closing line was -1.5, etc

[coolguy73739]

Its impossible to project a number unless you know what's the bookies line is..
My model based on total of 225 for an NBA game has no meaning if bookies posted line for the total is 224.5..
I already mentioned those will be falling in category A of my games and I will skip them
I can only compete against the 800 pound Gorilla the so called Bookies based on the offset of theirs and my number..
My Edge is relative compared to bookies number unless someone can teach me otherwise..

[ChuckyTheGoat]

Its impossible to project a number unless you know what's the bookies line is..
My model based on total of 225 for an NBA game has no meaning if bookies posted line for the total is 224.5..
I already mentioned those will be falling in category A of my games and I will skip them
I can only compete against the 800 pound Gorilla the so called Bookies based on the offset of theirs and my number..
My Edge is relative compared to bookies number unless someone can teach me otherwise..

Cool, I'm putting u on Ignore. I don't have extra brain-cells to burn. Good Luck.

[Kindred]

Cool...u want one CollBask angle to ponder? Consider this:

*The Duke/North Carolina is USUALLY very sharp.

Too much $$ on this game. House is motivated to put up a sharp #.

Compare that to the Woff/Samford game. 100x as much risked on Duke/UNC. Focus on small College games where you're more likely to identify a line error.

[Kindred]

Its impossible to project a number unless you know what's the bookies line is..
My model based on total of 225 for an NBA game has no meaning if bookies posted line for the total is 224.5..
I already mentioned those will be falling in category A of my games and I will skip them
I can only compete against the 800 pound Gorilla the so called Bookies based on the offset of theirs and my number..
My Edge is relative compared to bookies number unless someone can teach me otherwise..

https://www.sportsbookreview.com/betting-calculators/half-point-calculator/

[WireWire]

How to quantify your edge without numbers?
Cool, I think you should remove this idea from your list. It's a Numbers game. If somebody tells you they can beat the game with rationalization, we want to book that person.

I'm saying that you need to develop an unbiased method of projecting the final for today's game. It's best if u do it BLINDLY b4 u see the posted #. Then compare YOUR projection vs posted # over time. If you're to the +, you're good.

Also, winning bettors are constantly putting +EV tickets in their pocket. BTCL (Beat The Closing Line) Value is correlated to the winning bettor. Look at the winner's ticket stack as games are kicking off. Invariably, he will have a lot of tickets that look like:

+8.5, when closing line was +7
+1.5, when closing line was -1.5, etc

I think this is way more important say when you are betting NFL where the lines are sharp/tight, NHL not so much, While I do believe it is important obviously I don't think it's the end all be all. If your a model has a proven long term edge over Vegas at line X, line movement shouldn’t have an effect in the LR, Great thread btw.

[coolguy73739]

I think this is way more important say when you are betting NFL where the lines are sharp/tight, NHL not so much, While I do believe it is important obviously I don't think it's the end all be all. If your a model has a proven long term edge over Vegas at line X, line movement shouldn’t have an effect in the LR, Great thread btw.

[coolguy73739]

https://www.pinnacle.com/en/betting-articles/educational/revisiting-randomness-in-betting/S3R23DEMVZRLEQ9M