NHL Situational Plays

[emceeaye]

going with detroit on one of bases query's:

AD and 10>penalties>5 and streak<4 and p:goals!=3 and 48>p:shots on goal>15 and day!=Sunday and day!=Monday and tS(W, N=5)<4 and 6>o:streak>-3 and p:margin>-4 and 4.5

This is interesting...I am frequently confronted with the challenge of deciding between the results of a query with a nicely high sample size, relatively low WP, good ROI AND the results of a query with a significantly lower sample size, good WP, good ROI...I'm going to stick to my conservative approach with this that made me a good profit with MLB this year and make this a nonplay due to conflicting queries.

Thanks for posting Nash...didn't get to running bases' queries yet.

[otunj]

This is interesting...I am frequently confronted with the challenge of deciding between the results of a query with a nicely high sample size, relatively low WP, good ROI AND the results of a query with a significantly lower sample size, good WP, good ROI...I'm going to stick to my conservative approach with this that made me a good profit with MLB this year and make this a nonplay due to conflicting queries.

Thanks for posting Nash...didn't get to running bases' queries yet.

The most money you make are on the bets you avoid

[zxqwpdbg]

This is interesting...I am frequently confronted with the challenge of deciding between the results of a query with a nicely high sample size, relatively low WP, good ROI AND the results of a query with a significantly lower sample size, good WP, good ROI...I'm going to stick to my conservative approach with this that made me a good profit with MLB this year and make this a nonplay due to conflicting queries.

Thanks for posting Nash...didn't get to running bases' queries yet.

i posted this one. i did manage to copy and paste bases' stuff, for anyone interested.

there's also an option to disable smiley faces.

[zxqwpdbg]

it looks like the conference parameter is up. sportsdatabase also added division i didn't pay him for that though, i wonder why they put it up. just keep in mind that it is possibly in beta while he does the final adjustments. what are all the shortcuts for conference and division? like conference, non conference, can you select which division you want? things like that.

[otunj]

i posted this one. i did manage to copy and paste bases' stuff, for anyone interested.

there's also an option to disable smiley faces.

Maybe I'm an idiot, but where are bases' posts?

[escism]

it looks like the conference parameter is up. sportsdatabase also added division i didn't pay him for that though, i wonder why they put it up. just keep in mind that it is possibly in beta while he does the final adjustments. what are all the shortcuts for conference and division? like conference, non conference, can you select which division you want? things like that.

That's awesome! I'm not sure there shortcuts for conference or division. WHen looking at the Google group everyone always writes them out from I can tell.

[JMon]

i posted this one. i did manage to copy and paste bases' stuff, for anyone interested.

there's also an option to disable smiley faces.

id be interested. Before they were removed, I noticed they were loaded with smiley's.
did you happened to get them them fixed on your copy? Also how do you disable them before you post?

[escism]

Any home team with a total of 5.5 that’s playing in the month of January, and is coming off 2 or more home wins in a row, has seen 71 of those 105-games (67.6%) go under the total since the beginning of the 1996-1997 year.
(http://www.sportsbookreview.com/nhl-...onday-a-36667/)

H and total=5.5 and month=1 and site streak>=2

I think I did the site streak correct. But anyone edit if I messed that one up.

[zxqwpdbg]

Maybe I'm an idiot, but where are bases' posts?

bases' post was deleted. so you can't see it anymore.

id be interested. Before they were removed, I noticed they were loaded with smiley's.
did you happened to get them them fixed on your copy? Also how do you disable them before you post?

i will post all of bases' query's. i did not get a chance to go through the smiley's. some of his query's didn't look right either, like they had been clipped off or missing parts. not sure if it had anything to do with the copying and pasting he did or maybe was on a mobile app. either way i will copy and paste what i have. some might be missing or moved around but for the most part this is what it is.

you can turn off the smiley's by clicking on "reply to thread". then under your message in "additional options" there is a check box for "disable smilies in text". i seem to have to click it every time i make a new post though.

[zxqwpdbg]

bases' query's:

p:streak=-2 and p:margin<2 and site!=p:site and pp:margin!=-2 and ppp:margin>-6 and o:streak!=2 and o:streak!=1 and month!=5 and 6>site streak>-5 and n:rest<3 and on:rest<3

4 < po:goals < 7 and p:FL and p:goals>0 and o:streak<4 and rest<5 and rest!=2 and (day=Saturday or day=Sunday or day=Monday)

month = 11 and o:rest = 1 and rest = 1 and total<6 and p:rest<3 and op:rest<3 and H and streak>-4 and site streak<2 and 5>o:site streak>-4

pvertime>0 and p:L and A and day=Saturday and line<183 and streak>-4 and o:streak>-2 and op:margin>-3 and openalties!=6

total=5.5 and rest=1 and o:rest=1 and day!=Monday and p:total=5.5 and op:total<6 and po:goals<5 and opo:goals<5 and 34>p:shots on goal>25 and 34>op:shots on goal>19 and 3>streak>-2

HF and p:goals=2 and op:goals=3 and 4>pp:goals>1

season>2006 and total = 5.5 and day != Monday and 0 < p:goals < 6 and op:goals != 6 and 9 > penalties > 0 and 9 > openalties > 0 and penalties + openalties < 12 and 3>streak>-5 and 3>o:streak>-5 and power play goals + opower play goals>0 and (tS(ou margin, N=1)<.01 or tS(ou margin, N=1)>1) and (.5>opu margin>-3 or opu margin>1) and pp:goals<5 and 013 and rest + o:rest<5 and rest+o:rest!=3

total=5 and 1>streak>-5 and 1>o:streak>-5 and day!=Saturday and month!=4

penalties = 5 and p:shots on goal = 26 and day!=Saturday and day!=Monday

D and streak + o:streak=4 and 3>streak>0 and pvertime=0 and (month=11 or month<4) and line>102

AD and 0>streak>-3 and o:streak=-1 and 5>op:streak>0 and tS(W, N=3)=1

day = Tuesday and streak = 3 and total>5

total = 5.5 and p:goals + op:goals = 2 and p:goals - op:goals = 0 and H

A and streak=1 and o:streak=-1 and p:margin<5 and 3>P:margin>-5 and day!=Sunday and day!=Wednesday and rest + o:rest>0 and playoffs=0 and line<147

tS(W, N=3) = 3.0 and oS(W, N=3) = 0.0 and day!=Saturday and P:margin>-2 and rest - o:rest!=-1

total = 5.5 and H and ou streak = 1 and (2>streak>-3 or streak=4) and o:streak<3 and 4>site streak>-2 and day!=Wednesday and day!=Monday and day!=Friday

total = 5.5 and p:goals + op:goals = 6 and p:goals - op:goals = 0 and day!=Monday and day!=Wednesday

F and streak=-1 and o:streak<0 and day!=Monday and (H or rest<2) and (H or o:rest<2) and site streak>-3 and o:site streak!=2 and p:margin>-5 and P:margin>-4 and op:goals<4 and rest + o:rest>0 and game number<75

total = 5.5 and 14 < p:shots on goal < 32 and playoffs = 0 and day != Thursday and day != Friday and day != Sunday and rest < 4 and o:rest < 4 and rest + o:rest > 0 and 7 > streak > -4

total = 5.5 and 14 < p:shots on goal < 32 and playoffs = 0 and day != Thursday and day != Friday and day != Sunday and rest < 4 and o:rest < 4 and rest + o:rest > 0 and 7 > streak > -4 and (game number < 38 or game number > 51) and ou streak < 5 and pu margin > -3 and 4.5 > opu margin > -3 and 43 > op:shots on goal > 19

total=5 and 1>streak>-5 and 1>o:streak>-5 and day!=Saturday and month!=4

season>2009 and 0>power play goals - opower play goals>-3 and day!=Tuesday and -4-6 and month!=2 and month!=5 and month!=6 and (A or day!=Sunday) and (215>line>-212 or line>249) and penalties<10 and openalties>1 and p:shots on goal - op:shots on goal<18 and op:shots on goal<43 and p:assists!=5 and 14>op:assists>1 and p:shifts>326 and 83>op:shifts -p:shifts>-75 and po:goals>0 and rest + o:rest<7

p:streak>2 and p:margin<3 and site!=p:site and rest<3 and site streak>-5 and o:site streak<5 and p:rest<3 and pp:rest<6 and op:margin!=-3 and n:rest<4 and opo:goals>0 and (line<-275 or line>-219)

rest = o:rest and AD and rest = 1 and (P:AD or P:HF) and WP <= 60 and (130 >= line > 115 or line > 150) and P:rest - oP:rest <= 1 and 5>streak>-7 and o:streak>-5 and p:rest<5 and day!=Tuesday and 6>o:site streak>-3 and on:rest<3

A and 12 > tA(penalty minutes) > 11 and month < 7 and streak != -2 and o:streak != -3 and day != Saturday and 4 > site streak > -4 and rest < 8 and 9 > oS(W, N=10) > 1 and on:rest < 4 and line < 220

AD and 10>penalties>5 and streak<4 and p:goals!=3 and 48>p:shots on goal>15 and day!=Sunday and day!=Monday and tS(W, N=5)<4 and 6>o:streak>-3 and p:margin>-4 and 4.5

total=5.5 and p:faceoffs won=33 and day!=Thursday and day!=Tuesday and 7>streak>-2 and P:margin>-4

[otunj]

just got an email from sportsdatabase. they said they are gonna try to have a beta version of it up today.

division parameters:

http://sportsdatabase.com/nhl/query?output=default&sdql=division+and+H&submit=++S+D+Q+L+!++

conference parameters:

http://sportsdatabase.com/nhl/query?output=default&sdql=conference+and+H&submit=++S+D+Q+L+!++

some divisions only have 2 seasons inputed, but looks like a work in progress which is ok

To look at games that are not in the team's conference, you want to do:
o:conference!=conference

Same thing with division
o:division!=division

[zxqwpdbg]

division parameters:

http://sportsdatabase.com/nhl/query?output=default&sdql=division+and+H&submit=++S+D+Q+L+!++

conference parameters:

http://sportsdatabase.com/nhl/query?output=default&sdql=conference+and+H&submit=++S+D+Q+L+!++

some divisions only have 2 seasons inputed, but looks like a work in progress which is ok

To look at games that are not in the team's conference, you want to do:
o:conference!=conference

Same thing with division
o:division!=division

the reason is because there was a realignment in 2013. before 2013 there were 6 divisions and after 2013 only 4. there was also some conference realigning. if i am not mistaken detroit and columbus went from the west to the east and winnipeg went from the east to the west.

[zxqwpdbg]

my first conference query:

po:goals = 0 and op:goals = 0 and playoffs = 0 and p:goals >= opo:goals and o:conference = Western

[otunj]

Back Fitting

When I started doing queries, one of the first posts I read on the sportsdatabase group was from "weather" regarding back fitting queries. I've tried to adhere to his advice in my queries ever since. I hope you all enjoy it. I copied it below:

I need to speak out in terms of experience here before someone gets hurt.

The system included here has little liklihood of future success, and here is why. My experience, which is plenty, has shown me that when you backfit a situation with so many conditions (at least12 here), your no longer predictive in terms of the future, but creating the most perfect flow chart of the past.


Most of these systems are built on finding something with a modest ROI, and then experimenting with variables, meaningful, or meaningless, adding only those that make the situation appear better. If you experiment with a lot of variables, then you will "stumble" into those that make a certain situation look better, but in the process, you "contaminate" future predictability.


What you are left with is thinking and believing you have found the holy grail of sports betting, only to be fooled by a false premise of profitability. I gurantee you over the next 100 games the win rate is closer to 0% ROI than it is to what the backfitted data suggests it will do.


We are putting faith in a system that is built upon 12-15 conditions, many of which are used not out of known predictive advantages, but out of anything that improves the win rate. By definition you will find something that is better than the 1 condition variable.


Thing of this. You flipped a coin 200 times, and without adding any filters your results were 100 heads and 100 tails, no predictability of future flips.


You video taped every flip. Now you go back and analyse what happoened. You noticed when you flipoed with your left hand, heads came up 55 times out of 100, and the 100 flips with your left hand came up just 45.


So you add the variable, if flipped with left hand, 55% of the time you get heads!


Further looks see that if you placed the coin in your left hand from your right hand you got 30 heads and 20 tailes, but when you picked the coin up instead without any use of your left hand, you got 25 heads and 25 tails. So your now up to a way to get 60% heads!


next you noticed if you paused for more than 10 seconds after placing the coin in your left hand from your right, you got 17 heads and just 8 tails. You now have a situation that generates 68% heads, just by flipping the coin from your left hand after placing it there from your right hand, and waiting at least 10 seconds before you flipped it.


So my question is:


If you then flipped the coin 300 more times, doing everything the way you did to get 68%, what percentage of heads is expected from the 300 flips?

The answer is 50%!


See the variables used above were selected not based on anything that is predictive, but based on anything that made the system look better! THAT IS THE PROBLEM!

The heads scenerio here was built the same way, none of the added conditions are predictive in any way!


The moral of the story is this:


Build a concept on known +EV variables, not a system built to make +EV concepts.
What do I mean by meaningful variables? When you do a search in a given sport with 1 variable, and it shows an advantage, then you have found a meaningful variable.

Build a pile of meaningful variables for a given sport, then stack the meaningfu variables, to get meaningful situations.


Hope that helps everyone here.

[zxqwpdbg]

perfect avs query:

team=Avalanche and season>=2013 and p:HL

[escism]

perfect avs query:

team=Avalanche and season>=2013 and p:HL

I've got a couple of other queries from Bases up top that are pointing to the Avs today.

[zxqwpdbg]

I've got a couple of other queries from Bases up top that are pointing to the Avs today.

ya i saw that. i put 2 units on them tonight.

[otunj]

I've got a couple of other queries from Bases up top that are pointing to the Avs today.

Also pointing to Avs

WP<45 and o:WP>45 and op:L and month=10

[escism]

Also pointing to Avs

WP<45 and o:WP>45 and op:L and month=10

Yep that's one. I can't remember the others. Lets pick up some W's tonight!

[nash13]

When I started doing queries, one of the first posts I read on the sportsdatabase group was from "weather" regarding back fitting queries. I've tried to adhere to his advice in my queries ever since. I hope you all enjoy it. I copied it below:

I need to speak out in terms of experience here before someone gets hurt.

The system included here has little liklihood of future success, and here is why. My experience, which is plenty, has shown me that when you backfit a situation with so many conditions (at least12 here), your no longer predictive in terms of the future, but creating the most perfect flow chart of the past.


Most of these systems are built on finding something with a modest ROI, and then experimenting with variables, meaningful, or meaningless, adding only those that make the situation appear better. If you experiment with a lot of variables, then you will "stumble" into those that make a certain situation look better, but in the process, you "contaminate" future predictability.


What you are left with is thinking and believing you have found the holy grail of sports betting, only to be fooled by a false premise of profitability. I gurantee you over the next 100 games the win rate is closer to 0% ROI than it is to what the backfitted data suggests it will do.


We are putting faith in a system that is built upon 12-15 conditions, many of which are used not out of known predictive advantages, but out of anything that improves the win rate. By definition you will find something that is better than the 1 condition variable.


Thing of this. You flipped a coin 200 times, and without adding any filters your results were 100 heads and 100 tails, no predictability of future flips.


You video taped every flip. Now you go back and analyse what happoened. You noticed when you flipoed with your left hand, heads came up 55 times out of 100, and the 100 flips with your left hand came up just 45.


So you add the variable, if flipped with left hand, 55% of the time you get heads!


Further looks see that if you placed the coin in your left hand from your right hand you got 30 heads and 20 tailes, but when you picked the coin up instead without any use of your left hand, you got 25 heads and 25 tails. So your now up to a way to get 60% heads!


next you noticed if you paused for more than 10 seconds after placing the coin in your left hand from your right, you got 17 heads and just 8 tails. You now have a situation that generates 68% heads, just by flipping the coin from your left hand after placing it there from your right hand, and waiting at least 10 seconds before you flipped it.


So my question is:


If you then flipped the coin 300 more times, doing everything the way you did to get 68%, what percentage of heads is expected from the 300 flips?

The answer is 50%!


See the variables used above were selected not based on anything that is predictive, but based on anything that made the system look better! THAT IS THE PROBLEM!

The heads scenerio here was built the same way, none of the added conditions are predictive in any way!


The moral of the story is this:


Build a concept on known +EV variables, not a system built to make +EV concepts.
What do I mean by meaningful variables? When you do a search in a given sport with 1 variable, and it shows an advantage, then you have found a meaningful variable.

Build a pile of meaningful variables for a given sport, then stack the meaningfu variables, to get meaningful situations.


Hope that helps everyone here.

Great post, regarding overtweaking and overfitting a system, it is very important to adjust logical reason behind that.
I only choose systems and trends which i can understand in terms of sports psychology. Revenge Systems esp in the NBA are quite good for that.

in addition:
it is quite as important to know how much your are staking. only the optimal stake size in combination with optimal system works. if you risk to few units, you are wasting profits away, if you risk too much, it is the same.

[escism]

Pretty pumped went 5-2 last night with my systems. The only losses were the Sabres, I had two systems pointing to them, and they damn near won. Then the other loss was the Canadiens, which was an OT. Even hit a 3 team parlay. So overall a very profitable night! Yay for situational plays!

[nash13]

I have to add, i usually take ML bets nor in Regular Time after 60 mins. This changes a lot. Because the trends in queries are base on ML/SU

[escism]

10/31/14 Live PLay

This is a live play tonight w/ 2 games:

season>=2009 and H and o:rest=0 and op:goals=0

Got it from Ross Benjamin's pick today. Not overly confident with the trend, but it's worth monitoring.

[escism]

season>=2009 and AD and rest=0 and p:goals=0 and op:goals>=2

19-41 (-1.03, 31.7%)	avg line: 153.4 / -167.5	on / against: -$1,310 / +$1,056	ROI: -21.8% / +10.4%
20-36-4 (-0.23, 35.7%)	avg total: 5.4

Really like this one, it's got two on it tonight. Interesting, 0 rest after getting shut out the night before. They don't seem to bounce back the next night. It's hard to think that two of the top teams in the league (Ducks & Kings) won't be able to bounce back, but according to this it doesn't look good.

[otunj]

This is a live play tonight w/ 2 games:

season>=2009 and H and o:rest=0 and op:goals=0

Got it from Ross Benjamin's pick today. Not overly confident with the trend, but it's worth monitoring.

+1.8 units in 4 seasons? I'm not confident at all.

[otunj]

season>=2009 and AD and rest=0 and p:goals=0 and op:goals>=2

19-41 (-1.03, 31.7%)	avg line: 153.4 / -167.5	on / against: -$1,310 / +$1,056	ROI: -21.8% / +10.4%
20-36-4 (-0.23, 35.7%)	avg total: 5.4

Really like this one, it's got two on it tonight. Interesting, 0 rest after getting shut out the night before. They don't seem to bounce back the next night. It's hard to think that two of the top teams in the league (Ducks & Kings) won't be able to bounce back, but according to this it doesn't look good.

Better

[otunj]

Pretty pumped went 5-2 last night with my systems. The only losses were the Sabres, I had two systems pointing to them, and they damn near won. Then the other loss was the Canadiens, which was an OT. Even hit a 3 team parlay. So overall a very profitable night! Yay for situational plays!

went 4-4 (-.05 units) with 4 of them going to OT/SO (2-2 record). I was feeling the rush last night!!

[emceeaye]

season>=2009 and AD and rest=0 and p:goals=0 and op:goals>=2

19-41 (-1.03, 31.7%)	avg line: 153.4 / -167.5	on / against: -$1,310 / +$1,056	ROI: -21.8% / +10.4%
20-36-4 (-0.23, 35.7%)	avg total: 5.4

Really like this one, it's got two on it tonight. Interesting, 0 rest after getting shut out the night before. They don't seem to bounce back the next night. It's hard to think that two of the top teams in the league (Ducks & Kings) won't be able to bounce back, but according to this it doesn't look good.

It does look good, but be careful, due to this conflicting query from bases with a much larger sample size...not saying redwinga won't win but probably worth a 2nd look:

season>2009 and 0>power play goals - opower play goals>-3 and day!=Tuesday and -4-6 and month!=2 and month!=5 and month!=6 and (A or day!=Sunday) and (215>line>-212 or line>249) and penalties<10 and openalties>1 and p:shots on goal - op:shots on goal<18 and op:shots on goal<43 and p:assists!=5 and 14>op:assists>1 and p:shifts>326 and 83>op:shifts -p:shifts>-75 and po:goals>0 and rest + o:rest<7

However even though small sample size, look at results with both queries combined....favors rewings...but verrry small sample size...

season > 2009 and season >= 2009 and AD and rest = 0 and p:goals = 0 and op:goals >= 2 and 0 > power play goals - opower play goals > -3 and day != Tuesday and -4 -6 = -10 and month != 2 and month != 5 and month != 6 and (A or day != Sunday) and (215 > line > -212 or line > 249) and penalties < 10 and openalties > 1 and p:shots on goal - op:shots on goal < 18 and op:shots on goal < 43 and p:assists != 5 and 14 > op:assists > 1 and p:shifts > 326 and 83 > op:shifts - p:shifts > -75 and po:goals > 0 and rest + o:rest < 7

[JMon]

season>=2009 and AD and rest=0 and p:goals=0 and op:goals>=2

19-41 (-1.03, 31.7%)	avg line: 153.4 / -167.5	on / against: -$1,310 / +$1,056	ROI: -21.8% / +10.4%
20-36-4 (-0.23, 35.7%)	avg total: 5.4

Really like this one, it's got two on it tonight. Interesting, 0 rest after getting shut out the night before. They don't seem to bounce back the next night. It's hard to think that two of the top teams in the league (Ducks & Kings) won't be able to bounce back, but according to this it doesn't look good.

Eliminate rest on the oppt. o:rest<2

[emceeaye]

Here's one based on losing streak=-4 in January

Here's one from Ross Benjamin from 1/29/14 whose logic escapes me. Does anyone know why any team playing on the road on a 4-game losing streak (no more and no less) would see their current game go over the total ONLY in the month of January since 2006?

A and month=1 and streak=-4 and season>=2006

[emceeaye]

Ross Benjamin situation (1/27/14) tweaked

Here's SDQL from a Ross Benjamin situation with a small tweak. Sample size is at least not prohibitively small. Only losing season since 2009 was 2013, but only by a slight bit---perhaps, that fact makes it a deal-breaker for some folks?

F and oA(goals)>=2.76 and op:goals+opo:goals>=7 and opp:goals+oppo:goals>=7 and oppp:goals+opppo:goals>=7 and season>=2009

Thoughts?

[emceeaye]

Ross Benjamin situation (1/24/14) tweaked

HF and oA(o:goals)<=2.52 and opo:goals>=3 and oppo:goals>=3 and opppo:goals>=3 and season>=2007

[emceeaye]

goal per game differential

Any money line home favorite of between -120 to -170, that’s playing in the 2^nd half of the season, versus an opponent that has a goal per game differential of +0.3 or better, and they allowed 3-goals or more in each of their last 3-games, has gone 26-4 (86.7%) since the beginning of the 2009-2010 year.

Can anyone explain what "goal per game differential of +0.3 or better" means here? Is it saying that the average difference in goals scored per successive game at least 0.3? if so, does the following express it?

oA(pp:goals-p:goals@team and season)-tA(pp:goals-p:goals@team and season)>=0.3

I keep getting an error.

[escism]

Eliminate rest on the oppt. o:rest<2

Thanks Jmon! Still getting the hang of it. Been trying to read all of your 4k+ posts, take A LOT of time!

[escism]

+1.8 units in 4 seasons? I'm not confident at all.

Yeah I agree with on it. I wasn't overly confident. I need to do some tweaking to it and see how I can get it more in the favorable side.