Sportpunter NBA. Odds Analysis Part 3
Previously we looked at how our Sportpunter NBA model has gone betting throughout the season (http://www.sportpunter.com/2017/10/sportpunter-nba-odds-analysis-part-1/) as well as how it had gone when compared to finishing prices (http://www.sportpunter.com/2017/10/sportpunter-nba-odds-analysis-part-2/). But today we will look at the expected profits based on those finishing prices and how we would have gone betting throughout the season.
The Tables shown below will look very similar to the tables on the first article with two columns added. The first table is based on head to head betting, and the two extra columns are %ProfitFinPin and $DiffProfit. What this indicates is that if we bet the same amount as we did when we suggested the bet, but not at the suggested bet price, but rather the finishing Pinnacle price then how would we have gone. The last column, $DiffProfit looks at the difference between the actual profit, and the profit based on finishing Pinnacle Prices.
Sound complex? Maybe, But what this means is that if $DiffProfit is positive, then our profits are greater than what we would expected, in other words, they are good bets. If $DiffProfit is negative, then perhaps the profits are not great as on average the odds have moved against us. A positive $DiffProfit means that the odds have moved in our favour when compared to finishing price, and if it is negative the odds have moved against us.
| MinSeasonMatch | #Bets | $Bet | $Profit | %ROI | %ProfitFinPin | $DiffProfit | |
|---|---|---|---|---|---|---|---|
| 1 | 10 | 139 | $32,492.94 | $9,965.88 | 30.7% | $10,364.19 | -$398.31 |
| 10 | 20 | 535 | $100,861.12 | -$1,220.90 | -1.2% | -$1,144.30 | -$76.59 |
| 20 | 30 | 484 | $83,727.42 | $5,968.13 | 7.1% | $3,877.33 | $2,090.80 |
| 30 | 40 | 455 | $70,763.48 | -$4,567.31 | -6.5% | -$5,620.51 | $1,053.20 |
| 40 | 50 | 446 | $67,946.01 | -$2,323.13 | -3.4% | -$2,234.44 | -$88.70 |
| 50 | 60 | 465 | $69,879.15 | $6,693.77 | 9.6% | $5,951.19 | $742.58 |
| 60 | 70 | 439 | $62,298.43 | $7,574.05 | 12.2% | $7,209.49 | $364.56 |
| 70 | 80 | 439 | $59,678.20 | -$968.59 | -1.6% | -$1,507.79 | $539.20 |
| 80 | 90 | 466 | $80,206.26 | $2,826.39 | 3.5% | $1,336.25 | $1,490.14 |
| 90 | 170 | 269 | $42,808.97 | -$10,232.47 | -23.9% | -$9,720.84 | -$511.62 |
The table above looks at head to head betting and the minimum number of matches for each team. The $DiffProfit column indicates a small negative amount early, followed by large positive amounts and then a negative amount ($-511) at the end of the season. Firstly the sum of the $DiffProfit amount is positive, which is great and indicates a good healthy and profitable model . It is slightly negative at the start which could indicate that perhaps we have been slightly lucky early on in the season or that we are pushing a position with the model which is not favoured that much around smart money. Considering the amount of profit that has been made in the first months, I believe this is purely due to random variation and a little bit of luck in our favour.
The model is strong throughout the year but negative when it comes to playoffs, once again indicating that stopping betting at playoffs is preferred for head to head betting.
For line betting as shown by the table below it is a little different
| MinSeasonMatch | #Bets | $Bet | $Profit | %ROI | Av. Line Change | %ProfitFinPin | $DiffProfit | |
|---|---|---|---|---|---|---|---|---|
| 1 | 10 | 159 | $38,455.33 | $5,170.87 | 13.4% | 0.142 | $4,508.60 | $662.26 |
| 10 | 20 | 575 | $123,332.86 | $6,260.14 | 5.1% | 0.127 | $4,586.45 | $1,673.69 |
| 20 | 30 | 505 | $95,984.23 | $6,319.83 | 6.6% | 0.095 | $5,301.23 | $1,018.60 |
| 30 | 40 | 489 | $85,209.79 | -$1,133.59 | -1.3% | 0.082 | -$786.81 | -$346.77 |
| 40 | 50 | 496 | $86,026.81 | -$2,239.35 | -2.6% | 0.052 | -$3,389.21 | $1,149.85 |
| 50 | 60 | 471 | $88,711.70 | $3,716.12 | 4.2% | 0.128 | $5,120.80 | -$1,404.68 |
| 60 | 70 | 461 | $77,347.09 | $1,032.35 | 1.3% | 0.072 | -$1,201.93 | $2,234.28 |
| 70 | 80 | 453 | $81,855.98 | $254.56 | 0.3% | -0.029 | $260.88 | -$6.31 |
| 80 | 90 | 469 | $104,366.07 | -$4,165.02 | -4.0% | 0.057 | -$3,290.51 | -$874.52 |
| 90 | 170 | 264 | $51,980.24 | -$8,749.92 | -16.8% | -0.031 | -$8,390.16 | -$359.76 |
We have relatively positive amounts in the $DiffProfit column up to around minimum match 70. When the minimum match of each team is 70-80 then expected profits is basically zero, as was the actual profits. Once again this confirms that betting the line up to around a month out of playoffs might be the best way to go.
For totals betting as shown below, totals profits as shown in $DiffProfit are strongly positive throughout
| MinSeasonMatch | #Bets | $Bet | $Profit | %ROI | Av.%Change | %ProfitFinPin | $DiffProfit | |
|---|---|---|---|---|---|---|---|---|
| 1 | 10 | 148 | $36,282.58 | $5,195.24 | 14.3% | 0.370 | $3,989.31 | $1,205.93 |
| 10 | 20 | 561 | $120,678.39 | $4,220.35 | 3.5% | 0.445 | -$1,041.29 | $5,261.64 |
| 20 | 30 | 466 | $84,370.44 | $2,852.37 | 3.4% | 0.643 | -$617.48 | $3,469.85 |
| 30 | 40 | 462 | $78,393.49 | $4,409.70 | 5.6% | 0.430 | $568.47 | $3,841.23 |
| 40 | 50 | 440 | $67,681.26 | -$5,266.76 | -7.8% | 0.314 | -$9,201.31 | $3,934.55 |
| 50 | 60 | 442 | $69,122.96 | -$2,124.86 | -3.1% | 0.437 | -$6,021.21 | $3,896.35 |
| 60 | 70 | 471 | $77,317.46 | $1,255.55 | 1.6% | 0.340 | $1,230.25 | $25.30 |
| 70 | 80 | 417 | $64,610.29 | $3,659.75 | 5.7% | 0.418 | $1,392.24 | $2,267.51 |
| 80 | 90 | 394 | $63,503.74 | $2,995.83 | 4.7% | 0.385 | -$590.35 | $3,586.18 |
| 90 | 170 | 262 | $49,956.23 | $4,960.97 | 9.9% | 0.198 | $4,938.81 | $22.16 |
Note that for playoffs, the variable $DiffProfit is only slightly positive, indicating that at this stage the lines are not moving that much in our favour despite our bets and staying pretty constant. But it does show that on average throughout the season, the line on average for totals moves in our favour greatly.
So we have shown not only how profitable the model is, and how when the odds move in our favour is it even more profitable, but we have also shown when the odds move in our favour as well.
In the final of the four part article series, we will look at betting early, compared to betting late.
To sign up to the Sportpunter 2017/18 NBA model, click here for details. http://www.sportpunter.com/sports/basketball/nba/
