Why Americans always lay to win? "Risk" is much better option in my opinion..

[donjuan]

What makes you think it will be negative? You didn't see my record I guess

Because you fail to understand absurdly simple concepts that a 5 year old would understand.

[Karayilan9]

I think what Sawyer is trying to say isn't how its coming across

-120 has a better chance of winning than a +120 in a perfect world but when it comes down to a small sample you could lose all your -120's and win all your +120's. If you wagered every game in the market for a few seasons it is likely to be a different story.

I think he means, don't automatically think that your -150 will win and your +150 will just because of the odds.

There is a saying that Sawyer will know,

In life one person understood me but he understood me wrong

[djiddish98]

That doesn't explain why he's flat betting to risk regardless of odds, which was the original argument in the thread.

The optimal method would be to (Per Ganch's kelly article):

bet an amount such that the percentage of his bankroll he stands to win is the same as his percent edge.

[wrongturn]

Sawyer must consistently think his + odd bets have more edge than his - odd bets, even he may not realize that. But right answer is to bet more on higher edge bets no matter the odd is + or -. How much more? That is another topic, but nobody says you have to bet exactly 250 or 100 for -250 odd though.

[PAULYPOKER]

Is that Winnie Cooper from Wonder Years?

Holy Phuck Dad you have an eye for Pussy,,I think it just might be since you mentioned it which brought it to my attention................



Anyone else think this is Winnie??

[trixtrix]

that huge forehead detracts from her overall appeal imo

[Pokerjoe]

%65 x 1.59 (-170) = 103
%54 x 1.91 (-110) = 103

Odds can be misleading.. What makes you think -150 has a better chance to bring profit then -130? No, it's not! Don't get fooled by odds..

Short odds (1.50) may win in a higher rate then bigger odds (1.80) but it doesn't mean that they will give a better yield/return on investment.

But majority of bettors think their winning chance will be better if they bet on short odds like 1.20-1.30. However, these odds can be misleading. MAny times, you'll win your -110 and -120 bets but your -150 will burn you.

Well, this is true. All that matters is EV, and favorites aren't inherently more EV than dogs, and in fact tend to be slightly more -EV, as an average.

[Pokerjoe]

Here's the trap bookie's lay for recreational bettors (by definition, a professional bettor doesn't fall for it): by setting up software so that putting in $100 bets at -150 means actually betting $150, the bookie seduces extra action on faves, and generally that's where the bookie's profit is.

The idea that big faves are subject to less variance is an illusion. You are less likely to lose your bet with them, but lose more when you do.

[Thomas_Garber]

Sawyer you are completely missing what LT and few others are trying to tell you. Perhaps it is because English is not your first language.

Coffee isn't your cup of tea, eh?

[Ganchrow]

The idea that big faves are subject to less variance is an illusion.

No illusion here ... just rather very basic mathematical fact.

For a background see my old post on Calculating Wager Variance.

Let X = Bet Size
Let E = Bet Edge
Let O = Decimal Odds on Bet
Let V = wager variance

then we have:

V = (O-E-1) * (E+1) * X²

And so:

∂V ∂O = X²*(E+1)

So in other words, because E > -1, holding bet size and edge constant, variance clearly increases (linearly) with decimal odds.

If we restate the variance equation with X as the dependent variable then we have:

X = V / √(O-E-1)(1+E)

Giving us:

∂X ∂O = - ½ * (1+E) * V / ( √(E+1) * (O-E-1) ) ³ < 0

So if we held edge constant and varied odds, then in order to maintain constant variance, we'd need to decrease our bet size as a decimal odds increased. Another way of phrasing this is that if two bets are of the same size and at the same edge, then the bet on the bigger under dog will have higher variance.

And finally:

∂V ∂E = X² * ( O - 2 * (E + 1) )

So holding odds constant, variance increases with edge when:

O > 2 * (E + 1)

And variance decreases with edge when:

O < 2 * (E + 1)

So holding odds constant, variance is at a maximum (verify ∂²V ∂E² = -2 < 0) when:

E = O 2 -1

So at even odds variance is at a maximum (V = X²) when edge = 0. So with negative edge, increasing your edge towards zero will also increase your variance, and with positive edge increasing your edge away from zero will decrease your variance.

For odds of variance is at a maximum (V = 2¼ * X²) when edge = 50%. So variance increases as you increase your edge towards +50%, and decreases as you increase your edge away from +50%.

And for odds of variance is at a maximum (V = 9 16 * X²) when edge = -25%. So variance increases as you increase your edge towards -25%, and decreases as you increase your edge away from -25%.

[Sawyer]

The problem is, many times lines are not fair. Vegas offers inflated lines (like -180 for a -130 game or -290 for a -180 game) so they make more money when player lose. Players are in love with huge chalk favourites. They feel more safe when they jumped on a -320. That's what I wanted to tell you here but only few people were able to understand. -150 won't make you more money then a -130. If you think this way, you won't get ahead of books..

I don't say +150 dog and -180 have equal chance to win.

[LT Profits]

The problem is, many times lines are not fair. Vegas offers inflated lines (like -180 for a -130 game or -290 for a -180 game) so they make more money when player lose. Players are in love with huge chalk favourites. They feel more safe when they jumped on a -320. That's what I wanted to tell you here but only few people were able to understand. -150 won't make you more money then a -130. If you think this way, you won't get ahead of books..

I don't say +150 dog and -180 have equal chance to win.

IMPOSSIBLE, sharps would pound the shit out of the dogs in those cases and the price would move toward fair value.

[Sawyer]

Do you know what is Reverse Line Movement?

Sometimes, Sharp action moves a -280 fav to -250.

Not always but many times, lines are not fair. In addition, don't forget that every books has different odds. Of course, a -200 fav will win more ofthen then -120 but -200 fav won't give you a better return since you need to win %66.67+ to make profit if you lay -200.

The conclusion is,

If you liked a huge chalk favourite, don't lay -250. Just lay 100 to win 40.

Have a great monday everyone!

[PAULYPOKER]

Holy Fukk are you guys still arguing about this shit,I can confidently tell you that you guys will never come to an agreement with one another so why don't you just save your breath....................

[BillyCostigan]

This is the dumbest thread ever. You might lose less but you'll also WIN LESS if those bets hit. It's going to even out.

[BillyCostigan]

Nice NBA record though, Sawyer.

[LT Profits]

This is the dumbest thread ever. You might lose less but you'll also WIN LESS if those bets hit. It's going to even out.

Not if you only make +EV bets, then you are losing money by betting risk.

[Flight]

Uhh.... Holy Ganchrow sighting, Batman!

[BillyCostigan]

LT I was referring to the original example.

[Flight]

V = (O-E-1) * (E+1) * X²

And so:

∂V ∂O = X²*(E+1)

I had no idea you could do partial derivatives in BBCode. How you got the lower case delta is beyond me.

Great post.

[Ganchrow]

I had no idea you could do partial derivatives in BBCode. How you got the lower case delta is beyond me.

Great post.

It's part of some custom BBCode (see SBR's BBCode listing) I wrote back when I was in SBR's employ. It's nonstandard and to be honest I don't think I've ever seen anyone use them other than myself. LaTEX is of course much easier on the eyes, considerably more versatile, simple to use, and a universal standard.

You can display any html character code with the [hcc][/hcc] tags.

So you'd use [hcc]part[/hcc] to display ∂ and [hcc]delta[/hcc] for δ.

[Pokerjoe]

Ganchrow,

You said:
holding bet size and edge constant, variance clearly increases (linearly) with decimal odds.

Yes. But wouldn't increasing bet size (which is what happens when you bet $150 on -150 faves as opposed to, say, the $100 you'd bet on an even money play with the same EV) put the variance (per BR size) right back in?

However, if you bet that same $100 on a +150 dog, then, yes, you'd be suffering more variance.

I look at it this way: if you had a 10k BR, and bet $100 per game generally, and liked a bet on a 100-1 fave (same EV as all your other bets), would it be really low variance to bet your whole BR on it, which is what would happen if you just punched $100 into the typical SB software?

[Pokerjoe]

I'll say it again: US style odds and software is designed to increase betting. IOW, it's designed to increase the bookie's profits, not yours.

Typically, rec players will want to bet $100 on a game at -110 and will, without thinking about it, find they've bet $110 instead. That's fine, few of us here are going to bet anything we don't want. But the rec player, yes, ends up betting 10% more than they'd really thought about betting, which, yes, adds up, from the bookie's POV.

[Pokerjoe]

In fact, Ganch, tell me if my betting makes sense to you.
At 1.50, (-200), I'd bet .67 to win .33
At 2.00, (+100) I'd bet .50 to win .50
at 3.00, (+200) I'd bet .33 to win .67

The formula for this is obv, but I've never actually worked out whether this is the best approach. It's AN approach, and consistent, and objective, and is still multiplied by BR and adjusted for bet size by EV, of course.

Is it variance neutral?

[dogman]

I know alot of people here are knocking this thread but this question is something that has been on my mind for some time. I've tried all 3 ways of betting that has been mentioned here but never sure which one was the most profitable. There has to be a right answer for this, no?

I've heard all 3 arguments and they each have their points. In my career of sports betting(35 years), LT's way was the one I would mostly use, then in the last year or so I would play the way Saywer mentioned and also sometimes I would bet the way Pokerjoe mentions in the post above.

[Ganchrow]

But wouldn't increasing bet size (which is what happens when you bet $150 on -150 faves as opposed to, say, the $100 you'd bet on an even money play with the same EV) put the variance (per BR size) right back in?

However, if you bet that same $100 on a +150 dog, then, yes, you'd be suffering more variance.

Correct.

Let W = Win amount of bet

Then:

W = X * (O-1)

If we lock in the win amount at 1 unit then:

X = 1 O-1

And:

V = (O-E-1) * (E+1) / (O-1)²

∂ ∂O V(O,E | X = 1 O-1 ) = E+1 (O-1)³ * (2*E - O + 1)

The sign of which is strictly dependent on (2*E - O + 1).

So holding win amount constant, then provided E < O-1 2 (which would hold in all but the most extraordinary of circumstances -- consider that for a bet at -110 this would necessitate a win probability > ~ 76.19%) variance will in fact decrease with odds.

But as I've said before, the propensity for recreational and other bettors to set wager amounts differently on favorites than on underdogs is little more than an artifact of the arcane system of quoting US-style odds.

It's really pretty simple, if one keeps bet size constant then variance will increase with odds, while if one keeps win size constant then variance will decrease with odds.

And of course, as previously stated, in order to hold variance constant one would decrease bet size with increasing odds (just not by so much as to keep win amount constant). It works out so that at unit variance and zero edge, a player would need to bet 1 / √O-1 units to hold his variance constant. (So for example, assuming 1 unit at +100, that'd be 0.5 units to win 2 at +400, 2 units to win 0.5 at -400, etc.)

But this all really begs the question of why one would want to hold variance constant. Variance is just one statistic among many at which a quantitative bettor should be looking. And while such a player should certainly be aware of how his bet sizing methodology impacts the variance of his results, he shouldn't be held captive by it either.

In fact, Ganch, tell me if my betting makes sense to you.
At 1.50, I'd bet .67 to win .33 or lose .67
At 2.00, I'd bet .50 to win .50 or lose .50
at 3.00, I'd bet .33 to win .67 or lose .33

The formula for this is obv, but I've never actually worked out whether this is the best approach. It's AN approach, and consistent, and objective, and is still multiplied by BR and adjusted for bet size by EV, of course.

I agree ... it's AN approach, and it's consistent, and it's objective ... but that's a pretty low hurdle rate to overcome. In short, what's the point?

Two of the most important questions for a player to ask himself when considering a staking strategy are:

1. What are the player's goals in terms of acceptable risk and desired return (and more generally in terms of the shape of his results distribution)?
2. How does the betting scheme in consideration further that set of goals?

But if you're a positive EV bettor, then step 1 should really be to investigate Kelly. It's straightforward and provides for objective utility maximization across a wide range of risk preferences (via the Kelly multiplier).

Now Kelly certainly has its problems, but a positive EV bettor who isn't using it, needs to be very certain he staying away for the right reasons. It's my firm belief that Kelly (at a suitable multiplier level) should be the default betting strategy for all bettors with quantifiable edge and that deviations should only be attempted after very careful study and analysis.

So in other words if you're a +EV bettor and and don't know what staking strategy to use ... use Kelly.

(And if you're not a +EV bettor? Bet within your means, stick to reputable books, shop for the best lines available, take all the bonuses you can get, and save for this parenthetical feel free to ignore all my posts in their entirety and just have as much fun as possible. )

[Pokerjoe]

Ok, my simple understanding of Kelly is to bet the estimated % advantage. If you estimate a 1% advantage, bet 1%.

My method accounts for this, in that if 1% of my BR was $100, then,

At 1.50, (-200), I'd bet .67 to win .33, or $133 to win $67
At 2.00, (+100) I'd bet .50 to win .50, or $100 to win $100
at 3.00, (+200) I'd bet .33 to win .67, or $67 to win $133
(rounding errors, obv)

If I thought my EV was double that, I'd double those bet sizes.

This is just the method for dealing with bet sizing I've come up with by "feel" you might say.

[LT Profits]

Ok, my simple understanding of Kelly is to bet the estimated % advantage. If you estimate a 1% advantage, bet 1%.

Not quite. Kelly says if you have a 1% edge, you bet TO WIN 1% of your bankroll. Which goes right back to my original point.

[Pokerjoe]

So, LT, if you thought you had a 1% edge backing a 100-1 fave, you'd put your whole BR on the line.

[Ganchrow]

So, LT, if you thought you had a 1% edge backing a 100-1 fave, you'd put your whole BR on the line.

1% edge on a 100-1 fave would equate to a 100% probability of winning.

Realize Kelly doesn't account for the possibility of simply being wrong.

[JoeVig]

Holy cow this topic is still going on?

[Pokerjoe]

1% edge on a 100-1 fave would equate to a 100% probability of winning.

Realize Kelly doesn't account for the possibility of simply being wrong.

You're saying it isn't possible, then, to have a 1% edge on 100-1 fave (because it certainly isn't possible to have a 100% chance of winning).

Certainly you can only approach a 1% edge but never get there (true odds of 10,000-1 on a 100-1 prop would be 0.99% of an edge).

I suppose my point is this, and it comes from experience at estimating edges (which is the case in almost all real world handicapping, that is, in situations where you are only estimating edges, and don't actually know them): the greater the variance from a 50/50 proposition, the greater the magnitude of error, or penalty of error, or however you want to put it, if you're wrong. IOW, whether backing or laying on a 100-1 proposition, the effect of any estimation error on your part is huge, approaching catastrophe, and thus, as a proposition varies from 50/50, I lessen my risk.

Thus: (units)*1/(UK-style odds). Whatever, I'm comfortable with it. And I'd even go so far as to say that anyone thinking they can measure advantage with enough precision to warrant risking huge percentages of their BR is just waiting for a black swan to bite them on the ass.

[terpkeg]

I am failing to see how most of this discussion is relevant to risk/to win unless you assume that every risk bet is risking the same amount.

I am a risk better because, like Sawyer mentioned, the real question is: how much can you your bankroll afford to RISK.

What if I am a TO WIN better who finds an "edge" in a +100 line, he thinks the side has a 55% of winning and bets 1% or 1unit. The same to win better finds a 15% edge in another play, this play is -400, but in reality, it should be -1900 because 95% the play should win. I think a 15% edge is forth a 5% play, my max. So do I actually bet TO WIN 5%, risking 20% of my roll on ONE PLAY.

Even if the 95% is accurate, the damage that losing 20% of your roll on one play, even if only 5% of the time, would not be handled well by most.


Also, to say that you are losing out on value just because you are a risk better is making an assumption.

Just say you play all of your plays that have, in your mind, 1%-5% value TO WIN one 1% or one unit. Another player RISKS 2% of his bankroll on the same set of plays. As long as the majority of these plays are -199 and under, which is likely, the RISK better is maximizing value, right?

[richyrich8478]

i personally take a lot of risks...thats part of gamling and thats what make it so much fun...also a reason why i win so much money

[sharpcat]

Greek Sports Picks is now 28-5 in their "MLB Lock of the Day." Today, their pick was the Over of 9 1/2 in the Rockies and Brewers game. It easily covered with 15 runs scored combined between the 2 teams. Try their free pick for tomorrow by dialing their toll free number: 1 - 888 - 608 - 7487. Their free picks record is 14-5. TRY IT NOW! BREAK THE BOOKIES!!! ITS EASY MONEY!

Handicappers digest says these guys are on a 1-20 run