Lay It On The Line!
After the first 10 rounds we looked at the returns if you had bet on your team each week. Unsurprisingly, the Knights fans were found to be out of pocket the most, and still would be, after recording just the one win.
But what about if you bet the line each week?
Well firstly, let’s define what the line is. It is when a match is handicapped by the bookmaker to make it a fair fight. It means, if the bookmaker is correct, each team has a 50% chance of winning taking into account the handicap.
So looking at the Knights for example. They have only won one game and are a near certainty to get beaten each week. So to make it a fair fight they typically will receive a nice handicap to their score. In the last round this was a massive 24.5 points. This means that they get a 24.5 point head start, and as long as they lose (or win) by less than this, so by 24 points or less, you win the line bet. They actually nearly won (amazingly) last week, and as they lost by only 4 points, you would have won the line bet this week if you’d have gone with the Knights.
What would we expect?
If the bookmakers model (Sportsbet in this case) is correct, and the handicaps are accurate, you would expect a team to win 50% of the line bets for you, and conversely lose 50% too. So after 16 games (4 teams have only played 15) you would expect every team to have 8 line wins and 8 line loses.
So we have looked at this and the following is the state of play after 17 rounds, assuming you bet on each team each week (clearly betting on both teams in same game would be stupid but for this analysis we have):
What this tells us, is that for teams with a win percentage above 50%, they have done better than the bookmakers thought – ie the bookmakers have underestimated these teams. So Canberra (63%), Cronulla (67%) and Gold Coast (63%) have been the stand out teams – the bookies keep underestimating how good they are. All three teams have positive returns on our H2H betting, which is evidence our model has recognised this!
For teams below 50%, it tells us the bookmakers give the teams more credit than they deserve – ie the bookmakers have overestimated these teams. Manly (38%), Knights (38%) and South Sydney (38%) are all in this category. Manly and South Sydney are often over valued by our model too when looking at the H2H bets (apart from last bloody week!!!).
The bookmaker’s model has been spot on for Warriors, St George and Wests.
Overall, betting on each line would obviously result in a loss. The bookies only pay out 1.90 so the cumulative return would be negative $12.60 on 252 bets, or a loss of 5%.
For the mathematicians out there, in a fair (50:50) game, the probability of getting 6 out 16 (ie more Manly) is 23%, so shows you how far wrong the bookies model is.
But what if you had a model that was better at predicting the line?
What about the TipBetPro Model?
We have therefore looked at the Sportsbet line (so handicap), looked at what the TipBetPro predicts would be the result, and then assumed we bet on the team who we think would win taking into account the handicap.
So this would mean laying 126 bets to date (ie one per match). As shown below we would have won 74 of these bets (59%) and the returns would be 14.60, or 12%. Not too bad.
In terms of teams, you can see our model is really bad at predicting Manly. This is because we rated them relatively highly when they were losing and when the model corrected they win!! The model suggests never taking the line bet on the Knights (which appears reasonable) or Penrith (which means we are likely undervaluing them which is an experience issue, as our model backs experience and they are relatively low on this).
If we look at returns by round, it shows that the past few weeks (when our H2H has been poor) our line predictions have been getting better and better (which also ties in with our tipping being good).
Being selective
Can we increase the winnings by only betting on certain lines? Well you’d think if you bet on games when the difference between the prediction and the line is greatest, you’d increase the likelihood of a win.
So looking at games where the difference is over 10 points, we’d have bet on 45 games, winning 29 of these, so 64% of games. The return would be 22%, so close to our year to date H2H return of 26%.
Interestingly, the St George Dragons do stand out as a team we have predicted would win by 10 or more points against the line (so go against the bookies quite strongly) and they haven’t achieved the bookies line 4 out of 7 times. That said, they have done us proud at times on the H2H bets, but all up does indicate we favour them more than we should!
In the off season we will do more analysis, tweak some parameters and add some statistical rigor, with a view to offering line bet recommendations next season (or handicap betting).
Hope this was informative and any thoughts/suggestions much appreciated.
To see our betting recommendations, including our Bet of The Week, and to read more about our approach, and to compare our performance with your own, please visit www.tipbetpro.com or follow us (abuse us) on Twitter @tipbetpronrl.