[bettilimbroke999]

If you had only bet on dogs with at least 2/1 return (+200) in MLB 07' regular season betting 1 unit per game how would you have come out, ahead or behind?

[durito]

www.covers.com

[azrob]

After compiling all of the team data from Covers.com and writing a VB.NET application that lets me easily process it and run tests, I've already found some interesting things.

First, in answer to the OP's question... I get 648-921 and -12.92 units for that system throughout the 2007 season. But if you only bet on teams with a ML of +133 or better, the results are 319-502 and +24.12 units. If someone can easily verify that for me, that would be great. I'm still in debugging mode.

Also, there was a thread on here somewhere about the Baseball Underdog system, where you bet on dogs with odds of less than +250 (3.500) not currently on a 3+ game losing streak, not facing an opponent on a 3+ game winning streak, and not facing a pitcher in the top 20 when ranked by ERA. According to most, that was a losing system but I can't test it with my data because I don't have the historical pitcher rankings.

But just for fun, I simplified it a bit by eliminating opposing pitchers with a win/loss percentage of 0.500 or better. I also eliminated opposing teams with a win/loss percentage of 0.500 or better.

When betting on home teams, the results were:

2007 Season... 118-120, +18.18 units
2006 Season... 120-121, +17.25 units

When betting on away teams, the results were:

2007 Season... 106-140, -7.39 units
2006 Season... 85-119, -14.65 units

Pretty modest gains, but it looks like a working system if you stick to the home teams only. Again, if anyone can verify these results for me that would be great.

[pimike]

Originally Posted by bettilimbroke999

If you had only bet on dogs with at least 2/1 return (+200) in MLB 07' regular season betting 1 unit per game how would you have come out, ahead or behind?

behind, the fav wins 8 times out of 10 when dog is plus 200 or more.

[azrob]

Correction on my 2.000+ underdog results... it was 956-1303 and -19.02 units for 2007. As for the 2.333+ underdog results, I got 400-615 and +32.4 units.

My previous results were for Away teams only.

Also, I misread the OP's question, he was asking for +200 ML dogs, which is 3.000+ EU odds. For that I get 37-75 and +9.42 units. This does not match up at all to what pimike said, which makes we think something is amiss. It's late, that's all I can say. :-)

[pimike]

Understand its late, get your numbers straight then we will talk.

if you bet 100 3 times on plus 200 it loses your down 300. u win next one your down 100. after 4 plays.

if u were to bet 3 times at 270 it wins your up 300 you lose next one at 270 your up 30. stay off those bigl lines no value

big dan has played the last three hockey games plus 195 but lost all three. so if he takes philly tomorrow he wins he is still down because he will not get 300 plus on a line for that.

[azrob]

Well, I understand how odds work. But consider this...

My results were 37-75. Which means that if you straight bet $100 per game, you lost $7500. But you won at least $7400 (37 x $200). The +9.42 units is theoretically possible since it's only minimum odds of +200. If every game had the underdog ML at exactly +200, you would lose 1 unit. That is, assuming my win-loss numbers are correct. Tomorrow I can export the data for those 112 games to a spreadsheet for manual verification.

Yay?

[azrob]

Okay, I exported those games to a file and took a look at it. Adding up all of the units won on the underdogs, I got 82.29 units. Subtract out the 75 units you would have lost and you get +7.29 units. I'm not sure why the discrepancy between the +9.42 units my app came up with; it's possible I mis-keyed a number.

There were some pretty good upsets, like the Washington/Minnesota game on June 9, 2007. The underdog Nationals closed at +375 and won 3-1.

[Ganchrow]

Originally Posted by azrob

asking for +200 ML dogs, which is 3.000+ EU odds. For that I get 37-75 and +9.42 units.

At fixed stake I get 38-75 for +11.43 units (an ROI of 10.12%).

Based on the theoretical hold expected return was -1.92 units so results are 13.35 units better than expected.

Standard deviation over the 113 bets would be 13.33 units (based on theoretical hold -- 13.35 units assuming bets offered at 0 EV) so results are about 1.00 standard deviations better than expected and about 0.86 standard deviations better than breakeven.

[azrob]

Ganch - Your wins were close to mine, but I wonder why the huge discrepancy in the losses? I think you typed in 45 when you meant 75 (it IS late). The number of games I exported were indeed 112, and I can see that every one of them had ML odds of +200 or higher, predominantly for the Away teams. In fact, your numbers can't be right because you would only lose 45 units and gain at least 76 units at +200 or higher. Or am I missing something? I have to admit I don't understand the statistics talk. But I did understand your first sentence!

What is the "theoretical hold"?

[Ganchrow]

Originally Posted by azrob

Ganch - Your wins were close to mine, but I wonder why the huge discrepancy in the losses? I think you typed in 45 when you meant 75 (it IS late).

Yes 38-75. 45 was indeed a typo.

Originally Posted by azrob

What is the "theoretical hold"?

http://forum.sbrforum.com/players-ta...ical-hold.html

[azrob]

Hey, Ganch... I attached my 112-game data file, which is comma-delimited. The first three fields are date, away team and home team. If it's not too much to ask, can you tell me which game my app dropped? I think you're getting 113 games to my 112. I was doing regular season only, not playoffs. Maybe that's why.

[Ganchrow]

Originally Posted by azrob

Hey, Ganch... I attached my 112-game data file, which is comma-delimited. The first three fields are date, away team and home team. If it's not too much to ask, can you tell me which game my app dropped? I think you're getting 113 games to my 112. I was doing regular season only, not playoffs. Maybe that's why.

You attached the 2006 file.

[azrob]

I apologize, I'm retarded. I re-ran the app for the 2006 season and attached that file by mistake. Here is the correct file.

[Ganchrow]

Originally Posted by azrob

I apologize, I'm retarded. I re-ran the app for the 2006 season and attached that file by mistake. Here is the correct file.

Yeah it was a post-season game so no problem with your script.

The Indians were a +201 dog in the Bronx on 10/8/07 and won 6-4.

Take the game out of the equation and we're at 37-45 +9.42 for an ROI of 8.34%. That's 0.85 standard deviations better than expected and 0.71 better than breakeven.

[azrob]

Ganch - Thanks, I'm feeling pretty good about it now!

fgolly - Post it up, it will be interesting for all to see.