Tuesday, May 03, 2022

Decimal odds and heavy favorites

Continuing this series on sports betting and decimal odds. The topic of this post is when one team is a big favorite to win a game. Is it possible to take enough small wins and avoid the big losses to grind out a profit on big favorites.

Betting on heavy favorites, the better team in a game that looks like a mismatch. It's a bit like picking up pennies in front of a steamroller. You can win a small amount or lose a large amount. Seems like easy money until it isn't.

A game that looks like a mismatch might not be a lock though. some examples
Undertaker vs. Brock Lesnar, WrestleMania XXX, 2014
Mike Tyson vs. Buster Douglas, 1990
UNLV vs. Duke, NCAA Basketball Final Four, 1991
wish I had the link. A guy lost $17,000 on women's tennis laying 170:1 when an unranked beat a top player. The big name player had to pull out of the match with injury.

For this post, I will define heavy favorite as having decimal odds in the 1.05 to 1.10 range. On ALC anyway 1.05 is the floor payout on favorites. A bit more on that below. From my earlier post, I calculated at 1.05 the break even point is 95.24%  At 1.10 the break even is 90.91%

So even though these bets usually win, the player is laying heavy odds. A loss is unusually bad as each loss will take numerous consecutive wins just to cover it and get back to even. Still I wonder if there may be opportunity to find long-term profitable plays among heavy favorites. To grind a slow, steady profit and minimize disastrous upsets.

quantifying the strength of a favorite

One thing I noticed on ALC Pro-Line is that for a given favorite odds such as say 1.06, on different games the odds offered on the underdog can vary. From my earlier post I noted that house underdog prices serve two purposes. To reflect the probability that the underdog will pull an upset; and to draw bets on the underdog, to balance against heavy action on the favorite.

some recent examples from Pro-Line

NBA playoff series

Atlanta 7.25
Miami 1.05

Toronto 6.75
Philadelphia 1.07

Boxing

Dillian Whyte 4.70
Tyson Fury 1.10

So if we take a bit of a leap. use the underdog decimal odds as a kind of proxy, an estimate that the underdog will win. The lower the underdog probability, the safer the play on the favorite becomes. so a higher underdog number means a lower probability that the underdog will win. This can provide an extra piece of data when evaluating betting on a favorite. At different prices, 1.05, 1.06, 1.07, 1.10, with some data research there may be a floor underdog price that can make that play on the favorite good. when combined with other research on the game.

Now comparing various favorite options. As the price on the favorite increases from 1.05, the underdog price will go down. Of course as favorite prices increase, the winning payout increases, which can allows more risk or tolerance for upsets. For example 1.07 pays 40% higher than 1.05. 1.10 pays out double on 1.05.

I got thinking about how to compare or somehow quantify the favorite and underdog prices. I came up with a "strength of favorite" number. with favorites, we want to optimize the small winning payout against how often we have to eat a big loss.

Let Pf be the decimal odds on the favorite.
Let Pd be the decimal odds on the underdog.

then the strength of favorite number is defined as
(Pf - 1) * (Pd - 1) * 100

the examples above come out as

Game Favorite Decimal Odds (Pf)
Underdog Decimal Odds (Pd) Strength of favorite number
Miami vs Atlanta NBA Playoffs 1.05 7.25 31.25
76ers vs Toronto NBA Playoffs 1.07 6.75 40.25
Tyson Fury vs Dillian Whyte Boxing 1.10 4.70 37

 from there maybe 76ers at 1.07 might be the most promising of these three plays, as a starting point in research, in addition to other analysis on this play.

about 1.05

I noticed that 1.05 is the minimum favorite line that ALC will run. It is likely that at least some of these 1.05 are actually 1.02 or 1.03 "rounded up" to 1.05. So probably some of the 1.05 are mispriced for that reason. So if you can figure out where to look there may be an opportunity there. Still with 1.05, the break even point is 95.24%  which doesn't leave much room to work with to grind out a small profit. In some gross mismatches which can occur in international women's hockey and men's international soccer, ALC won't run a gift 1.05 moneyline. You have to lay a large points spread to play on the favorite.

in leagues that have achieved a lot of parity such as the NFL, NHL, and major league baseball, you don't see 1.05 lines. on any given Sunday is real and in general no single game is such a mismatch that the underdog has "no hope" and it's a minimum 1.05 play on the favorite.

----

Update: on the featured matches above

Miami won the playoff series over Atlanta 4-1

Philadelphia won the playoff series over Toronto 4-2

Fury wins over Whyte in a six round TKO

 

so this time all of the favorites held. those who played on the favorites came out ahead

Monday, November 29, 2021

How the house sets decimal odds

Subtitle: two ways to win at sports betting

In my previous post, I looked at decimal odds from the players perspective. In this post I will consider decimal odds, or the price in general, from the sportsbook perspective.

I'll use a running example for this post. Suppose for some upcoming football game the analysis desk at the sportsbook determines that the favorite for this game has an 80% chance to win the game. The underdog has a 20% chance to win.

How would the house post a line for players to bet against. From my previous post, from the players perspective

x = 1 / D

where x is the players break-even probability, and D is the decimal odds.

Now for the house it then is

D = 1 / x

The house is holding the probabilities x, and can use these to calculate its break-even odds D. So for this game the favorite is 4/5, and the underdog is 1/5. So the house is indifferent with the lines

favorite decimal odds line: 1 / (4/5) = 1.25

underdog decimal odds line: 1 / (1/5) = 5.0

If the analysis desk is right about the probability, then this line would be unexploitable to the house. The house would expect to break even.

Now the house wants to make a profit, not just break even. So the book could take the mathematical break even line above, and shave a bit off it to create some profit margin - a some margin of error in their analysis of the game.

So if the house takes a 10% discount off the mathematical line, then the players could see a posted line of

favorite line: 1.13

underdog line: 4.5

--

Now I promised to show two ways to win at sports betting. The first way to win is when your analysis of the game is better than the sportsbook analysis. Suppose you've studied this football game and you feel the favorite is stronger, 90% to win the game. 

At 1.13, your break even point x is 88.4% if your analysis is better than the house analysis then you can make an expected profit betting on the favorite.

--

When the house posts lines, one of their goals is to balance the action. When the action is properly balanced, the house will make a profit on the game no matter which side actually ends up winning.The house has none of its own money at risk. Winning wagers are paid out from losing wagers, and there is some left over for an operating margin for the sportsbook.

I imagine that when the profit is assured, the sportsbook can pour a cold beer, sit back and just enjoy watching the game along with all of the players.

Consider again the football game above. Suppose the betting action comes in at the same ratio as the mathematical line. 80% on the favorite, 20% on the underdog. Suppose 5 people each make a $10 bet on the game. So 4 bets on the favorite at 1.13 and 1 on the underdog at 4.5. The house takes in $50 from the five players.

If the favorite wins then the house pays out 4 $11.30 tickets, so it pays out $45.20

If the underdog wins then the house pays out 1 $45 ticket, so it pays out $45

So the action on this game is properly balanced. The house will make a profit no matter what happens in the game. All winning bets are paid out with the money from the losing bets. None of the sportsbook money is at risk in this game.

--

Trying to balance the action leads to the second way to win at sports betting. Suppose the risk management desk looks at the discounted mathematical line from above of 1.13 / 4.5  risk management is concerned there will be too much action on the favorite, putting house money at risk if the favorite wins. Cutting the line on the favorite to 1.10 might help some. But to get more action on the underdog to balance the action, the house moves the line to 5.5 on the underdog.

Now the break even line on the underdog is 5.0  at 5.5 it is profitable to take the underdog - even though you will lose the bet 80% of the time. So sometimes the house may post lines to balance the action that are profitable opportunities.

Friday, October 01, 2021

Calculating the break even point of decimal odds

 On ALC at least, the odds on the lines are expressed using decimal odds. So you will see things like 1.05, 1.90, 4.35, etc. This post is just to do a quick primer on decimal odds, and to demonstrate the break even point for a given decimal odds wager.

I find an easy way to think about decimal odds is this. Suppose the decimal odds for some event is D. You give the house $1 when you make your bet. If you win the bet the house pays you D. If you lose the bet then the house just keeps your dollar.

From there it should be clear that the decimal odds D will always be greater than one. After all nobody is going to give a bookie $1 in order to "win back" 85 cents, and that's only if you win the bet.

Some examples

if the decimal odds are 1.1, then you wager $100 to try to win $110

if the decimal odds are 1.5, then you wager $10 to hope to win $15, for a $5 profit

if the decimal odds are 2.5, then you wager $1 to profit $1.50

so thinking about it a bit, decimal odds are what the player is laying on a wager. so at 1.5, you lay $10 to make $5 profit. so laying 2:1 you need to win this bet 2/3 of the time to break even.

at D=1.1 you are laying 10:1, so basically you are laying 1 : ( 1 - D ) with decimal odds.

Or to put it into a formula. let D be the decimal odds. let x be how often you win the wager

then your break even point is

( D - 1 )x - ( 1 - x ) = 0

given D, solving for x yields

x = 1 / D

here's table based on some common decimal odds posted at ALC Pro-Line

Decimal Odds Break Even Point (Percent)
1100
1.05 95.24
1.1 90.91
1.25 80
1.5 66.67
1.75 57.14
1.85 54.05
1.89 52.91
1.9 52.63
1.91 52.36
1.92 52.08
1.95 51.28
2 50
3 33.33
4 25
5 20
6 16.67
10 10
12 8.33
13 7.69

Friday, June 04, 2021

Accounting for the value of lotto free tickets

At this time the Lotto Max jackpot is at the max $70 million. It was at 70 for the Tuesday draw earlier this week but it didn't go.

I had a ticket in the Tuesday draw. I had some free play money on ALC from a $50 for $25 deposit reload bonus. I bought a $5 ticket.

In the Tuesday draw, as sometimes happens, I won a free ticket into tonight's draw. That's okay with me, a second shot at the big jackpot, or one of 29 $1 million secondary maxmillions prizes, off one ticket purchase.

It did get me thinking, what's the value of winning a free ticket? Either for the player, or the house. (of course mathematically they must be the same) The free lotto ticket could be worth something, or improbably quite a lot. it could also roll over again into another free ticket. mostly of course they turn out to be worthless since most lotto tickets just lose.

From the house perspective, how to calculate the profit or loss on each individual draw when free tickets into future draws are awarded. How would especially winnings off free tickets that the player was given be accounted for.

I realized if you set it up and think about it properly it's not that hard. The way to think about and calculate the value of winning a free lotto ticket is like this. 

In the first draw, your ticket actually won a $5 prize which was paid out. The $5 win was then "intercepted" by ALC, and the house used the $5 to purchase a $5 ticket into the next draw. ALC then gave the ticket it just bought to the player who becomes the legal owner of the lotto ticket.

To the house, everything is clean. On the first draw your ticket won $5, which is accounted as a $5 winning payout in the profit and loss on that draw. ALC can close the books on the first draw and know exactly how much came in from ticket sales and how much was paid out to winning tickets. The accountant for the first draw does not care or have any interest in how the free tickets fared in the later draws.

On the second draw the same $5 is accounted as income as a regular $5 ticket purchase, the same as buying a ticket at a corner store or on alc.ca  To the accountant on the second draw, they received a $5 payment and issued a ticket. Everything is in order and they don't care if the ticket was then gifted to some customer as a free ticket.

To the player, it's a ticket into the second draw which would have cost $5 to buy at a gas station or online. So basically you won $5 on the first draw, but the house on your behalf used your $5 win to buy you a ticket into the next draw.

Saturday, February 27, 2021

ALC Pro-Line payouts have improved

In recent advertising, ALC Pro-Line has been emphasizing a more studious approach. The commercials talk about analyzing odds, studying matchups, etc. When you're ready to make your picks, then Pro-Line payouts are in line with anything out there.

This is a change in the tone of the marketing. Historically the Pro-Line advertising has featured this goofy, oafish guy and his friends. They go to corner stores and make their picks. The goofy guy is clumsy and always knocking over a display or something. He just makes silly picks and it's all good fun among friends.

In the new advertising it's a departure for Atlantic Lotto, normally a monopoly, to even acknowledge there are other options for sports betting. On the same mainstream channels and TV shows Pro-Line buys advertising, there are also commercials for William Hill, 888, Bodog, Party, Odds Shark, various online casinos. 

By appealing to the more serious player, ALC seems to realize they have lost online sports business to offshore books. In the past I calculated that Pro-Line was paying about 77 cents per dollar wagered, keeping a 23% vig for itself. This is excessive and unfair. Even recreational players who aren't good with math, who still like to play a few games every week, could instinctively tell something isn't right with the payouts.

I logged in recently due to an ALC promo. I have an ALC.ca account from promos in the past. Over on Pro-Line I noticed that payouts have indeed improved. On a two-game parley with even payouts - the house pays out the same on both sides of the bet on a totals or point spread wager. In these games, on a $10 wager the house pays out $36.10 to a winner.

So if you bought 4 $10 tickets to cover the four possible combinations. You spend $40. Three tickets lose, and one will win $36.10. So for $40 spent, $36.10 is paid out to winners, or a $90.25 payout for every $100 wagered. The house keeps $9.75 of every $100, a 9.75% vig.

This seems reasonable and is in line with standard 11/10 vig at sports books. So +1 to ALC for improving their payouts and truth in advertising.

is Pro-Line beatable

At 9.75% vig it may creates some opportunities for playing for fun and profit. The break even point can be calculated on a two-game parley at .9025. So on a $10 wager you can win $26.10 or lose $10. Let X be the fraction of the time you win your bet.

26.1x - 10(1 - x) = 0

x = 0.277

So the break even is about 28%. If you use a random number generator to make your picks you would win by pure luck 25%. Remember you need to win two bets to win the ticket.

So if from your analysis you have two good picks then the square root of 0.277 is .5263. So if you can run 53% and find two games priced at 1.9 (the .9025 payout) then you would be profitable.

If you only have one good pick there is still an opportunity. On a game where the house pays out the same on both sides you can use the random number generator to make the pick and you would win the bet half of the time. So you can run 50% on events you know nothing about such as Mexican soccer, Euroleague basketball, or whatever. As long as the payout is the same on both sides of the bet.

Or buy two tickets, with your good pick, and both sides of the even payout game. 

With one good pick, and one random pick at 50%, you need to be 55.4% on your good pick to be profitable.


There is also a limited form of single game betting now on ALC Pro-Line. For hockey it's win the game and score in regulation. These games are attractively priced. On evenly priced games Pro-Line pays out $19 for every $20 wagered. So only a 5% vig on these moneyline plays with single-game allowed.

Now there is a slight catch where the payout is not quite 95 cents on the dollar. If the game is scoreless 0-0 at the end of regulation, then the house pays out nothing and scoops all the wagers on the game. It's a bit like the 00 in roulette, but less common. So be aware of that small risk. Still 0-0 games are uncommon in the NHL, and scoring is up this year. I haven't done the research but I'm guessing it reduces the payout to maybe 94 cents on the dollar.

Still good to know moneyline NHL at low vig. It's a good format for the players and I hope it sticks. 

These games are handy for clearing play though bonuses as they pay 0.95 instead of 0.9025 on the two-game ticket. Also you only need to win one bet to win the ticket. 

 

If you're a more casual fan and not sharp action, then a trick with ALC Pro-Line is use bonuses to improve your odds. Pro-Line and ALC run bonus promotions fairly frequently so it's good to sign up for the emails so you know when something's running. 

I wonder if this will be looked back on as the good time on Pro-Line. Payouts are competitive, it is easy to get money on or off, and there are frequent promos and bonuses.

Tuesday, October 17, 2017

Free Play on ALC Pro-Line

It's been a while. I haven't been active in any area recently. Poker or sports. It's not looking great that I will be doing much in either area in the foreseeable future. It is what it is I guess. I may have something somewhat later to post on but will see.

Anyway after a long dormant period there has been a noticeable increase of late in emails from Atlantic Lotto. I'd guess there is some new manager there and is driving this email outreach.

Usually this is just noise that is easily ignored. However there was an interesting Pro-Line promotion. If you make a $10 bet, then you get a $10 credit from the house for another bet.

Well I couldn't pass up that opportunity. The question becomes how to structure my plays. The way Pro-Line works is you have to parlay with at least 2 plays, betting on individual games is not available.

I unfortunately don't have time for a lot of research so I was thinking of a way to realize the EV of the free play using random picks. I found some "even" NHL totals games to pick. By "even", I mean the price from the house is the same on the over as well as the under. So if the house is indifferent to what you pick, then that's even. With the same price on the over as the under, it's well suited to using a randomized approach to picking, as there's no difference in expected outcome either way.

Although I couldn't help but do a quick cursory research on the games I picked. I ended up taking the OVER at 5.5 for Colorado at Nashville, and the UNDER at 5.5 on Carolina at Edmonton. I didn't do enough research to explain my reasoning here.

I can calculate the EV of the free play. With no research and a random selection there's a 50% chance of winning on each game. I have to win both games to win the bet. So there's a 25% chance I will win the bet using random selection for the games.

Now if I win the bet (25%) I win $20.60 is the payout on these two games. And if I lose (75%) I lose $10. So the EV is -$2.35. So for a $10 wager where the house is indifferent the expected end result in my account is $7.65. So 23.5% vig, thanks ALC. it's horrible they exploit the players like that.

Anyway on the $10 free play it would also have an EV of about $7.65. So the expected end amount in my account is $15.30, so a healthy 53% profit, and it was fun to be back in the game.