sw

[danniego]

Just bought the program and fiddling around.
In the Stakewizard i put in the selections i want to bet on including the line odds.
Bank is: 2000 and end target is 20000
I get bet recommendations in the range of 10-20% of the bank. Is this not to steep?. I thought around 1-5% of your bank total on any bet is golden rule.
And with 10 bets is the list i have to bet almost the total bank on the events.
Also the complex betting is not clear to me what it is exactly and how to do this type of betting then as advised by sw.
i.e. It combines C with E for amount of 18. How to place this bet then?

[AimHigher]

These are my 2 cents, based on my knowledge of sports betting, statistics, money management theory and my experience as a user of Odds Wizard/Stake Wizard. How much you should be willing to risk on a bet, or a series of bets, is a function of the expected value AND the probability of winning. As an example, let’s say we have two bets:
Bet A: Fair odds 1.5 Line odds 1.8
Bet B: Fair odds 4 Line odds 4.8
Both bets have the same expected value of 20% (1.8/1.5 - 1 = 4.8/4 - 1). However, the probability of winning is very different. Bet A has a probability of winning of 1/1.5 = 66.67% while Bet B only has a probability of winning of 1/4 = 25%. Hence, the variability of your bank roll if you play a series of Bet B will be a significantly greater than for Bet A. Also, because of this variability, if you bet the same amount on both bets the "risk of ruin" (chance that your bank roll goes to 0 before realizing the long run return) is a lot greater for Bet B than for Bet A. So if you want to achieve a good balance between "risk of ruin" and long term return, you would bet less on Bet B than on Bet A. Presumably, the algorithm Stake Wizard uses will accomplish this goal better than the standard Kelly formula (or variants thereof) and I don't doubt that it does. So with this in mind, how much SW will recommend that you bet is a function of the absolute level of the fair odds (theoretical probability of winning that bet) and the expected value for the bet. For Bet A (using your 2000 and 20000 numbers) SW will recommend to bet 292.76 while for Bet B it will recommend betting 58.90. If the fair odds calculated by Odds Wizard are absolutely correct I have no reason to don't doubt that these levels are the optimal ones. However, my personal preference is to be less aggressive since the fair odds are only a theoretical number and other factors (injuries, motivation, weather etc) play in as well. So if my actual bankroll is 2000, I might use a bankroll of 500 with a target of 5000 when calculating the bet sizes. If, after several hundred "live" bets, Odds Wizard's fair odds have proven to be very accurate (on the leagues that I am betting on) then I might consider using a “fund” amount closer to my actual bank roll when calculating my bet sizes.

I also know that some successful sport bettors will be risking a certain % of their bankroll on a given bet as you mention. If you only place bets where the probability of winning and expected value is similar e.g. your estimated fair odds is always around 2 and expected value is always around 10% , then this method can make sense. However, if you place bets with varying probability of winning OR with varying expected values, it does not make sense to me for the reasons mentioned above.

For your last question regarding combining bets, it works differently for different bookmakers. Some of the bookies I use will automatically give me the option to place a bet on a “double” if I select two separate bets before entering a bet amount (or “triple” if I select three bets etc). “Double” would mean that both bets would have to win for the bet to be graded as a winner and the odds of the bet would be the individual odds multiplied with each other. Other bookies, such as Pinnacle Sports, will have a separate section (for Pinnacle it is called Parlays/Accumulators) where they have certain bets where they offer the opportunity to combine bets. You would then simply select the bets you would want to combine and then enter the combined bet amount (18 in your example) in the next step. Keep in mind that some bets can’t be combined (in the case of Pinnacle you can only combine bets in the same league as far as I know) and you can only combine bets (in a simple way anyway) by placing them with the same bookmaker. If you have two bets, where the first bet ends before the second one starts, you could "combine" the bets using different bookmakers by placing the suggested bet amount on the first bet and then if it is a winner, place the total initial bet plus the winnings on the second bet. Since I personally find placing combined (parlay) bets to be a hassle (more effort that it is worth), I set the number of accumulators to 1 in Stake Wizard so that Stake Wizard only calculates bet sizes for individual bets. I don’t know how much this will impact the long term expectancy versus allowing for combined bets but since I can’t place many of the combined bets anyway with my bookmakers, it is a moot point (for me).

Hope it helps,
Aim

[danniego]

Thank you Aim for the extensive reply and information within.
In between writing my post and reading your reply I was delving further in the subject and found out what you have explained also. So the consensus and understanding is there
I have read "Optimal Betting by Marvin L. French" about it. Maybe to some other forum readers interest.
Before using OW I used a bet rating allocation. Higher betting on higher (lower odd) probability bets on a 1-3 scale. But using the SW the optimum betting should be more accurately according to the theorem behind it.
Combination bets I also find more an hassle and have set the accumulator to 1 also.
I am using successfully SW for soccer, MLB and ATP Tennis.
Just recently started ATP in play betting with SW but it is to early to tell if SW is spot on with the changing odds but it is in the green for now. Will continue the IP betting this weekend to gather more bets and its results.

[AimHigher]

Glad to hear you are having success and thanks for the tip about the book. IMO, OddsWizard should work best on sports like soccer there the team players are the same (or close) from each game to the next, at least during the season.

The same would be true for baseball, if it wasn't for the pitcher since the pitcher generally has a greater impact on the outcome than any other player. I am still trying to figure out some statistically valid way of adjusting the odds provided by OddsWizard to reflect the pitching matchup using expected/normalized ERA, WHIP etc for each pitcher. The problem I have there is that I don't know if a teams edge (as calculated by OW) is a function of better average offense, defense or pitching. I would be very happy to hear about good solutions to this.

For tennis, my biggest concern with the in-play odds is that they don't seem to take into account (I asked a Alex a question about it but have not received a response) if one player is up a break or if they are both on serve. Since in men's tennis (particularly on fast surfaces) a player is much more likely to win a game if he is serving, the probability of winning a set e.g. if the score is 4:3 varies significantly if the server is up a break or not. I'd be very happy to hear about solutions to this as well, since tennis is one of my favorite sports to watch and it would be fun to be able to confidently bet on it during games as well.

Aim

[danniego]

Watching the ip games I watched with ATP Tennis Men, using live feed of OW, it seems it is doing a good job of presenting the ip odds as since they are most of the time in line with Betfair odds offered. When fair<>offered the offered goes quickly towards the fair given by OW. During tie-breaks, serving for game/set point, etc. It is these short moments when fair<offered thus out of line with OW that I take the offer. Until now it is working good and encouraged me to continue this way. I just finished IP betting Golubev-Istomin but offers where all the time not good enough (to my critera/rules) and hovering around the fair given by OW!
No trades done but increasing confidence of the algorithm OW uses. Let's see for tomorrow ip tennis how it goes with it.
The thing with ip betting is you can gather rather fast a decent sample size if it is profitable or not which is good to quickly abandon the strategy or go with it with full bankroll dedicated to it making bigger and meaning full bets.

Regarding the MLB Pitching stats. I have made a copy of some bookmarked url of interest to you (Maybe you know them, but anyway I post them for you and other forum members bearing an interest in it)
http://www.baseball-reference.com/about/wpa.shtml
http://www.baseballprospectus.com/artic ... cleid=1480
http://www.baseball-almanac.com/stats2.shtml
http://www.baseball-reference.com/about ... sary.shtml
http://www.thespread.com/mlb-baseball-p ... cher-stats

Danniego

[AimHigher]

Thanks for the links. Unfortunately it is not so much that I don't have the data as it is I don't know of a statistically valid model for using that data to adjust OW's fair odds to reflect the pitching match up. Regardless whether I "fix" that or not to be able to bet on baseball, there are more than enough soccer games to keep me busy. I really like that most soccer leagues have started again now that the World Cup is over (which also was fun to watch).

As you implicitly point out, I can definitely see that there are times where the in-play odds in tennis could be really useful and since it would certainly make watching tennis even more exciting, I will start tinkering with it. Your point about being able to quickly get enough data points to evaluate it is also very relevant so I appreciate you pointing that out.

Hope the good fortunes continue,

Aim