Sunday arvo and this week was looking like a losing one, but Tennis and Formbet came to the rescue over the last few nights to put the week well into profit. AFL was so frustrating again this week, so many good starts and close games.
Brett left a very good comment on last weeks post. He said
Mate, pay a coder, or code this yourself. You probably would have saved yourself a lot of money earlier on if you had done this:
Record the line and/or price (for straight up betting) you took a bet at, and compare it to the closing line and/or price. Record the proportion of the time the line moved for/against you, and the magnitude of that move. The ‘Beat the Closing Line’ evaluation method (BTCL).
Do this for every selection service. Figure out the standard deviation of movement for the closing line in every sport/market you’re into. Calculate the statistical probability of selection service X in having a long term edge against the closing line (calculate the z score).
Bet peanuts / shadow bet (just to generate records to analyze) until the selection system goes ‘live’ once you have a z score worth considering (>2).
This way, you wouldn’t have to subjectively ‘reassess’ anything (as I don’t really know what other hard statistical data you are going to use to make decisions about keeping or dropping a service.)
Now, I did do statistics at University, but it was only a first year subject and I had Jonathan Lowe from Sportpunter (he used to teach statistics at Uni) do pretty much all my homework. So I don’t know anything about Z scores and whatsits. But the basics I can do and should do as Brett is right. A simple ROI is not going to help me decide where to continue betting and when to stop a service. This following week I will be tracking all closing lines for baseball, Tennis and AFL. I have been keeping an eye on them this week and it does look like I am getting much better odds then the closing lines, which is a great thing.
MLB recorded another losing week with around $3,000 lost. AFL after what seemed like a terrible week only resulted in a $1,200 loss. NRL and Golf lost $600 and $700 respectively.
Tennis had another great week with $2,000 profit in ATP and an even more impressive $3,800 in the WTA. This is the first time in 11 months both models have been in profit. To think that at one stage early in the year the ATP model was at -$15,000. An amazing $27,000 turnaround in about 4 months. Coming into the short grass season, and traditionally the model has not performed to well on this surface. I won’t be changing my stakes or betting any different.
A big well done and thank you to Dave at Formbet. He has put in a massive amount of work into the new autobot and if this first weeks worth of results is any indication, then every minute was worth it. A $8,800 profit for the week puts me back into profit with this model. I am not going to get too excited just yet as I have had weeks like this before and then the next week turns it all around, but things are looking quite positive.
Now that i’ve started something and it seems to be working, i’ll have to continue it. HERE’S HOPING FOR THE WORST BETTING WEEK I HAVE EVER SEEN
Total profit for the week was $9,177.64
Baseball $313,775.01 $14,445.58 4.6%
Week profit/loss: -1193.46
Baseball Line $379,673.00 $27,796.49 7.3%
Week profit/loss: -1715.07
AFL $53,223.00 $(1,803.50) -3.4%
Week profit/loss: 848.19
AFL line $113,456.00 $6,741.90 5.9%
Week profit/loss: -351.68
AFL Totals $77,466.00 $(3,651.00) -4.7%
Week profit/loss: -1729
Tennis – ATP $304,438.00 $12,558.07 4.1%
Week profit/loss: 2062.47
WTA – Tennis $295,160.00 $65.21 0.0%
Week profit/loss: 3785.89
NRL $23,350.00 $(6,826.39) -29.2%
Week profit/loss: -613.70
Golf $16,477.00 $(8,635.25) -52.4%
Week profit/loss: -700
Formbet $84,055.00 $2,695.00 3.2%
Week profit/loss: 8784
Stats
Total bet amount: $2,808,891.37
Profit/Loss: $26,825.71
POT: 1.0%
Bets Placed: 7,013
Won: 2,824
Lost: 3,936
Half Payout: 4
Refunded: 249
Winning %: 41.8%
Profit per day: 26,825.71 / 342 = $78.44