-
Comment -
will transfer them later to the sheet. must crosscheck if anything is mentioned double.OK, this is what I have on the menu today....
from jmon sa cover vs lac ATS: 12-44-2 (-5.59, 21.4%) P:L and Po
oints >= 110 and op:FL and op:division = opo:division and 2009 <= season
zeeman21 ATS: 308-228-9 (0.71, 57.5%) showing atl covers vs ny site=away and line<0 and 0
from Ronald in sitpost #1404 ATS: 63-41-2 (2.69, 60.6%) sa covers vx lac -- works better 2nd half of season -- A and p:L and total>=200 and p:margin>-4 and opoints>100 and oppoints>100 and opppoints>100
ut covers vs ind from Ronald in sitpost #1389 ATS: 37-12-0 (3.71, 75.5%) AD and p:AD and rest = 0 and o:rest > 0 and month = 11 and line < 4
ny covers vs atl from Ronald in sitpost #1472 -- ATS:71-40-2 (2.42, 64.0%) HD and rest = 1 and p:rest = 0 and p:A and pp:A and season >= 2009Comment -
Not much of a sample, but against any logic. I tried this to prove the opposite and was stunned:
p:site=away and site=away and rest=0 and 0
On the other hand, Utah after upset win on the road, playing as a dog on the road again in B2B:
team=Jazz and p:site=away and site=away and rest=0 and p:line>0 and line>0 and 20121001<=date
2 - 9 ATS...
</p:line<=4>Last edited by dmitean; 11-10-14, 12:27 PM.Comment -
What you're seeing here is the basis of a few chase systems that people play (successfully most of the time) on SBR and other boards, where you're betting against streaks using a Martingale-esque amount of units.
Essentially once a team gets to X amount of wins in a row, the chase begins by fading that team, with the first betting unit. If the team wins again, fade with 2x the initial unit, wins again, 4x the initial unit, and if they win again, cease betting that particular team (to protect against a wipeout).
As you see in your results, 90% of the time the team does not win three in a row past whatever "start" point the system favors, and usually people do well with it. SBR member Grinder12000 runs an annual chase following a similar format for the NBA and it's done well over time.
Comment -
And that's the way it should be.
Team trends, and even worse, player trends, don't hold up over time, go obsolete the fastest, and have a much higher chance of never being real to begin with due to WAY too many variables.
Lurkers and people new to SDQL (FlyinAir): Don't fall for this nonsense from the various touts and info pages you see, they simply don't hold up. Google "correlation does not imply causation" for more information as to why something "looks" real, when it simply isn't.
Last edited by Mako-SBR; 11-10-14, 03:53 PM.Comment -
going off Ronald's, if you add AD and p:AD and rest = 0 and o:rest > 0 and month = 11 and line < 4 and p:W
the p:W, it becomes 12-1 ATS trend for utahComment -
@nash: how exactly does the spreadsheet work? if I come up with a new trend, should I just add it there or post it here too? Also sent you a request from email very similar to my username.Comment -
I am willing to contribute any chance I can get access to the spreadsheet?Comment -
I am willing to contribute as well. Like I said earlier I have been back tracking every single code here since Page 1, and I'm about half way done now. Any chance nash13 you and I can work together on this one?Comment -
No problem for me, just write me a personal message, I will give you the rights via Google docs. A simple message with your e-mail address is all I need.Comment -
Sdql
Hey nash13,
Chicagocover3@aol.com is my email.
Like I said, I've been backtracking all the way to post one (last season NBA) and everybody SDQL's. Since there are so many I didn't take the time to give credit to where it's due, I did create a list of everybody who contributed. I saw you were making a spreadsheet...is it like the same thing I'm doing? Or?
I'm more than willing bro, I stay up till 6 am handicapping everynight.
Thanks,Comment -
It should be nearly the same, but I have been selective, just add what you like. I will give you access in 5 mins. Please delete your mail adress after getting access, to avoid spamming.
The sheet tracks other sports as well, just get used to the format that's all.Comment -
Ronald and Golf, please write a PM. I will be away at night, but activate everyone Asap.Comment -
Results:OK, this is what I have on the menu today....
from jmon sa cover vs lac ATS: 12-44-2 (-5.59, 21.4%) P:L and Pooints >= 110 and op:FL and op:division = opo:division and 2009 <= season
zeeman21 ATS: 308-228-9 (0.71, 57.5%) showing atl covers vs ny site=away and line < 0 and 0 < p:margin and p:streak<=-2 and 19951030 <= date
from Ronald in sitpost #1404 ATS: 63-41-2 (2.69, 60.6%) sa covers vx lac -- works better 2nd half of season -- A and p:L and total>=200 and p:margin>-4 and opoints>100 and oppoints>100 and opppoints>100
ut covers vs ind from Ronald in sitpost #1389 ATS: 37-12-0 (3.71, 75.5%) AD and p:AD and rest = 0 and o:rest > 0 and month = 11 and line < 4
ny covers vs atl from Ronald in sitpost #1472 -- ATS:71-40-2 (2.42, 64.0%) HD and rest = 1 and p:rest = 0 and p:A and pp:A and season >= 2009
Jmon's query won the sa cover
Zeeman's query won the atl cover
Ronald's query won the sa cover
Ronald's query lost the ut cover
Ronald's query lost the ny cover
I bet the sa cover and the ut cover and went 1-1Comment -
Cashed.Worth the read..
Luis Scola is 23-5 life time when coming off a game scoring less than 6.5 points, then playing at home vs opp under 50 wp
Roy Hibbert is 20-8 life time when coming off a game scoring less than 6.5 points, then playing at home vs opp under 50 wp
Scola and Hibbert play for the same team
Roy Hibbert : p : points<6.5 and H and o:WP<50
TUESDAY:
ATS & OVER
p: points>88 and p: points<100 and p:margin<7.5 and AD and total > 191.5 and op:L and season>=2012 and total>200Last edited by FlyinAir; 11-11-14, 02:28 AM.Comment -
2 Wins - 8 Losses last night.
Is everyone on board? Anything to add or something you'd like to change?Comment -
I've noticed some of these queries go way back into the 1990s but the game must have changed immensely since then. I've been filtering to as far back as 2005 depending on the query but what would you guys recommend?Comment -
i think there could be added a level of trust to each query.
i do that by calculating the years of profit from 2006-14
three periods of trust: 1-3 years profit --> low level
4-6 years profit --> middle level
7-9 years - high level
and i will give a extra point if all of the last 3 years gone positive.Comment -
That sounds awesome! Maybe even an extra point if the last 10 or 20 instances have had a positive outcome?
I hope these scores could be calculated automatically though, because updating them manually would be kind of a chore, and if they aren't updated, the scores could become misleading.Comment -
BOLD are Overall NBA data. Not bold are from 2004 until now. We have small sample queries but very solid percetages during the years. This queries are good base to play around with them and could be additional help to more solid sdql.p : LHD and pp : WAF 46-32-0 (0.99, 59.0%) 79-59-2 (0.64, 57.2%)
p : LHF and pp : LHF and ppp : HF 40-21-2 (2.37, 65.6%) 80-49-4 (2.26, 62.0%)
p : LHF and pp : LHF and ppp : WD 15-34-0 (-4.32, 30.6%) 26-50-1 (-3.46, 34.2%)
p : LHF and pp : LHD and ppp : WF 19-10-0 (-0.36, 65.5%) 35-20-1 (2.67, 63.6%)
p : LHF and pp : LHD and ppp : LA 25-34-0 (-0.98, 42.4%) 47-66-0 (-2.04, 41.6%)
p : LAF and pp : WHF and ppp : WF 120-81-7 (0.74, 59.7%) 222-168-9 (0.79, 56.9%)
p : LAF and pp : LHD and ppp : H 7-13-1 (-1.05, 35.0%) 12-26-1 (-2.41, 31.6%)
p : LAD and pp : LAF and ppp : H 57-41-3 (2.84, 58.2%) 105-73-5 (2.74, 59.0%)
p : WAD and pp : WHD and ppp : WD 17-8-0 (3.24, 68.0%) 32-11-0 (3.60, 74.4%)
p : WAD and pp : LAF and ppp : F 38-28-1 (2.05, 57.6%) 81-56-1 (2.62, 59.1%)
p : WAD and pp : LHF and ppp : LA 40-26-1 (1.80, 60.6%) 69-49-3 (1.88, 58.5%)
p : WAD and pp : LHD and ppp : WD 18-30-2 (-1.45, 37.5%) 31-52-2 (-2.68, 37.3%)
p : WAF and pp : WHD and ppp : W 7-18-0 (-3.64, 28.0%) 16-32-0 (-3.09, 33.3%)
p : WAF and pp : LAD and ppp : WAF 15-27-0 (-4.73, 35.7%) 29-49-0 (-3.12, 37.2%)
Comment -
Only finding three query options for today, really only two:
from mako showing sac and cha cover vs dal and por -- ATS: 202-137-9 (1.48, 59.6%)
A and p:AFL and 1<= rest<=2 and season>=2005
From Eser Orhan (same as mako's) showing sac and cha cover vs dal and por -- ATS:154-94-4 (2.03, 62.1%)
A and p:AFL and p:line>=-4 and season>2007 154-94-4 (2.03, 62.1%)
from jmon -- okc, sac, and cha cover vs mil and dal and por -- ATS:14-36-0 (-5.22, 28.0%)
H and tA(FTA)<=21 and month=11 and F and -8.5<=line<=-3 and 2008<=season and rest<4 and o:rest<3 and losses>0 and o:wins<7 and o:losses<9Comment -
p:margin>=25 and line>p:line and season>=2011Comment -
5 plays added to Plays in the sheetComment -
Nash can you please add daveandmarie at gmail.com to access? Sent you an email and a request through google but still no access. ThanksLast edited by Consigliere; 11-11-14, 12:06 PM.Comment -
No play for today but:
F and p : margin>=15 and p : WA and pp : L and P : WA
Strong favorite who win with 15 or more (blowout) at road last game but lost previous one (here i eliminate really hot teams with good winning streaks) and win previous matchup at road get worse line. Here people are affected by favorite last blowout at road game and of course recieve bad line by the books cause it's more probable favorite in the game will be the public play.
As you see the favorite won more than 70% but covers in only 40% with 150 games behind that trend. It's not great finding but i see some logic behind this numbers. Almost every losing season in this trend is 50:50 mark (like 4-4 or 6-6 for example) so if the season is bad it didn't generate huge loss.... last season it's 1-9-0!107-42 (5.93, 71.8%) 58-86-5 (-1.63, 40.3%) avg line: -7.6 Comment -
Question for those out there much smarter than me....the Sac and Cha picks come from the trend A and p:AFL and 1<=rest<=2, so that is the situational angle for Sac and Cha coming in...do you guys go and look at that exact angle (site and previous result and rest filter) for the opposing team (ie Dal and Por) and see how that plays out for the same time frame? I did it, they were 50-50 scenarios essentially for both...so what that would tell me is that the angle on Sac/Cha is 60%, the same angle applied to Dal/Por is 50%, so the net angle is 55%...I'm sure the probability is not a simple average but wondering if you guys look at the angle from the opposites team perspective to do additional filtering?Comment -
We missed thisone yesterday:
H and 102<=tA(points)<=110 and 93<=oA(ooints)<=96 and total<=193 and playoffs=0
Comment -
yes i will do. sorry for the wait.Comment -
The way I frame the SDQL bettor's mission statement is that we are constantly in search of chronic mispricings in betting markets based on specific situations due to the lines makers model omitting certain information that the overall market does not seem to be picking up on (hence the chronic part). The fact that most situations do not show a bias either way points to them being modeled as an error term (providing randomness only) meaning they could equally help you or hurt you and therefore should be lumped in with all other error terms.
The underlying premise is that there is causation in the situations we uncover like "previous win was by 25 or more" which may point to an overreaction by the market to the actual strength of that team and provide value going the other way in their following game. Otherwise, you end up with situations that were exhibiting data mining which can be very costly if assumed to be causation instead.
Just my two cents.Comment
