NBA Stituational Bet, SDQL

dmitean · 12-11-14, 02:46 AM

conference = Western and o:conference = Eastern and WP<30 and o:WP<40 and playoffs=0 and season>2010

13-25-4 (-6.17, 34.2%)

avg total: 194.2

No special way to explain it, but when two bad teams from different conferences meet, bet the Under.
The key here is very bad Western team - the moment you put WP<40 instead of WP<30, the Under looks a lot worse.

It's also very active lately:
conference = Western and o:conference = Eastern and WP<30 and o:WP<40 and playoffs=0 and season>2013
O/U:2-10-1 (-6.15, 16.7%) avg total: 201.5

JMon · 12-11-14, 08:37 AM

<!--[if gte mso 9]><xml> <w:WordDocument> <w:View>Normal</w:View> <w:Zoom>0</w:Zoom> <w:TrackMoves/> <w:TrackFormatting/> <w:PunctuationKerning/> <w:ValidateAgainstSchemas/> <w:SaveIfXMLInvalid>false</w:SaveIfXMLInvalid> <w:IgnoreMixedContent>false</w:IgnoreMixedContent> <w:AlwaysShowPlaceholderText>false</w:AlwaysShowPlaceholderText> <w

oNotPromoteQF/> <w:LidThemeOther>EN-US</w:LidThemeOther> <w:LidThemeAsian>X-NONE</w:LidThemeAsian> <w:LidThemeComplexScript>X-NONE</w:LidThemeComplexScript> <w:Compatibility> <w:BreakWrappedTables/> <w:SnapToGridInCell/> <w:WrapTextWithPunct/> <w:UseAsianBreakRules/> <w

ontGrowAutofit/> <w:SplitPgBreakAndParaMark/> <w:EnableOpenTypeKerning/> <w

ontFlipMirrorIndents/> <w:OverrideTableStyleHps/> </w:Compatibility> <m:mathPr> <m:mathFont m:val="Cambria Math"/> <m:brkBin m:val="before"/> <m:brkBinSub m:val="--"/> <m:smallFrac m:val="off"/> <m:dispDef/> <m:lMargin m:val="0"/> <m:rMargin m:val="0"/> <m:defJc m:val="centerGroup"/> <m:wrapIndent m:val="1440"/> <m:intLim m:val="subSup"/> <m:naryLim m:val="undOvr"/> </m:mathPr></w:WordDocument> </xml><![endif]--> NBA SDQL "Best Bet" 2014-15

Overall Group Record - 65-58-2 52.8% (+3.43) (max wagered -to win 1 unit)

Yesterday's recap 2-3 (-1.16)

Official Entries: Max 10 of 10:

1. JMon - 12-10 (+1.90)

2. pip2 - 11-10 (+.38)

3. nash13 - 6-4 (+1.82)

4. Ronald S. - 9-10 (-1.65)

5. hyahya - 6-4-1 (+1.55)

6. Mako-SBR - 10-10 (-1.00)

7. dmitean - 0-2 (-2.04)

8. Consigliere - 8-4-1 (+3.77)

9. GolfAddict - 3-4 (-1.30)

10. open spot

emceeaye · 12-11-14, 01:27 PM

Kings ATS 1st quarter

active again today:

p:M3 >= 3 and (A or D) and p:L and rest = 0 and [M1 > - line / 4] = [False]

As long as away team isn't too much of a favorite, the trend still applies:

p:M3 >= 0 and AF and -2<=line<0 and p:margin<0 and rest = 0

dmitean · 12-11-14, 04:51 PM

NBA best bet:

Oklahoma -4, -111 odds

SDQL: conference = Western and o:conference = Eastern and WP>60 and o:WP>60 and playoffs=0 and season>2010
SU:52-44 (2.54, 54.2%)ATS:56-38-2 (2.06, 59.6%) avg line: -0.5O/U:43-50-3 (0.77, 46.2%) avg total: 196.1

English: When good teams from different conferences with similar records meet, bet the one from the West.

pip2 · 12-11-14, 05:19 PM

NBA SDQL 'BEST BET' 2014-15

12/11/14 11:00am pst NBA Basketball 704 Sacramento Kings 1st Quarter -½ +110* vs Houston Rockets

SDQL:

Generic to find an applicable game: p:M3 >= 1 and (A or D) and p:L and rest = 0

For today's game to see if it is a bet: p:M3 >= 1 and (A or D) and p:L and rest = 0 and -5 < line < 3 and [M1 > 0]
(At which point you need to compare the number of 'falses' with the number of 'trues'.)

SDQL: English

For the first quarter fade the away or dog team that lost just last night after holding a lead at the end of the third quarter.

JMon · 12-11-14, 06:32 PM

Originally posted by dmitean

NBA best bet:

Oklahoma -4, -111 odds

SDQL: conference = Western and o:conference = Eastern and WP>60 and o:WP>60 and playoffs=0 and season>2010
SU:52-44 (2.54, 54.2%)ATS:56-38-2 (2.06, 59.6%) avg line: -0.5O/U:43-50-3 (0.77, 46.2%) avg total: 196.1

English: When good teams from different conferences with similar records meet, bet the one from the West.

OKC doesn't have a win % over 60

Ronald S. · 12-11-14, 07:13 PM

Sorry should have said earlier... no best bet for me today. Been rough goings lately

emceeaye · 12-11-14, 07:54 PM

Originally posted by pip2

NBA SDQL 'BEST BET' 2014-15

12/11/14 11:00am pst NBA Basketball 704 Sacramento Kings 1st Quarter -½ +110* vs Houston Rockets

SDQL:

Generic to find an applicable game: p:M3 >= 1 and (A or D) and p:L and rest = 0

For today's game to see if it is a bet: p:M3 >= 1 and (A or D) and p:L and rest = 0 and -5 < line < 3 and [M1 > 0]
(At which point you need to compare the number of 'falses' with the number of 'trues'.)

SDQL: English

For the first quarter fade the away or dog team that lost just last night after holding a lead at the end of the third quarter.

Another way to determine the likelihood that the home team will win in this situation is to compare the differences (margin) between the average Q1 score and average Q2 score in this situation with that of other situations where the margin is more or less to differing degrees and when the previous game's 3rd quarter margin was more or less to differing degrees.

dmitean · 12-11-14, 09:54 PM

Originally posted by JMon

OKC doesn't have a win % over 60

Yes they are. Since getting Durant and Westbrook back, they are 3 - 1 = 75%.

dmitean · 12-12-14, 07:37 AM

Originally posted by dmitean

NBA best bet:

Oklahoma -4, -111 odds

SDQL: conference = Western and o:conference = Eastern and WP>60 and o:WP>60 and playoffs=0 and season>2010
SU:52-44 (2.54, 54.2%)ATS:56-38-2 (2.06, 59.6%) avg line: -0.5O/U:43-50-3 (0.77, 46.2%) avg total: 196.1

English: When good teams from different conferences with similar records meet, bet the one from the West.

Finally a winner, after two bad first attempts

Heart · 12-12-14, 10:33 AM

Can someone please be kind enough to translate the below query?

Average(margin@team and season)>=3

pip2 · 12-12-14, 11:00 AM

NBA SDQL 'BEST BET' 2014-15

12/12/14 7:50am pst Reduced Basketball 803 Los Angeles Clippers/Washington Wizards Over 204½ -105*

SDQL:

playoffs=0 and season > 2005 and A and oA(dpa,N=3) > 3 and oA(dps,N=3) > 1 and tA(dps,N=3) > 1 and tA(dpa,N=3) > 3 and 3 > line > -8 and 197 < total < 208 and month=12

SDQL: English

Since 2005, 2 teams on a mini-streak (3) where they have both scored more than expected and allowed more points than expected, during the month of December where the game has a line and total similar to tonight's LAC/WSH game..

dmitean · 12-12-14, 12:07 PM

As for the offer to build a program that will give us queries results daily, I will chip in gladly.

Wanted to do the same, but few guys that promised me to build it, couldn't deliver and others asked really big fee.

To be honest, I got offers lower than 2,000$, but they from guys I don't know, so if this is someone that Nash knows and trusts...

dmitean · 12-12-14, 12:41 PM

Didn't check them deep - but here the queries that apply for today. Maybe someone can write it in to the file - just don't have the time today...

155 Under 203.5 in Washington
148 Atlanta -8
146 Under 206 in SA
143 SA -12.5
69 LAL +12.5
137 Boston -6.5
112 59 Under 205.5 in Minny
111 53 45 Under 204 in Phoenix
87 Over 204 in Phoenix
107 Clippers -2.5
92 Under 191.5 in Memphis
76 Pacers +8.5
76 Heat +1.5
75 Under 199.5 in Chicago
72 Suns -9.5
63 4 Portland +2.5

Heart · 12-12-14, 12:47 PM

I would like that open spot if it is still available.

Going through the queries with a fine tooth comb..

I think we should get rid of NBA16 in the spreadsheet. Add to the query 'and season!=2012' and it's only hitting 50%. 2012 was an anomoly where it went 35-17 or something making this one look better than it is.

dmitean · 12-12-14, 12:52 PM

I think that many queries are no good and I just ignore them. I think that we can delete about 20 queries easily...

Cutler'sThumb · 12-12-14, 01:12 PM

Originally posted by dmitean

I think that many queries are no good and I just ignore them. I think that we can delete about 20 queries easily...

I think this is a really important thing to do. With so many queries its important to keep refining where the long term quality is.
As I said previously, I'm also up for chipping in on the program.

Heart · 12-12-14, 01:24 PM

yup i'm seeing a lot of redundancy and based off of others with parameters that are irrelevant... which is why we have a lot of queries pointing to the same teams often.. because the queries are similar.. should probably scrub them at some point to get rid of the overhead.

edit: if we scrub the queries we'd have a lot less and it'd take a lot less time to go through. Should weigh whether or not it would then make sense to pay to automate the process

JMon · 12-12-14, 01:25 PM

Originally posted by dmitean

Yes they are. Since getting Durant and Westbrook back, they are 3 - 1 = 75%.

your query doesn't state that, WP is for the season. Didn't show up in sdql db as a play. Good win, but your play will not be documented as a sdql bet

JMon · 12-12-14, 02:12 PM

Originally posted by Heart

I would like that open spot if it is still available.

Going through the queries with a fine tooth comb..

I think we should get rid of NBA16 in the spreadsheet. Add to the query 'and season!=2012' and it's only hitting 50%. 2012 was an anomoly where it went 35-17 or something making this one look better than it is.

You are in Heart!

JMon · 12-12-14, 02:15 PM

NBA SDQL "Best Bet" 2014-15

Overall Group Record - 66-58-2 53.2% (+4.43) (max wagered -to win 1 unit)

Yesterday's recap 1-0 (+1.00)

Official Entries: Max 10 of 10:

1. JMon - 12-10 (+1.90)

2. pip2 - 12-10 (+1.38)

3. nash13 - 6-4 (+1.82)

4. Ronald S. - 9-10 (-1.65)

5. hyahya - 6-4-1 (+1.55)

6. Mako-SBR - 10-10 (-1.00)

7. dmitean - 0-2 (-2.04)

8. Consigliere - 8-4-1 (+3.77)

9. GolfAddict - 3-4 (-1.30)

10. Heart

Mako-SBR · 12-12-14, 02:23 PM

Originally posted by dmitean

I think that many queries are no good and I just ignore them. I think that we can delete about 20 queries easily...

Originally posted by Cutler'sThumb

I think this is a really important thing to do. With so many queries its important to keep refining where the long term quality is.
As I said previously, I'm also up for chipping in on the program.

Unfortunately, each person's version of what a winning query is, or is not, even despite say poor numbers pointing the way, is still mostly opinion and subjective to each person's own views.

I personally don't want any of the queries touched on the sheet please, we each can dismiss or disqualify queries we may feel aren't performing well on our own, thanks.

A better long-term solution is to determine a set of rating criteria that is equivalent to a grade, or in nash's case a "confidence" rating, that can be applied to each new query that is added to the DB, as well as previous queries already in the system.

The grading metrics are also unfortunately subject purely to opinion, in this case hopefully one based on a group consensus feeling of what makes a "winning" query versus a "losing query", but that is a better solution than simply wiping queries out of the DB that a few people may not like or in their own view is a "bad" query.

Mako-SBR · 12-12-14, 03:11 PM

NBA SDQL 'Best Bet' 2014-15

12-12-14 - 12:05p PST

Play: OKC (-9) MIN -110

SDQL:
HD and po:blocks>=10 and 4<=line<=10 and season>=2006

English:
Fading a home dog that had 10 or more shots blocked in their last game, against a line that reflects a tough opponent by getting between 4 and 10 points. Let's get it!!

nash13 · 12-12-14, 03:14 PM

i can tell you about my confidence rating system:
* i use a relatively active confidence system = that queries are compared for each sport separately. what works and good numbers are for NHL should not be the criteria for NBA and vice versa.
* i take several points depending on the type of the query and use percentile divisions for each of them. ATS and TOTALS have:
1. ROI
2. Total Profit
3. Last 3 Years Profit
4. Last Years Profit
5. Z-Score
6. Percentage of Years in Profit (7 out of 8 etc)

After that I calculate best performing third (percentile above 33%), least performing and middle performing. This gives each section a point from 3 for best up to 1 for least.
I sum the points up and divide them by 6. The scores give you a hint about how good a query is performing on several levels.

Mako-SBR · 12-12-14, 03:27 PM

Originally posted by nash13

i can tell you about my confidence rating system:
* i use a relatively active confidence system = that queries are compared for each sport separately. what works and good numbers are for NHL should not be the criteria for NBA and vice versa.
* i take several points depending on the type of the query and use percentile divisions for each of them. ATS and TOTALS have:
1. ROI
2. Total Profit
3. Last 3 Years Profit
4. Last Years Profit
5. Z-Score
6. Percentage of Years in Profit (7 out of 8 etc)

After that I calculate best performing third (percentile above 33%), least performing and middle performing. This gives each section a point from 3 for best up to 1 for least.
I sum the points up and divide them by 6. The scores give you a hint about how good a query is performing on several levels.

That is great nash, thanks for breaking it down.

When you added them to the DB I went into the MLB section to explore, because I never got a good handle on Baseball this past season in terms of feeling like I could develop or identify winning queries.

Just by doing a simple sort of the MLB sheet by nash's rating, high to low, then looking at each of the top results in detail, it was clear that he has some nice gems that I'll definitely be using next season.

That's what I prefer, and wouldn't then move to delete the lower confidence queries or something, because not only do the lower performers still offer an opportunity to learn more nuanced parameters within SDQL itself, they may still provide winning season-over-season results that I can tweak or change to produce a better rating down the road.

Heart · 12-12-14, 03:32 PM

NBA SDQL 'BEST BET' 2014-15

12/12/14 9:35pm Los Angeles Lakers +13 -108 vs San Antonio Spurs

SDQL:
AD and (o:rest=0 or rest>=1) and (ats streak)<0 and 0<(o:ats streak) and line>=12 and playoffs=0

Translation to english
An away dog on rest or opponent off of no rest who is on an against the spread losing streak and whose opponent is on an against the spread winning streak where the line is 12 or more in the regular season.

Ronald S. · 12-12-14, 03:41 PM

NBA Best Bet

12/12/14 3:30 PM

LA Clippers -2.5 -105 (@Washington) (5dimes)

HD and p:AW and p

oints < 105 and season >= 2011 and rest=o:rest
ATS: 16-40-1 (-3.28, 28.6%) avg line: 3.5

Will go with JMon's query today. Fade the home dog on equal rest with their opponent and coming off an away win where they scored fewer than 105 points.

nash13 · 12-12-14, 03:44 PM

Originally posted by Mako-SBR

That is great nash, thanks for breaking it down.

When you added them to the DB I went into the MLB section to explore, because I never got a good handle on Baseball this past season in terms of feeling like I could develop or identify winning queries.

Just by doing a simple sort of the MLB sheet by nash's rating, high to low, then looking at each of the top results in detail, it was clear that he has some nice gems that I'll definitely be using next season.

That's what I prefer, and wouldn't then move to delete the lower confidence queries or something, because not only do the lower performers still offer an opportunity to learn more nuanced parameters within SDQL itself, they may still provide winning season-over-season results that I can tweak or change to produce a better rating down the road.

and a rating of a system can vary from year to year. based on last years performance etc

JMon · 12-12-14, 04:25 PM

NBA Best Bet

12/12/14 3:20p CT

Play : PHX UN 204 -110
SDQL:

208 >= total >= 201 and tS(ou margin, N=5) >= 30 and 60 >= WP >52 and 2009 <= season and site = home and rest>0

English:

Since 2009, play under with a total of 201-208-a winning home team on rest that has went over the total by 30 or more in their last five games.

dmitean · 12-12-14, 05:10 PM

Have many thoughts on my own about how to rate queries, but I come from a bit different approach, since I add basketball factor in to it and not basing just on the numbers, but will write my thoughts later on.
I firmly believe, that if you play with SDQL long enough, you will easily find a query that will match any scenario. Give me 20 minutes and I will find 3 queries that in the same game will support Over, 3 queries that support Under, 3 queries that will support home team and three queries that will support road team.
The key in my eyes is to find query that reflects the best, situation in which both teams come to the match.

If one team is hot as fire and other is cold as ice and the query doesn't have some reference to streaks, it can't be considered in my eyes.

But I will elaborate more on it a bit later tonight or tomorrow at the latest.

As for deleting queries - it depends. If we will have software, that will do the job of going through queries for us, than it's a no problem./

If we will continue to do it manually and keeping in mind that queries database continues to get bigger, it's a problem.

I had no problem to do it with 60 queries, but now we have 170+ and counting.

What happens when we reach 200+? 250+?

dmitean · 12-12-14, 05:38 PM

NBA best bet
Play: Clippers -2.5 -105

SDQL: conference = Western and o:conference = Eastern and WP>60 and o:WP>60 and playoffs=0 and season>2010 and site=away and line<0 and line>-5
Very small sample, but larger one is also good:
conference = Western and o:conference = Eastern and WP>60 and o:WP>60 and playoffs=0 and season>2010 and site=away
including 9 - 0 ATS (the push is a bit laying, because I bet that game and it was -4.5 most of the evening) this season.

English: I continue to expoit the giant advantage that West has over the East. Back good Western team, playing on the road against good Eastern team.

Notes: Clippers offense was best in the league last season and after huge struggles in the beginning of the season, now reached new heights. I like Wizards a lot and I'm not sure that I would back Clippers against Wizards in playoff series, even if Clippers had HCA.
But, Clippers are one of the best regular season teams in the league this season and while they are in that great form on offense, it's better to back them, than fade them imho.

dmitean · 12-12-14, 09:36 PM

Originally posted by JMon

your query doesn't state that, WP is for the season. Didn't show up in sdql db as a play. Good win, but your play will not be documented as a sdql bet

Can you show me a link, where it states that it has to appear in sdql db?
If the play fits both the query and the logic behind it, who cares that it didn't appear on sdql db, because Thunder started the season without half their roster???
It makes zero sense...

dmitean · 12-12-14, 09:57 PM

You know what? Doesn't matter. I don't want to argue and waste my time over things like that. That's why I never do such kind of contest and so on... You end up wasting energy on things that really don't matter.
I will quit the best bet contest. Please find someone else, that has more time and effort to put in to this, besides making the plays.

pip2 · 12-12-14, 10:41 PM

Originally posted by pip2

NBA SDQL 'BEST BET' 2014-15

12/12/14 7:50am pst Reduced Basketball 803 Los Angeles Clippers/Washington Wizards Over 204½ -105*

SDQL:

playoffs=0 and season > 2005 and A and oA(dpa,N=3) > 3 and oA(dps,N=3) > 1 and tA(dps,N=3) > 1 and tA(dpa,N=3) > 3 and 3 > line > -8 and 197 < total < 208 and month=12

SDQL: English

Since 2005, 2 teams on a mini-streak (3) where they have both scored more than expected and allowed more points than expected, during the month of December where the game has a line and total similar to tonight's LAC/WSH game..

Heh! I forgot the n:rest>0 and on:rest>0 parameters!

JMon · 12-13-14, 08:46 AM

Originally posted by dmitean

You know what? Doesn't matter. I don't want to argue and waste my time over things like that. That's why I never do such kind of contest and so on... You end up wasting energy on things that really don't matter.
I will quit the best bet contest. Please find someone else, that has more time and effort to put in to this, besides making the plays.

Not sure why you think the Best Bet is a contest- it was never intended to be (hence why I never put anyone in order for most units!!!); rather it is for us to have fun and track NBA SDQL 'Best Best' 2014-15 for ourselves and on-lookers. You're right, I won't waste another moment trying to explain, for accurate tracking, why your play wasn't documented as a SDQL play for that day. geez,lol

.