4% ROI possible in NBA??

[339955]

i just backtested a model i made against NBA totals and i got 4% roi over 2,000 bets. is this possible?? i didn't use any complicated math like bayes just some simple ideas but i thought NBA was hard to beat. what do you think?

[matthew919]

Can you describe your training and testing procedure in more detail?

[Miz]

All the standard comments apply....
Success rate is easier to digest than ROI.
Is it a sim, regression based, logistic regression based, etc?
Did you do better against openers or closers?
Did you ensure you had no data from the "future" in your testing?
Did your success rate improve as the difference between your projection and line increased?

Those are the kinds of things you'd like to know

[Miz]

Let me add something to that. People are under the impression that the math has to be complicated to be successful. Not true at all. I do far more complicated math at work all day than what I do in sports betting. Basic linear algebra does wonders.

[339955]

hey miz thanks again for the response. i understand some of your comments but am unclear on what you mean by a few of them. could you explain

"sucess rate is easier to digest than ROI"

what do you mean by "sim"?

[Miz]

Success rate just means winning percentage. For example if you had 55% over 2000 plays, that would be very good. ROI can be defined differently by different people, and be staking dependent. It is simpler if you just state success rate for the readers. Sim means simulation. i.e. Monte carlo simulations

[oversnow]

[hutennis]

Success rate just means winning percentage. For example if you had 55% over 2000 plays, that would be very good.

Just to make sure while we at it...

That would be very good only if you are expected to be right less (to a statistically significant degree) than 55% of the time.
Otherwise, where is the beef?

[Miz]

Yep, fair point. I was assuming a point spread where the expected success rate is 50%

[HUY]

i just backtested a model i made against NBA totals and i got 4% roi over 2,000 bets. is this possible?? i didn't use any complicated math like bayes just some simple ideas but i thought NBA was hard to beat. what do you think?

If you didn't split your data into training and testing segments then your results are meaningless.

[fistacuff83]

how many seasons do these 2000 games go back

[elchristo]

Success rate just means winning percentage. For example if you had 55% over 2000 plays, that would be very good. ROI can be defined differently by different people, and be staking dependent. It is simpler if you just state success rate for the readers. Sim means simulation. i.e. Monte carlo simulations

if you bet $20 a game, how much would you make over 2000 plays? im tryin to figure out what id make with a 1000 budget

[HeeeHAWWWW]

if you bet $20 a game, how much would you make over 2000 plays? im tryin to figure out what id make with a 1000 budget

Starting bankroll of $1k, staking at 1/3 Kelly, 55% at -110, yields a median of around $4.5k. If you can get -105 (eg at pinnacle), median rockets up to $24k.

[Miz]

20*[(1100)-1.1(900)] for -110 for flat betting

[elchristo]

Starting bankroll of $1k, staking at 1/3 Kelly, 55% at -110, yields a median of around $4.5k. If you can get -105 (eg at pinnacle), median rockets up to $24k.

can you guys explain to me what Kelly means? I've seen justin7's videos on youtube and i still dont get it. I'm sorry i'm really bad at math but i'm trying to understand it. I got Sharp Sports Betting coming in the mail, does that book explain it step by step?

[matthew919]

In Kelly, your wager size is determined by your bankroll size, as well as the odds given and the true estimated odds for the event (the difference of which determines your perceived edge). It is derived as an optimization problem using basic calculus. The hazy part of the formula is estimating the true odds - something which no one ever really knows for sure.

[Miz]

i'm really bad at math

You can change this.

[byronbb]

Do totals ever move over 1.5 points from open to close?

[hutennis]

In Kelly, your wager size is determined by your bankroll size, as well as the odds given and the true estimated odds for the event (the difference of which determines your perceived edge). It is derived as an optimization problem using basic calculus. The hazy part of the formula is estimating the true odds - something which no one ever really knows for sure.

The calculus is in deed very simple.
Basically, Kelly tells you that for optimal growth you should wager to win the amount that equals to your BR times your edge.

If your BR is $10000 and your edge is 2% then should be wagering to win $200.
If your odds are -200 then you wager $400.
If your odds are +200 then you wager $100.

So calculations are really easy.

Be very mindful about that "hazy" part. Make sure that your 2% edge is not just a fantasy in your head but a real number or you will get
slaughtered by using Kelly.

[elchristo]

You can change this.

i am tryin to thats why i am asking for help

[Sawyer]

%4 ROI is not something sensational..even in NBA. Please open a thread when you make %12-15.

[HUY]

%4 ROI is not something sensational..even in NBA. Please open a thread when you make %12-15.

Sawyer you are a clown.

[matthew919]

Why not shoot for 30

[339955]

sorry sawyer i am not good as you bro i wil lkeep working to get there though u an inspirationt o us all

[Raizen]

Success rate just means winning percentage. For example if you had 55% over 2000 plays, that would be very good. ROI can be defined differently by different people, and be staking dependent. It is simpler if you just state success rate for the readers. Sim means simulation. i.e. Monte carlo simulations

That must be in USA cause in the rest of the world the win rate is almost the more insignificant stat.

[Miz]

That must be in USA cause in the rest of the world the win rate is almost the more insignificant stat.

You're correct. I was assuming a point spread sport with an expected win pct of 50%; and this is nearly exclusive to the US.

The only sports dominated by moneyline play is baseball and hockey here, although moneylines are posted for many sports.

[PAULYPOKER]

4% ROI possible in NBA??

Only if you are a personal friend of Stern.................