Posted:
#26
2014-05-19
HOU @ LAA, u 7.5, +105
CIN @ WSH, u 6.5, -100
0
Over under betting system
5-2
4-3
Posted:
#30
@King Creole: Thank you, sir!
@lakshow: Thank you, too! It is really as easy as I told before. I don't do anything different than looking at the recent over/under trends for both teams on the road/at home and the recent over/under trends for the starters. Then I sum it up and depending on how well the game or system pick meets the criterias from post no. 18 I place a bet or stay away from it. If you like I can give you yesterday's games and calculations as an example which should explain the system as a whole. It's really not very difficult.
0
Posted:
#31
There were 5 games yesterday:
- CIN @ WSH
- DET @ CLE
- MIL @ ATL
- CWS @ KCR
- HOU @ LAA
Mike Minor (ATL) and Scott Carroll (CWS) both had less than 5 starts over the last 55 days, so both the Braves and the Sox game were no allowed plays.
This leaves the following games:
1) CIN @ WSH.
We need to go to sportsdatabase.com.
First we query the stats for the Reds' away record over the last 35 days. They had 6 overs, 9 unders, 0 pushes.
Then we query the stats for the Nats' home record over the last 35 days. They had 6 overs, 9 unders, 2 pushes.
After that we look into the recent (55 days, so in our case the whole season) overall stats of Mike Leake and see that 2 overs, 5 unders, 1 push.
Then we do the same for Stephen Strasburg, in his starts the scores were 6 overs, 3 unders, 1 push.
Then we sum it up: 20 overs, 26 unders, 3 pushes.
After that we sum up the overs and the pushes and divide it by the sum of overs+unders+pushes: (20+3)/(20+26+3)=0.469.
Next we do the same for the unders: (26+3)/(20+26+3)=0.592. So the system suggests an under, as the (let's call it) chance for over is 0.469 and for under is 0.592.
We meet all the criterias for a under bet (see post no. 29: both starters at least 5 starts, total line offered by the bookie 6.5 and therefore higher than than 6.0), so we place a bet on under.
Result was 4-3 in extra innings (2-2 after 9 innings, sadly), we lost the bet.
2) DET @ CLE.
Again we go to sportsdatabase.com.
First we query the stats for the Tiger's away record over the last 35 days (5 o, 8 u, 0 p).
Then we query the stats for the Indians' home record over the last 35 days (8 o, 9 u, 0 p).
After that we look for Drew Smyly's stats: 3 o, 2 u, 0 p.
Same for Corey Kluber: 4 o, 5 u, 0 p.
We sum it up: 20 o, 24 u.
We calculate the chance for over: (20+0)/(20+24+0)=0.455.
We calculate the chance for under (24+0)/(20+24+0)=0.545.
The system suggests under. The starters would have met the criterias, but the chance for under is too low. We don't place a bet.
3) HOU @ LAA.
Sportsdatabase.
Astros away: 4 o, 8 u, 1 p.
Angels home: 7 o, 7 u, 2 p.
Keuchel: 5 o, 3 u, 0 p.
Richards: 2 o, 5 u, 1 p.
Sum: 18 o, 23 u, 4 p.
Chance for over: (18+4)/(18+23+4)=0.489.
Chance for under: (23+4)/(18+23+4)=0.600.
The system suggests under. Starters meet the criteria. Chance for under is higher than 0.590. Total line by the bookie meets the criteria (was 7.5 and therefore higher than 6.0). We place a bet on under.
Result was 5-2, we won the bet.
That's basically all what I do and what I have backtested. I was really surprised that it worked so well, but I do like this system (or what you ever wanna call it very much).
It has a ton of pros IMO:
- comparing it to conventional (let's call it) betting by brain and picking your bets randomly on your feel it gives you fixed criterias
- you don't need a chase system which seemingly a lot of guys use to make profits on baseball (what I don't like about chases is the fact that you need a certain bankroll and you have not enough control IMO; I know a lot of guys make profit with chasing but I am a different person)
- you don't need much time to calculate it (as I am no Excel-crack I query everything manually and then put it into an Excel-sheet which sums up and calculates everything for me)
- if you like to, you can bet every game in a season after (approximately) mid May
- if you like to, you can take some days off and come back when you like to (that's another advantage over chasing)
- to win with conventional betting one has to outsmart the bookie somehow IMO; with the system above you don't have to, at least it doesn't feel like doing so
- when thinking about over and under stats and lines in a different way, one could come to the conclusion that everything can be found in them: starters performance, whole teams pitching performance, whole teams offensive performance, overachievers and underachievers, streaks, and so on
That's all for now which I could think of. If you have any further questions please feel free to ask. I don't expect to carry a 70% win rate until the end of the season, but 60% are highly possible and this would be very nice compared to the very low amount of work I have to put in it.
0
I had no luck lol
Posted:
#34
In general the queries look like this (the brackets have to be filled of course):
1. Query for teams:
"date>=[first date=35 days ago] and date<=[second date=today] and team=[team name] and site=[home or away]"
2. Query for starters:
"date>=[fist date=55 days ago] and date<=(second date=today] and starter=[starter name]"
Example: BAL @ PIT today (2014-05-20), starters M. Gonzalez and F. Liriano:
1. Query for teams:
1a. Baltimore:
- query: "date>=20140415 and date<=20140520 and team=Orioles and site=away"
- result: 8 over, 9 under, 0 push
1b. Pittsburg:
- query: "date>=20140415 and date<=20140520 and team=Pirates and site=home"
- result: 10 over, 7 under, 0 push
2. Query for starters:
2a. Miguel Gonzalez:
- query: "season=2014 and date<=20140520 and starter=Miguel Gonzalez" (normally we would have to use a "first date" here, but the season isn't old enough, so we can still use only the second date as a filter)
- result: 5 over, 2 under, 0 push
2b. Francisco Liriano:
- query: "season=2014 and date<=20140520 and starter=Francisco Liriano"
- result: 3 over, 5 under, 1 push
3. sum it up:
- 26 over, 23 under, 1 push
- 0.54 over, 0.48 under
- no bet
Got it?
0
Posted:
#35
Another example in wich I switched out the "season=2014" for starters, maybe it could be misleading, although the results aren't any different to the query results below.
SFG @ COL, M. Bumgarner vs. F. Morales:
1. Teams:
1a. Giants:
- query: "date>=20140415 and date<=20140520 and team=Giants and site=away"
- result: 4 over, 11 under, 1 push
1b. Rockies:
- query: "date>=20140415 and date<=20140520 and team=Rockies and site=home"
- result: 9 over, 6 under, 0 push
2. Starters:
2a. Bumgarner:
- query: "date>=20140325 and date<=20140520 and starter=Madison Bumgarner"
- result: 5 over, 4 under, 0 push
2b. Morales:
- query:"date>=20140325 and date<=20140520 and starter=Franklin Morales"
- result: 4 over, 4 under, 0 push
3. sum it up:
- 22 over, 25 under, 1 push
- 0.48 over, 0.54 under
- no bet
0
Posted:
#38
Because of the results from the backtesting. 
Let's assume that the average of all odds you pick over the course of the season is about -125 (indeed they should come closer to -115 but let's take this as an assumption).
When you bet 1 unit on odds of -125 you have to win at least .56-.57 pct of your bets to make profit. My goal was to find criterias (see post #29 for my current ones) which come at least close to a winning pct of ,59, because the profit of a .56-.57 win pct would only be marginal.
I looked around and tried different criterias for chosing bets in my backtest and ended up with the ones I posted.
Therefore I have to say, because I want to be honest, that I still was not able to backtest a whole season! But: My backtest consists of a little bit more than 1200 games right now, which leads me to the assumption that the backtest of the whole season won't be much different in the end.
Indeed I tried a few different things regarding criterias:
1) Only play bets where the calculation gives the over or under at least.700. The problem was here that they indeed won on a really high average, but since the calculated pct is more than .700 you would play about (exaggerated) 1 or 2 bets per month if at all.
2) Then I drove down the percentage required for a bet to take place down more and more in a few steps, from .650 to .625 and ended up with .550. But those bets didn't show a win pct of at least .590 which I was aiming for. But .590 seemed to be something like a "sweet spot".
3) After that I looked for some other criterias, e.g. that starters should have at least 5 starts over the last 55 days and was able to increase the win pct a little bit more. I also found out that betting over on totals of higher than 8.5 would hurt more than it helps, same thing in the other direction with betting on unders when the total was 6.0 or below.
Hope this helps. I am still (indeed right now) doing more backtesting. The results arent't any different than what I have told before. It still shows me a win pct of about .600 when using the criterias I made public. If you have any more questions feel free to ask.
0
Posted:
#39
Quote Originally Posted by lakshow25:
Thanks I'll try it out, any picks for today.
The only pick was the Cards game from post no. 32.Thanks I'll try it out, any picks for today.
There are some games which won't be looked into because of some starters have less than 5 starts. All the other possible games don't have at least a over or under pct. of .590 or higher.
Funny thing (at least 'somehow' funny) is that those accumulated trend numbers and percentages indeed point towards a certain direction.
When we look at the Cards game of today, which is my only pick, we get:
- ARI away: 6 o, 8 u, 2 p
- STL home: 6 o, 3 u, 3 p
- Arroyo: 3 o, 4 u, 1 p
- Wainwright: 4 o, 2 u, 2 p
We get a lot of pushes. Therefore I expected the line to not be a whole number. But I got the line at 7.0. Misstake by the bookie? Maybe, don't know.
Besides that I got the over 7.0 for an odd of +105 early in the morning at 8:50 a.m. (take into account that I am from Germany and right now as I write this post it's 4:13 p.m. over here). The under was a dog.
I don't know what's your opinion on it, but punters in Germany have the assumption that "beating the closing line" is a good indicator for picking value. And guess what? Right now the numbers have changed. Meanwhile the over was favored and is at -111.
I have noticed that quite some time since starting this blog, that when I pick early, the numbers often change in my favor.
Besides that look at this:
- Wainwright got a run support on average of 6.28 in his last 7 starts
- the Cards scored 5.1 runs on average in their last 10 at home
- the Cards scored not less than 3 runs in their last 10 games (9 times at least 4, 6 times at least 5)
- when Arroyo started, the Dbacks conceded 5 or more runs in 4 of his 8 starts
- in the last 15 Dbacks away games the total score was 7 runs or more 10 times and they conceded 4.53 runs on average in those games
- the Dbacks scored more than 2 runs in 17 of their last 22 away games and at least 3 in each of their last 8 away games
Now please look only at those numbers. Would you expect a total of 7? To be honest I wouldn't. I would at least expect a bookie total of 7.5.
The nice thing is I would have chosen the over if I had looked into those numbers. But I didn't have too. It was all 'hidden' in the recent over/under stats. Somehow the bookies didn't get it.
0
Posted:
#41
Maybe two other recent examples. Both picks from yesterday.
I still am little bit grumpy because of Aroldis Chapman. If he had been able to close this damn game, I would have won both bets. The next foolish things happened in extra innings. I have lost the CIN @ WSH pick, but 1) it was close, very close and 2) I call it bad luck.
My pick was: CIN @ WSH, u 6.5, -100
The closing line was: u 6.5 -114
Meaning that I was able to beat the closing line.
Stats before the game:
Reds:
- scored only 3.07 runs per game over the last 14 overall and only 2.2 runs per game over the last 9 away games
- with Mike Leake starting the Reds have conceded 3.38 runs per game
Nationals:
- scored only 2.86 runs per game over the last 14 overall and only 3.5 runs per game over the last 8 home games
- with Stephen Strasburg starting the Nats had conceded 4.00 runs per game overall but only 2.2 runs per game over his last 5 starts
- scored only 3.07 runs per game over the last 14 overall and only 2.2 runs per game over the last 9 away games
- with Mike Leake starting the Reds have conceded 3.38 runs per game
Nationals:
- scored only 2.86 runs per game over the last 14 overall and only 3.5 runs per game over the last 8 home games
- with Stephen Strasburg starting the Nats had conceded 4.00 runs per game overall but only 2.2 runs per game over his last 5 starts
I hadn't looked into those numbers when placing the bet, I only looked into the over/under 'hidden trends'. The odds changed into my direction and the stats (which I didn't know when picking) would also suggest a bet on the under IMO. I still consider picking the under as a good bet, sadly I lost it.
Similar to this goes the other pick.
Second pick was HOU @ LAA, u 7.5, +105.
Closing line at pinnacle: u 7.5 for +103. Not big of a difference, but I got the better line early in the morning.
Taking a look into the stats pregame we could see:
Astros:
- scored 4.08 runs per game over their last 13 but only 3.56 runs per game in their last 9 away games
- with Keuchel starting the Astros have conceded 4.25 runs per game but only 2.67 in away games
Angels:
- scored only 3.63 runs per game over their last 16 and 3.7 runs per game in their last 10 home games
- with Richards starting the Angels have conceded only 2.75 runs per game
- scored 4.08 runs per game over their last 13 but only 3.56 runs per game in their last 9 away games
- with Keuchel starting the Astros have conceded 4.25 runs per game but only 2.67 in away games
Angels:
- scored only 3.63 runs per game over their last 16 and 3.7 runs per game in their last 10 home games
- with Richards starting the Angels have conceded only 2.75 runs per game
IMO those numbers point towards a bet on u 7.5, too. But I didn't have to look into them, I got it from the 'hidden trends' of overs and unders.
Know what I mean? I found this out randomly, too. I don't know if it's coincidence or not. But somehow this is eye-catching when taking a deeper look at it.
0
Posted:
#44
Question???...I have been trying to query to receive the information on sportsdata...I am entering the same formula you suggest here but the data it is returning is for every single major league game played between the dates requested...I will cut and paste the formula below as an example...Any idea why this might be happening?...thanks for your time man
'team=Orioles and site=away and date>=20140415 and date<=20140520'
0
Posted:
#45
@lakshow:
Yeah you're right. It wouldn't be realistic to expect to be able to carry a 0.7 win pct over the whole season.
@BenchCoach:
Yeah, that's what it's supposed to do. When you look at the upper left you got to move your eyes to the second line which starts with
'OU' which means 'over/under'. The next numbers are the relevant ones. For the O's you get '9-9-0', meaning 9 over, 9 under, 0 push.
0
Posted:
#47
Additional facts for the game:
Athletics:
- with Milone starting the total score went over 7.5 in 5 of his 7 starts
- in the last 14 road games of the As the total went over 7.5 10 times
Rays:
- with Bedard starting the total score went over 7.5 in 5 of his 6 starts
- in the last 14 home games of the Rays the total went over 7.5 10 times
- with Milone starting the total score went over 7.5 in 5 of his 7 starts
- in the last 14 road games of the As the total went over 7.5 10 times
Rays:
- with Bedard starting the total score went over 7.5 in 5 of his 6 starts
- in the last 14 home games of the Rays the total went over 7.5 10 times
0
Hot Forum Topics
