RickJ

Hacker,

These are line on Pinnacle. Usually I have just placed these bets before posting on this site.

I am going to post my early ncaa fb plays now. And I just placed these bets on pinnacle

Oh...on the knicks...pinnacle had +5 +103 But other books had the line +5.5 or I even saw a +6 at -110

RickJ

RickJ

I just finished my initial review of ncaa fb this weekend..and these 3 stand out to me as clear percentage plays. I suspect the line will move against us as it did with ucla and arizona st last weekend...ucla went to pick and arizona st moved to -3.

I just finished betting these 3.

Good Luck!

13-Nov-06 4:30pm
16-Nov-06 Football
Pittsburgh U (+10.5) +101

13-Nov-06 9:00am
18-Nov-06 Football
Northwestern (+2.5) +104


13-Nov-06 5:05pm
17-Nov-06 Football
No Illinois (+3.5) for Game -108

I suspect these lines will move against me so want to get these 3 posted and place now.

RickJ
http://rickjshandicappingpicks.blogspot.com/

Record to date:

mlb
Sides: 17- 14 +6.83
o/u 3-8-1 -5.09
ncaafb
Sides 24-20-1 +4.35
o/u 3-5 -2.28
nfl
Sides 13-3-1 +9.85
o/u 4-3 +.83
nhl
sides 17-16 +2.39
Nba
Side 9-8 +.79
o/u 4-3 +.94

Total 177 Bets + 18.47 units +10.43% Return

Hacker

thanks, best of luck with your plays. looks like you know what you are doing from that record

RickJ

Tuesday nba plays

14-Nov-06 7:35pm
14-Nov-06 Basketball
Total Toronto Raptors/Golden State Warriors (U 212) for Game -104

RickJ
http://rickjshandicappingpicks.blogspot.com/

Record to date:

mlb
Sides: 17- 14 +6.83
o/u 3-8-1 -5.09
ncaafb
Sides 24-20-1 +4.35
o/u 3-5 -2.28
nfl
Sides 13-3-1 +9.85
o/u 4-3 +.83
nhl
sides 17-16 +2.39
Nba
Side 9-9 -.21
o/u 4-3 +.94

Total 178 Bets + 17.47 units +9.81% Return

RickJ

Tuesday nhl plays

Selection 1: NHL OT Incl Hockey
Phoenix Coyotes 14-November-2006 6:05 PM PST
Money Line for Game +157

RickJ
http://rickjshandicappingpicks.blogspot.com/

Record to date:

mlb
Sides: 17- 14 +6.83
o/u 3-8-1 -5.09
ncaafb
Sides 24-20-1 +4.35
o/u 3-5 -2.28
nfl
Sides 13-3-1 +9.85
o/u 4-3 +.83
nhl
sides 17-16 +2.39
Nba
Side 9-9 -.21
o/u 4-3 +.94

Total 178 Bets + 17.47 units +9.81% Return

RickJ

Wedn nba,nhl and ncaafb plays

A great day Tuesday. Phoenix came through in the nhl for a +157 win. Those are always nice. And another close one in the nba but we were on the winning side with under on Golden State.

We have hit a new high in percentage return..now over 11%. I have to remind you that my goal is over 5% in the long haul. But its nice to be increasing total plays with %return increasing also. Thats a very good sign!

Good Luck


15-Nov-06 4:35pm
15-Nov-06 Basketball
Boston Celtics (+2) for Game +100

15-Nov-06 4:05pm
15-Nov-06 Hockey
Money Ottawa Senators +153

15-Nov-06 6:05pm
15-Nov-06 Hockey
Colorado Avalanche for Game +128

Selection 1: NCAA Football
Miami Ohio 15-November-2006 4:30 PM PST
Spread +4 for Game -105

RickJ
http://rickjshandicappingpicks.blogspot.com/

Record to date:

mlb
Sides: 17- 14 +6.83
o/u 3-8-1 -5.09
ncaafb
Sides 24-20-1 +4.35
o/u 3-5 -2.28
nfl
Sides 13-3-1 +9.85
o/u 4-3 +.83
nhl
sides 18-16 +3.96
Nba
Side 9-9 -.21
o/u 5-3 +1.94

Total 180 Bets + 20.04 units +11.13% Return

RickJ

Friday nba+nhl+all ncaafb plays

Finally a losing day. I was beginning to wonder. Pittsburgh made a fight of it in the first half but really couldnt hang close. So -1 unit yesterday.

Lot of plays tonight.... 2 nhl plays, 2 nba plays, 2 ncaafb plays and then for sat..5 ncaafb plays.

Good Luck!

17-Nov-06 4:35pm
17-Nov-06 Hockey
Money Dallas Stars +107

17-Nov-06 4:35pm
17-Nov-06 Hockey
Money Ottawa Senators +110

17-Nov-06 5:05pm
17-Nov-06 Basketball
Total Washington Wizards/Detroit Pistons (U 201.5) for Game -107

17-Nov-06 5:05pm
17-Nov-06 Basketball
Memphis Grizzlies (+5.5) -110

17-Nov-06 3:00pm
17-Nov-06 Football
Spread Eastern Michigan (+11.5) -110

17-Nov-06 11:00am
18-Nov-06 Football
Memphis (+17) -108

17-Nov-06 11:00am
18-Nov-06 Football
Total Oklahoma State/Texas Tech (U 69.5) for Game -105

17-Nov-06 5:00pm
18-Nov-06 Football
Total California/USC (U 55) for Game -109

17-Nov-06 8:00pm
18-Nov-06 Football
Total San Jose State/Hawaii (U 73) for Game -105

Also remember I previously posted northern Il and northwestern as ncaafb plays.

Good Luck!

RickJ
http://rickjshandicappingpicks.blogspot.com/

Record to date:

mlb
Sides: 17- 14 +6.83
o/u 3-8-1 -5.09
ncaafb
Sides 25-21-1 +4.35
o/u 3-5 -2.28
nfl
Sides 13-3-1 +9.85
o/u 4-3 +.83
nhl
sides 19-17 +4.49
Nba
Side 10-9 +.79
o/u 5-3 +1.94

Total 185 Bets + 21.57 units +11.66% Return

RickJ

A good day Friday. Split on the nba for -.10 units, Split on nhl for +.07 units, and 2-0 in ncaafb for +2.0 units..so over all a +1.93 day.

A full slate in the nhl but I couldnt find anything there. 3 plays in the nba today. Nfl is tough this Sunday. I have 2 other possibilities but right now they are not plays. And we have 5 ncaa fb plays left for today. So a busy day today.

Good Luck

18-Nov-06 5:35pm
18-Nov-06 Basketball
Spread Indiana Pacers (+3.5) +102

18-Nov-06 5:35pm
18-Nov-06 Basketball
Spread Memphis Grizzlies (+10.5) -107

18-Nov-06 5:35pm
18-Nov-06 Basketball
Total Memphis Grizzlies/Dallas Mavericks (U 190.5) for Game -106

18-Nov-06 1:15pm
19-Nov-06 Football
Total Indianapolis Colts/Dallas Cowboys (U 49.5) for Game -105

18-Nov-06 10:00am
19-Nov-06 Football
Total Cincinnati Bengals/New Orleans Saints (U 51) for Game +100

18-Nov-06 1:05pm
19-Nov-06 Football
Arizona Cardinals (-2.5) for Game +105

Ncaa football plays posted prevously are as follows:


13-Nov-06 10:00am
18-Nov-06 Football
Northwestern (+2.5) for Game +104

17-Nov-06 11:00am
18-Nov-06 Football
Memphis (+17) for Game -108

17-Nov-06 11:00am
18-Nov-06 Football
Total Oklahoma State/Texas Tech (U 69.5) for Game -105

17-Nov-06 5:00pm
18-Nov-06 Football
Total California/USC (U 55) for Game -109

17-Nov-06 8:00pm
18-Nov-06 Football
Total San Jose State/Hawaii (U 73) for Game -105

Good Luck!

RickJ
http://rickjshandicappingpicks.blogspot.com/

Record to date:

mlb
Sides: 17- 14 +6.83
o/u 3-8-1 -5.09
ncaafb
Sides 27-21-1 +6.35
o/u 3-5 -2.28
nfl
Sides 13-3-1 +9.85
o/u 4-3 +.83
nhl
sides 20-18 +4.56
Nba
Side 10-10 -.31
o/u 6-3 +2.94

Total 191 Bets + 23.54 units +12.32% Return

thegreat

Rick just had a look at you blog.

How are you finding the way you are being treated by your government regarding punting or are you ok with the sportsbooks you can use in Vegas ?

__________________
___________


I'm a working man from Lancashire

And I want a Chippy Tea !!!!

Hacker

also, any good websites to learn a little about neural net analysis, mentioned on your blog? everything i've found so far is very hand-wavey, no real explanation of how it works. the more maths the better.

RickJ

I use pinnacle sports book and have never had a problem with the us govt. You could bet the casinos lines but they are -110 and 20cent lines in money lines. So your giving up probably at least 3% in vig over pinnacle.

Neural net is a vague term. But without disclosing anything about my methodology it is best used uncovering inefficiencies in the lines. So far using this method I have a 15% return on 200 bets. I cant imagine the lines being that inefficient so I am sure variance has something to do with it. My goal when starting this was to produce over a 5 % return.
When I hit 500 to 1000 bets it will be interesting to see the % return!

There are literally 100s of ways to use neural net analysis the trick is to find one that works. Of course that is just one of the factors I use..just one variable in the equation in coming up with plays.

RickJ

RickJ

8-0 SAT!!!!

I am still in shock. An 8-0 day. 5-0 in ncaafb and all were easy wins. 3-0 in nba also easy wins. Scary!

This morning I looked over the nfl again and would really have to stretch to find any more plays so the 3 below that were previously posted are it. 1 nba play and 1 nhl play today also. So 5 plays in all. We are at the 200 play mark and over 15% return. Thats pretty scary also considering if the lines were close to even each way it would be very tough to get this type of return.

My method of uncovering inefficiencies in the lines seems to be working well so far and works in all sports!

Good luck today


18-Nov-06 6:05pm
19-Nov-06 Basketball
Sacramento Kings (+2.5) for Game +102

18-Nov-06 7:05pm
19-Nov-06 Hockey
Money Chicago Blackhawks +184

Previously posted nfl plays:

18-Nov-06 1:15pm
19-Nov-06 Football
Total Indianapolis Colts/Dallas Cowboys (U 49.5) for Game -105 210.00 200.00


18-Nov-06 10:00am
19-Nov-06 Football
Total Cincinnati Bengals/New Orleans Saints (U 51) for Game +100 200.00 200.00

18-Nov-06 1:05pm
19-Nov-06 Football
Arizona Cardinals (-2.5) for Game +105

RickJ
http://rickjshandicappingpicks.blogspot.com/

Record to date:

mlb
Sides: 17- 14 +6.83
o/u 3-8-1 -5.09
ncaafb
Sides 29-21-1 +8.39
o/u 6-5 +.72
nfl
Sides 13-3-1 +9.85
o/u 4-3 +.83
nhl
sides 20-18 +4.56
Nba
Side 12-10 +1.71
o/u 7-3 +3.94

Total 199 Bets + 31.60 units +15.88% Return

Hacker

jay-sus - you'll have me following you in blind if you carry on like that. congratulations.

no problems about not wanting to share your methods in detail, totally understandable. just wondering if i can find out anything more in that area

most articles i have come across seem to cast neural net analysis in a dubious light - the claim that it "predicts the most likely result" has come up several times

i can see why non-punters would be drawn to such an imaginary pot of gold, but i feel certain that this is a misleading description of what neural nets are really about.

anyone who knows anything about handicapping would hardly spend all their time trying to work out which result was favourite. my (naive) understanding is that it finds overlays by not getting drawn into generally held biases - but surely this would be the aim of any decent handicapping method?

patrickstar

There is a lot of cloud built up in my memory between now and my uni days, so hopefully I won't say anything too wildly inaccurate here..

Essentially, there is nothing 'special' about neural nets. Much like genetic algortihms they got a fancy reputation in artificial intelligence circles because they drew from biology for inspiration.. at one time they were even seen as being a candidate for creating an artifical human brain.

But just as genetic algorithms are merely one (and rarely the best) method of iterative searching of a solution space, neural nets are just one type of learning algorithm, which can be used when you want to know the 'answer' to the next 'question' and you have a historical sample of data containing a lot of old 'questions' and their corresponding 'answers'.

A neural net, from the outside, works in exactly the same way as a normal regression model - you have input variables (questions), and you have output variables (answers).
And just like a normal linear regression, on the 'inside' you obtain the outputs from the inputs via some mathematical formula / set of weightings.

You train the model (i.e. in a normal simple linear regression model learn the weightings in the equation y=ax+b) by feeding it a set of examples where both the inputs and outputs are known.

An advantage of neural nets is that they are capable of representing nonlinear functions (in fact you could argue that all neural nets are is a representation of a non-linear function). Thus when you are dealing with data which follows an underlying non-linear relationship it is obviously possible to better model it than you could with a linear model.

Ironically the disadvantages of neural nets are brought about by the same complexity that gives them their strength: As with all learning algorithms, the more degress of freedom (i.e. variables to be learned) you have in a model the more you run the risk of modelling the noise and not the signal, hence the more quality and size you need in your training set of data in order to fit them.
Furthermore, they can be so complex that you can't really tell what is going on inside them.. any sort of attmepts to fine-tune or tinker with, or repair a neural net if it starts going wrong can be absolutely futile - you change one weighting a bit and suddenly everything changes and you have no idea why.



Yes it would.. and it all just depends how you use neural nets really. If you feed in a set of examples of runners (say their form figures for example) together with who won the race/match and then ask it to predict who will win the next race then it will just spit out the favourite at you (assuming it's learned enough from the data you've put in to manage to correctly identify who that is).

You can of course set a neural net up to give probabilistic outputs, but that is hardly anything special or unique, or anything specifically to do with it being a neural net.

To summarise, neural nets can and have been used to good effect in real world data analysis problems (including sports handicapping as seems to be being done here), but there is nothing 'magical' about them - they are powerful, but at the same time they are notoriously tricky and unstable, and they still must obey the same golden rules of all data analysis techiques:

1. In the real world, there is no perfect model. All you can do is get as close an approximation as possible to the real underlying effect given the (inherrently limited) data you have.

2. GIGO - garbage in garbage out - no matter how fancy your model is (and in real life often the fanciest models are the worst performers) it can only ever be as good as the data you train it on.

3. If you don't understand how your model works then you don't understand how (and when and why) it doesn't work, hence you are on dangerous ground using it.

patrickstar

just to add, keep it up RickJ - going great so far for you 5% return sounds like a sensible and achievable goal to me.

RickJ

I agree entirely with neural nets and using models for predictive purposes. But they provide a good start to a handicapping routine. But just a start! I fooled around with various models over the years and really have not had much success. It wasn't until I started a new methodology with using a neural net model as a starting point..and then using approximately 6 other steps following to filter and refine predictability that I feel I have been able to uncover inefficiencies in the lines in just about any sport.

But like I said...I dont imagine in the long run you can do much over a 5% return. Right now im at 204 bets with this approach and a 16% return. Nothing like variance to help you out! As it does with many poker players. But its always nice to be on the right side of variance once in awhile.

It will be interesting to see where I come out after 500 to 1000 bets.

RickJ

RickJ

Monday nba + nhl plays

Here they are. I actually thought there would be a few more based upon the overnight lines but it didnt pan out.

A good day yesterday. 3-0 in the nfl, but lost the nhl play and nba play for +1.05 units.

Good Luck!

20-Nov-06 4:35pm
20-Nov-06 Hockey
Money New York Islanders +135

20-Nov-06 5:05pm
20-Nov-06 Basketball
Memphis Grizzlies (+3.5) -105

RickJ
http://rickjshandicappingpicks.blogspot.com/

Record to date:

mlb
Sides: 17- 14 +6.83
o/u 3-8-1 -5.09
ncaafb
Sides 29-21-1 +8.39
o/u 6-5 +.72
nfl
Sides 14-3-1 +10.90
o/u 6-3 +2.83
nhl
sides 20-19 +3.56
Nba
Side 12-11 +.71
o/u 7-3 +3.94

Total 204 Bets + 32.65 units +16.00% Return

RickJ

A 1-1 day yesterday for breakeven. Memphis came through in the final quarter and won big as a dog. Islanders really were never in the game. But were within 1 goal late.

No nhl plays tonight. 1 nba play. 3 early ncaafb plays.

Good Luck


20-Nov-06 5:35pm
21-Nov-06 Basketball
Total Washington Wizards/Dallas Mavericks (U 205.5) -105

21-Nov-06 9:00am
24-Nov-06 Football
Miami Ohio (+3) -107

21-Nov-06 5:00pm
25-Nov-06 Football
Total Notre Dame/USC (U 57) for Game -111

21-Nov-06 1:00pm
25-Nov-06 Football
Nevada (+2.5) -102

RickJ
http://rickjshandicappingpicks.blogspot.com/

Record to date:

mlb
Sides: 17- 14 +6.83
o/u 3-8-1 -5.09
ncaafb
Sides 29-21-1 +8.39
o/u 6-5 +.72
nfl
Sides 14-3-1 +10.90
o/u 6-3 +2.83
nhl
sides 20-20 +2.56
Nba
Side 13-11 +1.71
o/u 7-3 +3.94

Total 206 Bets + 32.65 units +15.85% Return

bambam

I see you are Gold now Rick, welcome on board.

I'm an ignaramus on US sports\lines etc - so it's all goobldygook to me

RickJ

I know what you mean ...I am as ignorant trying to figure out what 1.75 means