[danshan11]

I guess I am asking what you use to determine if someone is sharp or not. Sharp to me means a long term winner and has a clear understanding of how betting really works! I am just trying to get some peoples ideas of what the benchmarks are, thanks

[SBR Drew]

Knowing when to place a bet, when to not place a bet, and mostly when to walk away .

[danshan11]

I think one thing that sticks out to me is a guy who has too many plays on a small card

say the MLB card has 10 games that means 10 sides and 10 totals
usually average roughly about a guy has 1 or maybe 2 plays out of that
and I see these guys with 12 plays, that to me is usually a big sign "this guy dont know shit" or "he just throws a bunch of shit up" but he definitely is not getting value on that many plays out of 20
I mean like I found 1 play on the WNBA playoffs and I was shocked there was one with only 2 games and all the posts I see guys picking 3 or 4 plays out of 4 available plays, that to me tells me they are FOS.

[Rich Boy]

Anyone who can consistently get an edge over the market

[danshan11]

Anyone who can consistently get an edge over the market

yeah but how do you gauge that edge? Line movement?

[Waterstpub87]

I guess I am asking what you use to determine if someone is sharp or not. Sharp to me means a long term winner and has a clear understanding of how betting really works! I am just trying to get some peoples ideas of what the benchmarks are, thanks

It is all closing line value on major stuff, can be w/l on smaller stuff. I've killed Japanese baseball and MLB props over the summer, can't really measure line value on that stuff.

I've seen you say that "not many plays" thing before. It isn't uniform. Could have 10 plays today, and then none for the week. Plenty of days I get 1 play out of both sides and totals, others days I might have 5 or 6. Still beat the close enough that I should be up money in baseball totals and sides. Depends on the time of the season. When you get to august, I tend to get very few plays, but in April it was easily 4 to 5 sides a day.

College football last year, beat the close by 1.5 pts on both the totals and sides over the entire season, had roughly 10-15 sides and 10-15 totals a week, which would be roughly 25% of the options.

[danshan11]

so you get roughly one "value bet out of 4? you dont think that seems high in general?

[Rich Boy]

yeah but how do you gauge that edge? Line movement?

Tracking a large sample of data, either the player has positive expectation or they dont, compare actual vs implied win probabilities, also compare price of wager vs closing odds

[danshan11]

Tracking a large sample of data, either the player has positive expectation or they dont, compare actual vs implied win probabilities, also compare price of wager vs closing odds

so line value over a decent sample size is what you are saying? what do you consider a decent sample size?

[Waterstpub87]

so you get roughly one "value bet out of 4? you dont think that seems high in general?

Sure, it was high. Creating a model that was closer to lines has been a project for this season. Still was generating value against the lines.

Plenty of people make 1 square play a day. Doesn't mean they are a sharp.

[danshan11]

I am not saying you cant have 10 one day, I am saying guys that are turning out huge game numbers dont usually see volume. I am not doubting your talent because your number is high to me, you might be extremely talented and be able to narrow down value plays way tighter.
example
If I make the Yankees -150 if they are -162 I cant take them because my math is not close enough to say for with any certainty that -160 is closer than me. I am in baseball about 14 cents off from a +100 game. so for me to have an MLB play its got to be more than 14 cents off on a +100 game, so that limits my plays dramatically. What I am saying is MOST guys who put out 4 plays on a 8 option card are FOS 99% of the time. that does not decide he is FOS but it is a symptom to me of someone FOS.

[Rich Boy]

so line value over a decent sample size is what you are saying? what do you consider a decent sample size?

Depends, if your average winner is around 5% then you will need a much larger sample than if its near 50%. Also depends on how confident you are in your system/backtests etc. I would be confident in a sample of 100+ games of results vs odds and prob 50+ games odds vs closing numbers

[danshan11]

Depends, if your average winner is around 5% then you will need a much larger sample than if its near 50%. Also depends on how confident you are in your system/backtests etc. I would be confident in a sample of 100+ games of results vs odds and prob 50+ games odds vs closing numbers

I dont understand what you mean, could you please give me an example?

[Rich Boy]

I dont understand what you mean, could you please give me an example?

Its about variance and probabilities. If your theoretical probability is very low or high (<10% or >90%) then it will require a larger sample in order to eliminate chance or luck from the sample. You could have a sample of 100 games with a 5% expected win rate and end up with a sample of 1-10%, huge range

If your expected win rate is around 50% then 100 games is more than enough, you wont get much more variance than 45-55%.

But if you can compare your odds vs pinny closing odds then it wont take a sample much more than 25-50 games to say you can out-handicap them consistently.

I would track both and see how they compare, also compare open vs closing line and see if they generally move in the direction you pick

[danshan11]

straight line value bet line to closing line is perfect for me but it seems like MOST people dont use that for some reason.

I also was reading a ripple theory. Big bettors with books eyeing their plays create ripples and this in turn creates a ride it or not effect. So example
Big bettor A with book respect bets a nickel on a game the line will move at that book and probably a few affiliate sites as well, this creates a ripple that might not have been there before he bet. So because other smart bettors see this bet, they tail it and this turns that ripple into a wave and can influence the closing line. You think this is true?

we are not even talking about syndicate plays or big capper plays with 1000s of followers or big money followers on their picks!

[Waterstpub87]

I am not saying you cant have 10 one day, I am saying guys that are turning out huge game numbers dont usually see volume. I am not doubting your talent because your number is high to me, you might be extremely talented and be able to narrow down value plays way tighter.
example
If I make the Yankees -150 if they are -162 I cant take them because my math is not close enough to say for with any certainty that -160 is closer than me. I am in baseball about 14 cents off from a +100 game. so for me to have an MLB play its got to be more than 14 cents off on a +100 game, so that limits my plays dramatically. What I am saying is MOST guys who put out 4 plays on a 8 option card are FOS 99% of the time. that does not decide he is FOS but it is a symptom to me of someone FOS.

I don't take things personally, no need to worry.

So you take things that are roughly 3.25% off or so. I am about 5% or so.

Depends on the sport. College lines, especially something like Toledo/Akron or UL Monroe/UL lafayette is going to be less efficient.

Most people are FOS. But betting more plays is not necessarily indicative of being a square unless you don't beat the close.

[Barrakuda]

Its about variance and probabilities. If your theoretical probability is very low or high (<10% or >90%) then it will require a larger sample in order to eliminate chance or luck from the sample. You could have a sample of 100 games with a 5% expected win rate and end up with a sample of 1-10%, huge range

If your expected win rate is around 50% then 100 games is more than enough, you wont get much more variance than 45-55%.

But if you can compare your odds vs pinny closing odds then it wont take a sample much more than 25-50 games to say you can out-handicap them consistently.

I would track both and see how they compare, also compare open vs closing line and see if they generally move in the direction you pick

Add a zero to the numbers, and you'd be a lot closer. 25-50 games means jack.

[Barrakuda]

Its about variance and probabilities. If your theoretical probability is very low or high (<10% or >90%) then it will require a larger sample in order to eliminate chance or luck from the sample. You could have a sample of 100 games with a 5% expected win rate and end up with a sample of 1-10%, huge range

If your expected win rate is around 50% then 100 games is more than enough, you wont get much more variance than 45-55%.

But if you can compare your odds vs pinny closing odds then it wont take a sample much more than 25-50 games to say you can out-handicap them consistently.

I would track both and see how they compare, also compare open vs closing line and see if they generally move in the direction you pick

So if someone hits 56 out of 100 even money bets, you think they have an edge going fwd?

[Rich Boy]

So if someone hits 56 out of 100 even money bets, you think they have an edge going fwd?

No, but if they beat the pinny closer on average over those 100 plays I would say its concrete.
What sample size you want for data is subjective and based on personal preference.

Try running simulations and see what kind of variance you are comfortable with
http://www.shodor.org/interactivate/...stableSpinner/
http://www.equitycurvesimulator.com/

[danshan11]

No, but if they beat the pinny closer on average over those 100 plays I would say its concrete.
What sample size you want for data is subjective and based on personal preference.

Try running simulations and see what kind of variance you are comfortable with
http://www.shodor.org/interactivate/...stableSpinner/
http://www.equitycurvesimulator.com/

I agree with you 1000 wins or losses does not mean shit to me compared to 100 line beaters

[danshan11]

man I just got scared to death by that simulator I set it up for my fair win% minus vig and spun that baby 10 rolls at a time and it took me 1200 rolls to get to my fair win% I would have lost for lord knows how long that is to get to profit on my edge

[tsty]

money earnt

minimum games 10k

[Barrakuda]

No, but if they beat the pinny closer on average over those 100 plays I would say its concrete.
What sample size you want for data is subjective and based on personal preference.

Try running simulations and see what kind of variance you are comfortable with
http://www.shodor.org/interactivate/...stableSpinner/
http://www.equitycurvesimulator.com/

LOL. Beating a closing line on 100 plays means nothing -- unless you're beating it by 10%.

[Barrakuda]

I agree with you 1000 wins or losses does not mean shit to me compared to 100 line beaters

Where's your evidence that 100 "line beaters" foretell future profits?

[Barrakuda]

If your expected win rate is around 50% then 100 games is more than enough, you wont get much more variance than 45-55%.

What does this mean?

[TheMoneyShot]

I always thought being a SHARP meant... you place your wager moments before the line moves against your initial play. Creating the best opportunity possible... best bang for your buck... and winning that play.

I've been doing this for nearly 20 years now... I technically play hungry dogs 80% of the time with my formulas...

Beating the closing line is sort of useless for me... 60%+ of the time you'll get the best line near closing time.

[danshan11]

Where's your evidence that 100 "line beaters" foretell future profits?

I dont have any evidence of anything that is why I asked. I THINK that I value the 100 beaters more than I value a 1000 game winner. I mean do it yourself go play with the simulator and you will see real quickly that skill is far far away and 1000 games is not enough it can take 10000 or more to get to your true win%. http://www.shodor.org/interactivate/...stableSpinner/
So if wins and losses mean zero which the simulator proves we got to use beating the line. I think if you bet 100 games and can by average beat the margin you have a chance but man its thin and forever.

[tsty]

100 games lol

best thing ive ever heard

i clear 100 bets a day fairly often

100 aint shit

[danshan11]

100 games lol

best thing ive ever heard

i clear 100 bets a day fairly often

100 aint shit

I think that is smart if you can create an edge on that many games daily and bet them you will cover variance very quickly and get to your true win% faster and more steady.

[tsty]

I think that is smart if you can create an edge on that many games daily and bet them you will cover variance very quickly and get to your true win% faster and more steady.

if you can model one sport why cant you do another?

the only difference is the variables in play

[Rich Boy]

What does this mean?

My point is that with lower win rate systems you need more precision in the data to have confidence since variance has a large impact on the results. 1 more win or loss in a 100 game sample at 10% expected win rate represents a huge difference, compared to 50%. But even if you have that data it doesnt guarantee that system will continue to work. Thats the biggest problem with systems and backtesting, in an efficient market past success doesnt guarantee future results and I have seen that first hand in sports over the years, books are quick to catch on to positive expectancy players, they watch you, stalk you like a fkn hawk and all the bets you place you are essentially giving them free info on their weaknesses at a small cost to them, they shore up those lines and you are on to the next system that works for a while and the cycle repeats.

That is the reason why I decided to give up sports betting, the edges are not as good as they used to be and books are constantly adjusting and becoming more efficient.

[danshan11]

if you can model one sport why cant you do another?

the only difference is the variables in play

I do NBA, WNBA, MLS, Liga MX, CFL, MLB and even with that I only get maybe 5 or 6 plays a day average.

[Thicht]

I think Danshan that a big reason that you are hesitant when a guy has a lot of volume is because you may be used to touts/services releasing very selectively and so you may have been conditioned to believe that is correct. But what you don't see is behind the scenes these touts/services may be unloading a great deal more volume. Most services aren't doing much with 1st halves and 2nd halves and even the main side/total action may be something like 10-20% of their total action.

[tsty]

My point is that with lower win rate systems you need more precision in the data to have confidence since variance has a large impact on the results. 1 more win or loss in a 100 game sample at 10% expected win rate represents a huge difference, compared to 50%. But even if you have that data it doesnt guarantee that system will continue to work. Thats the biggest problem with systems and backtesting, in an efficient market past success doesnt guarantee future results and I have seen that first hand in sports over the years, books are quick to catch on to positive expectancy players, they watch you, stalk you like a fkn hawk and all the bets you place you are essentially giving them free info on their weaknesses at a small cost to them, they shore up those lines and you are on to the next system that works for a while and the cycle repeats.

That is the reason why I decided to give up sports betting, the edges are not as good as they used to be and books are constantly adjusting and becoming more efficient.

you're delusional if you think the books are following you lol

even more so if you think they care about "systems"

the real reason is that your system never worked

systems in general never work

the only way to model sports is with regression , that's it

none of this coincidence bs

[reigle9]

i thought consistently beating the closing line was the textbook definition