NHL Situational Plays

[Cutler'sThumb]

Here's another of yours JMon, although I'm a little shy of it, as it's gone 11-2 so far this season and a regression to the mean seems likely:

H and ou streak>=2 and 2.9>=tA(goals)>=2.5 and 2.9>=oA(goals)>=2.5 and 2010<=season

[emceeaye]

HF and tA(o:goals)>=3 and po:goals<=2 and ppo:goals<=2 and season>=2010

and month=1 and day=Saturday together and separately.

[emceeaye]

Here's another of yours JMon, although I'm a little shy of it, as it's gone 11-2 so far this season and a regression to the mean seems likely:

H and ou streak>=2 and 2.9>=tA(goals)>=2.5 and 2.9>=oA(goals)>=2.5 and 2010<=season

H and ou streak>=2 and 2.9>=tA(goals)>=2.5 and 2.9>=oA(goals)>=2.5 and 2010<=season and month=1

Small sample size, but still not very compelling evidence in favor of under in January since 2010.

[emceeaye]

Take a look at this one supporting Wild ml tonight:

AD and streak + o:streak=4 and o:streak>1 and 3>streak>0 and pvertime=0 and (month = 11 or month < 4) and line>102

looks pretty good with a 56.2% ROI, right?

Just to illustrate the dangerousness of taking any query at face value...

look what happens when you add day=Saturday. Yes, the sample size decreases which clearly limits interpretability to some extent, but the results of this 17-game sample subset can't be ignored... down to an ROI of 2.8%.

[moshi]

Take a look at this one supporting Wild ml tonight:

AD and streak + o:streak=4 and o:streak>1 and 3>streak>0 and pvertime=0 and (month = 11 or month < 4) and line>102

looks pretty good with a 56.2% ROI, right?

Just to illustrate the dangerousness of taking any query at face value...

look what happens when you add day=Saturday. Yes, the sample size decreases which clearly limits interpretability to some extent, but the results of this 17-game sample subset can't be ignored... down to an ROI of 2.8%.

That really depends on how much weight you think "day=Saturday" holds. Note that ROI increases to +29.3% if you change it to "month=1".

[emceeaye]

That really depends on how much weight you think "day=Saturday" holds. Note that ROI increases to +29.3% if you change it to "month=1".

#1) The point is that it's critical not to remove the influence of Saturday for a home team. Hence, as shown by adding it, Saturday does have a lot of weight. Replacing day=Saturday with month=1 ignores the influence of Saturday for a home team.
#2) Good point that the influence of month=1 is also important. However, while ROI is still good at 29.3%, the ROI was practically cut in half from what it is for the original query.
#3) Now when you add day=Saturday and month=1 to the original query, although you have an extremely small sample size, it's interesting to note the WP favors the Stars and ROI is gone...Its just a little more evidence to support that day=Saturday may in fact trump the influence of January in this particular situation in terms of % influence of each of those variables. We will have to see sample size increase to determine that definitively. However results are suggestive and also dovetail intuitively with what one would think.
#4) when you compare the influence of the different months you see that March has the largest influence. When you add day=Saturday, ROI is around 30% bit sample size is still super low (WP 50%). I suspect that March might end up dampening the influence of Saturday more than January would.

[Cutler'sThumb]

Imho, sample size is probably the most influential factor when looking at this stuff. If I have a sample less than 100, I need to see a really impressive ROI/WP to take it seriously. Anything less than 50 better be walk on water good.

Here's NHL 36, which I feel is one of the best mixes of sample size and ROI and also favors the Wild tonight:

DA and total=5.5 and day!=Wednesday and day!=Friday and day!=Sunday and month<5 and (-3<streak<3 or streak=-5) and 5>o:streak>-3 and 4>p:margin>-4 and p:goals<6 and line<245 and pp:goals>0 and 4>P:margin>-2 and season = 2013

[emceeaye]

Avalanche ML

total=5.5 and p:faceoffs won=33 and day!=Thursday and day!=Tuesday and 4>streak>0 and P:margin>-4


p:goals=2 and op:goals=3 and H and streak<3 and o:streak<3 and day!=Friday and op:margin<3 and 4>rest - o:rest>-3 and rest - o:rest!=-1 and 2<opp:goals<4 and line<141


season>=2010 and HF and po:goals<=1 and opo:goals>=6

another one on avalanche and also the under...but with respect to the under, we have to see which goalie they put between pipes--one is solid, and the other not so good.
A and po:goals>=5 and op:margin=1 and opp:margin=1 and season>=2010

[JMon]

^..emc I believe there was a solid one favoring the Blue Jackets, but do not recall now.

[Cutler'sThumb]

Bruins -125

p:L and pp:L and (p:shots on goal)>(po:shots on goal) and (pp:shots on goal)>(ppo:shots on goal) and A

[Cutler'sThumb]

Stars +200

A and 12 > tA(penalty minutes) > 11 and month != 11 and month != 12 and game number >= 5

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

[emceeaye]

^..emc I believe there was a solid one favoring the Blue Jackets, but do not recall now.

Wow, Avalanche is my most solid play today. I didn't see that query favoring blue jackets but would love to if someone had it handy.

[emceeaye]

A and 12 > tA(penalty minutes) > 11 and month != 11 and month != 12 and game number >= 5

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

Yes saw these too. I saw some others I don't have handy on the Blackhawks though.

ok, found this Blackhawks query that gave me pause:

H and p:AL and op:HW and opp:HW and op:margin>=2 and opp:margin>=2

Here's one that at face value favors Blackhawks, but when you narrow it down by month and day, you get starkly conflicting results:

D and streak>3 and season>=2010
add day=Sunday and you see it favors Stars,
but add month=1 instead, and see it favors 'Hawks

[JMon]

Wow, Avalanche is my most solid play today. I didn't see that query favoring blue jackets but would love to if someone had it handy.

I see if I can't find it.

[JMon]

I see if I can't find it.

Well I just went through all mine and couldn't find it. I was sure I saw one this morning as I even have the Jackets written down as a possible play..any way bol.

[emceeaye]

Well I just went through all mine and couldn't find it. I was sure I saw one this morning as I even have the Jackets written down as a possible play..any way bol.

Well thanks for checking JMon. We have a little time until game time, so I have some time to check.

[Cutler'sThumb]

Yes saw these too. I saw some others I don't have handy on the Blackhawks though.

That's really the balancing act, imo: how to stay consistent with which trends to follow. Especially considering how many trends we have, it's easy to find contradictions. For now I'm consistently playing about 20 large/consistent long term trends along with a few really strong small trends sprinkled in. I'm sure there will be adaptations, but trying to subjectively figure out each day's trends wasn't working very well for me.

Obviously at -225, there a pretty good chance that the Blackhawks win.

[JMon]

Well thanks for checking JMon. We have a little time until game time, so I have some time to check.

Found it..reason why I couldn't find it b/c the line moved...

(oA(p:goals,N=2)) / (tA(p:shots on goal,N=2))<.09 and 154>line>145 and rest<3 and o:rest<2

[emceeaye]

That's really the balancing act, imo: how to stay consistent with which trends to follow. Especially considering how many trends we have, it's easy to find contradictions. For now I'm consistently playing about 20 large/consistent long term trends along with a few really strong small trends sprinkled in. I'm sure there will be adaptations, but trying to subjectively figure out each day's trends wasn't working very well for me.

Obviously at -225, there a pretty good chance that the Blackhawks win.

Yeah, that sounds like respectable approach. I especially like your idea of high volume betting (i.e., 20+ bets a day) using those high N, high ROI trends. In the long run, you will probably make a profit...I'm just not sure I have the patience for waiting that long. I think that one would definitely have to bet in high volume when using those types of trends predominantly, because as with a lot of basesloaded's trends like that, the ranges of values included for many of the parameters in each query are so wide (e.g., -4<=o:streak<=4), that unless you go into each one and narrow down the range to better match the teams playing that day, then a good number of them will be misses, especially when those queries in their original forms have poor WPs and favor dogs.

With respect to contradictions, regardless of how easy they are to come by, when I notice them, and unless they are an overwhelming minority relative to the number of queries favoring the other side, it's usually a no-play for me.

[emceeaye]

Found it..reason why I couldn't find it b/c the line moved...

(oA(p:goals,N=2)) / (tA(p:shots on goal,N=2))<.09 and 154>line>145 and rest<3 and o:rest<2

Ah yes, thanks JMon! Line really changed dramatically since yesterday night, moved around -20 since then.

update: wouldn't be surprised if line moved even more in these last few hours, because just announced McElhinney as goalie for Blue Jackets, who kinda sucks.

[emceeaye]

bol tonight, everyone!

[JMon]

AD and p:AW and p:division = po:division and op:HW and rest not in [2] and o:rest > 0

Very solid, but tough taking a large fav today, but they are 10-1 SU - not so good 1.5

Anyway solid overall, so wanted to share...cheers.

Boy I have cashed of this one!

[emceeaye]

Boy I have cashed of this one!

Nice, JMon, congrats! Don't know how I missed that one.

[Cutler'sThumb]

H and conference!=o:conference and total>5 and po:goals<3 and ppo:goals>1 and p:goals>0 and 0<pp:goals<5 and ppp:goals>0 and playoffs=0

Got the wrong version in earlier today. Had some of my back test parameters in there so it didn't show correctly.

[JMon]

Fade a fav revenging a loss and is off a loss of 4 or more

F and P:L and p:L and p:margin<=-4

Off a divisional loss

F and P:L and p:L and p:margin<=-4 and p:division=po:division

Make it same season revenge

F and P:L and p:L and p:margin<=-4 and p:division=po:division and P:season=season

[emceeaye]

month!=12 and month!=2 and month!=3 and tA(penalty minutes)<=.87*A(penalty minutes) and oA(penalty minutes)<=.87*A(penalty minutes) and season>=2009 and (day=Tuesday or day=Wednesday) and tA(penalties)<=.80*A(penalties) and oA(penalties)<=.80*A(penalties)

[emceeaye]

H and (day=Tuesday or day=Saturday) and conference =Western and o:conference=Eastern and conference!=o:conference and total>5 and po:goals<3 and ppo:goals>1 and p:goals > 0 and 0 < pp:goals < 5 and ppp:goals > 0 and playoffs=0

[JMon]

Fade a fav revenging a loss and is off a loss of 4 or more

F and P:L and p:L and p:margin<=-4

Off a divisional loss

F and P:L and p:L and p:margin<=-4 and p:division=po:division

Make it same season revenge

F and P:L and p:L and p:margin<=-4 and p:division=po:division and P:season=season

nice doggie came in...

[Cutler'sThumb]

Here's the more inclusive version of the trend emceeaye posted yesterday. Loses the super high hit rate, but gains significantly in sample size. Still very solid imo.

H and conference!=o:conference and total>5 and po:goals<3 and ppo:goals>1 and p:goals>0 and 0<pp:goals<5 and ppp:goals>0 and playoffs=0

[nash13]

Finished my analysis of 136 queries from the spreadsheet for this current season so far.
Just like NBA I will list them up with the same standards I use back there.

Overview:
All Picks with double plays and contradictions:
ALL: 1626-1379 (P:152.44 Y:5.07% SR:54.11%)
SU: 1133-1012 (P:100.39 Y:4.68% SR:52.82%)
TOT: 493-367 (P:52.05 Y:6.05% SR:52.05%)

Without Contradictions (Double Plays allowed)
ALL: 700-513 (P:146.39 Y:12.07% SR:57.71%)
SU: 313-256 (P:74.44 Y:13.08% SR:55.01%)
TOT: 387-257 (P:71.95 Y:11.17% SR:60.09%)

Single Plays only, Contradictions and Double Plays eliminated
ALL: 298-233 (P:46.44 Y:8.75% SR:56.12%)
SU: 123-106 (P:24.64 Y:10.78% SR:53.71%)
TOT: 175-127 (P:21.75 Y:7.2% SR:57.95%)

My Own System (Double Plays allowed, Contradictions disallowed, Only 3-Star Plays)
ALL: 374-282 (P:78.55 Y:11.97% SR:57.01%)
ATS: 166-150 (P:33.75 Y:10.68% SR:52.53%)
TOT: 208-132 (P:44.80 Y:13.18% SR:61.18%)

Profit by month:
Oct: 40.9
Nov: 22.4
Dec: 12.75
Jan: -0.26

[nash13]

if some queries are missing please put them in the google document and i will include the plays later

[emceeaye]

if some queries are missing please put them in the google document and i will include the plays later

OK, I'll go through what's in the google sheet and add any I have that aren't in there.

[perryhs]

po:shots on goal>p:shots on goal*2 and p:UD


po:shots on goal>p:shots on goal*2 and p:F

[perryhs]

D and streak>3 and p:shots on goal < op:shots on goal

or

D and streak>3 and p:shots on goal < po:shots on goal

[perryhs]

streak>=3 and p:margin+pp:margin+ppp:margin>=10