GAM 470
Probabilities in Derby and Sports Futures
Derby is a horse racing game found near many sport books. The game has five horses are the player bets on one of the combin(5,2)=10 possible quinellas. A quinella is a bet on the first and second place horses, in no particular order. The wins are different every time but following is an example.
|
Quinella |
Pays (for one) |
|
1, 2 |
5 |
|
1, 3 |
39 |
|
1, 4 |
200 |
|
1, 5 |
19 |
|
2, 3 |
4 |
|
2, 4 |
20 |
|
2, 5 |
2 |
|
3, 4 |
160 |
|
3, 5 |
15 |
|
4, 5 |
79 |
To determine the expected return of Derby first assume that it is the same on all bets. Then imagine placing a bet of 1/x on each quinella, where x is the payoff odds on that horse, on a "for one" basis. In this case regardless of which horse won the player would get back 1 unit. The amount spent on wagers, in the above example, would be 1/5 + 1/39 + 1/200 + … + 1/79 =
1.168847. So as in any bet the expected return is the ratio of the amount the player can expect to get back to the amount bet. In this case 1/1.168847 = 0.855544 = 85.54%. The house edge is thus 1 - 0.855544 = 0.144456 = 14.46%.The next table shows all the details for this particular race.
|
Quinella |
Pays (for one) |
Inverse of Win |
Fair Probability |
Fair Odds |
|
1, 2 |
5 |
0.200000 |
0.171109 |
5.844237 |
|
1, 3 |
39 |
0.025641 |
0.021937 |
45.585052 |
|
1, 4 |
200 |
0.005000 |
0.004278 |
233.769500 |
|
1, 5 |
19 |
0.052632 |
0.045029 |
22.208102 |
|
2, 3 |
4 |
0.250000 |
0.213886 |
4.675390 |
|
2, 4 |
20 |
0.050000 |
0.042777 |
23.376950 |
|
2, 5 |
2 |
0.500000 |
0.427772 |
2.337695 |
|
3, 4 |
160 |
0.006250 |
0.005347 |
187.015600 |
|
3, 5 |
15 |
0.066667 |
0.057036 |
17.532712 |
|
4, 5 |
79 |
0.012658 |
0.010830 |
92.338952 |
|
Total |
1.168847 |
1.000000 |
Inverse of win: How much the player must bet for a return of 1.
Fair probability: The actual probability of winning.
Fair odds: What the bet should bet to have a 100% return.
The expected return is the inverse of the total "inverse of win" column. In this case 1/1.168847 = 85.54%.
Next let’s look at a futures bet. Unlike Derby sports futures pay on a "to one" basis. I think it is easier to understand them if you convert the odds to a "for one" basis, by adding 1 to each win. Then take the inverse of each win. The sum of these inverses is how much you have to bet to be guaranteed a return of 1. So in the example below the expected return is 1/1.3336 = 74.99%, i.e. the house edge is 25.01%. The "fair probability" column is the actual probability of winning, assuming each bet had the same house advantage. Both "fair odds" columns are what the bet should be to have a 100% expected return.
2005 National League Championship
|
Team |
Wins (to one) |
Wins (for one) |
Inverse of win |
Fair probability |
Fair odds (for one) |
Fair odds (to one) |
|
Arizona Diamondbacks |
25 |
26 |
0.0385 |
0.0288 |
34.6736 |
33.6736 |
|
Atlanta Braves |
6 |
7 |
0.1429 |
0.1071 |
9.3352 |
8.3352 |
|
Chicago Cubs |
4 |
5 |
0.2000 |
0.1500 |
6.6680 |
5.6680 |
|
Cincinnati Reds |
30 |
31 |
0.0323 |
0.0242 |
41.3416 |
40.3416 |
|
Colorado Rockies |
250 |
251 |
0.0040 |
0.0030 |
334.7333 |
333.7333 |
|
Florida Marlins |
6.5 |
7.5 |
0.1333 |
0.1000 |
10.0020 |
9.0020 |
|
Houston Astros |
13 |
14 |
0.0714 |
0.0536 |
18.6704 |
17.6704 |
|
Los Angeles Dodgers |
15 |
16 |
0.0625 |
0.0469 |
21.3376 |
20.3376 |
|
Milwaukee Brewers |
150 |
151 |
0.0066 |
0.0050 |
201.3734 |
200.3734 |
|
New York Mets |
5 |
6 |
0.1667 |
0.1250 |
8.0016 |
7.0016 |
|
Philadelphia Phillies |
13 |
14 |
0.0714 |
0.0536 |
18.6704 |
17.6704 |
|
Pittsburgh Pirates |
90 |
91 |
0.0110 |
0.0082 |
121.3575 |
120.3575 |
|
San Diego Padres |
12 |
13 |
0.0769 |
0.0577 |
17.3368 |
16.3368 |
|
San Francisco Giants |
6 |
7 |
0.1429 |
0.1071 |
9.3352 |
8.3352 |
|
St Louis Cardinals |
5 |
6 |
0.1667 |
0.1250 |
8.0016 |
7.0016 |
|
Washington Nationals |
150 |
151 |
0.0066 |
0.0050 |
201.3734 |
200.3734 |
|
Total |
1.3336 |
1.0000 |
Source: Bodog