[Pokerjoe]

good-stuff....
all-you-mention-is-injurys-though...
there's-10+-other-factors-to-look-at...
okay-here's-one;;;;road-record

okay...good-luck.

Yes, there are many factors to look for. I have not "only mentioned injuries." Travel, HCA are important, too. But injuries are more important in basketball than any other sport, and proper analysis of them is a source of edge. Consider the points I made about Atlanta's game against NJ being a sign that they're back in full gear after losing high-minutes guard Joe Johnson. That's a source of edge.

As to most other factors, I sum them in my PRs. Or, try. That's the challenge.

Road record? I've done studies and find no predictive value in it at all. At any given moment, by normal distribution, you would expect some teams to have great road records, some terrible, and the home records as well, and it doesn't mean anything other than that they got a few more or less of their wins or losses at home or on the road, just as a distribution, not as a predictive variable.

The default position in any sports betting observation should be that an outlier is just random distribution. The default mistake of most sports bettors, though, is to do the opposite.

But you're welcome to consider it in your handicapping. I always remember that I might be wrong. Hell, I was wrong to have thought there to be no HCA when LAL plays LAC, when actually there is a bit of one, apparently.

And even better: show us some work in which you establish, not as an old wive's tale, but in fact, that road record has predictive value ATS. We'd all like to see it.

I well remember a debate at a card game once where some guy got hysterically angry. He insisted that the secret to success in NFL betting was in fumbles, he'd seen soooo many teams lose ATS because of fumbles, he'd seen a study which proved that the team that fumbles the least covers the most, etc.

I said I'd much rather bet on the teams that have been fumbling the most, and against those who have fumbled the least. He went crazy. I thought he might short circuit his pace maker. He called me an idiot. I am an idiot sometimes, but not in that instance. It might not any longer be true that betting on the fumbling teams is stone cold profitable in the NFL (I haven't checked the past few years), but for a long time it was.

And I'm sure no one here fails to understand that "teams that fumble the most cover the least" is not contradicted by "teams that have fumbled the most cover the most."

[Pokerjoe]

Talking about a top 5 team beating a bottom 5 team as "getting the monkey off their back" is much more Players Talk IMO. And I do believe in motivation, but the carry over from jacked up game to run of the mill game is limited.

They had a players only meeting before they smoked Washington, what is that indicative of to you? That meeting must of really helped them get focused and motivated, or ... it could be they were playing Washington.

All in who's telling the story, sorry if you can't see that. If you were to say they've worked through the newness period that woulda been logical, but to think their early season struggles against teams over .500 was because of the impending Cleveland game is ... silly

I still don't understand why you think playing bad teams helps good ones cover the spread. It increase their chance of winning games, but that isn't the issue. Covering the spread is the issue. Is sounds like you're saying to just bet on the best teams against the worst teams. ATS? That's what squares do, almost by definition: they bet on the best teams, without regard for the spread. But the spread--this is what spreads do--negates a good team's advantage. So tell us how being good is helpful in beating the spread? I don't get it.

Neither MIA or CLE's personnel changed in their game of Dec 2. Yet their perf ATS since has been dramatic, both of them. And it is so illogical to think that their perf ATS is because of SOS, not change in attitude, that I'm not going back and forth on it with you. It is just so fundamental that spreads negate advantage that I can't debate this further.

[Pokerjoe]

BTW, guys, this has been fun, but I don't have time to keep it up. Thanks for the ride.

Justin, do I get some points for pissing off Data? I think he took his ball and went home.

[Flying Dutchman]

"You're wrong, but I'm not saying how" because now you pissed me off.

Never mind.

Data, are you or are you not the guy behind AdvancedNFLStats?:

http://www.advancednflstats.com/

If you are, then back the fvk off everyone, as it is my all-time favorite sports site, bar-none with some seriously decent statistics to boot.

If you are not, then I see everyone's point on you being a non-contributing sour-pus.

[Data]

If you are not,

I am not.

then I see everyone's point on you being a non-contributing sour-pus.

I mostly post for smart people and am at peace with the fact that for the majority of posters here my posts fall into deaf ears. I would even go farther and say that this is an indicator that I am doing a better job for my target audience because the posters who are liked by the masses are very boring for the thinking people.

[Flying Dutchman]

I am not.

I mostly post for smart people and am at peace with the fact that for the majority of posters here my posts fall into deaf ears. I would even go farther and say that this is an indicator that I am doing a better job for my target audience because the posters who are liked by the masses are very boring for the thinking people.

OK folks, now he's calling us stupid because you certainly ain't posting for us.

You're now reaching for the ranks of Monkey-wacker & Math-dot-chump...

...OK on second thought, maybe not quite that bad...

[jgilmartin]

non-contributing sour-pus

Non-contributing in this thread, maybe.
Overall, no.

[Flying Dutchman]

Non-contributing in this thread, maybe.
Overall, no.

JG stands for Lieutenant-Junior Grade? Isn't that another one of the lower ranks on the Enterprise? Certainly below "Lieutenant-Commander Data."

Actually, I'm really bumping the thread as I see it as a fine piece of work by Poker Joe. Frankly I don't post stuff as extensive as this as I see it a not worth the time, but I certainly want to let Poker know I appreciate it FWIW.

By the way Poker, you were mentioned in that food-fight over in the variance thread. "Data" is showing his ass-pirations over there too.

[Meestermike]

Pokerjoe I admire your tenacity and your knowledgeable insights. From post # 23 I thought I would back up your point about CLE being so terrible both before the Miami pasting and after. The difference is quite profound.

The first set of numbers is their point differential of median scores of their last 4 games before Miami and the second is point differential of median scores of their last 4 games from Miami onward to date...

-9.00 
-8.00 
1.00 
-1.00 
1.50 
-7.00 
-4.00 
-3.00 
-7.00 
-7.50 
-13.50 
-15.00 
-2.50 
-6.50

 
-17.50 
-21.00 
-16.50 
-19.50
-20.50 
-20.00 
-14.50

[classhandicapper]

I'm a 45 year basketball fan that got more serious about understanding the game, advanced statistics, and valuing players over the last few years. I have a background of playing horses for about 35 years and am a winning player at that sport (modest amounts). So I know a lot about gambling and finding value. I am starting to play basketball now (very small bets to start).

Right now my model is mostly based on point differential adjusted for quality of competition, home/away, days rest or lack thereof, and injuries. I also note whether the team has been playing better or worse lately without actually building it into the line. I use that as a kind + or - to give more or less weight to a potential bet.

I am currently considering adding won/loss record to the model.

Since I'm a beginner, I am curious if the "core" of most people's line is point differential adjusted for quality of competition, record, or other factors that then get adjusted?

So far I am hitting 50% of my bets (21 of 42). That would not be very encouraging except that a few times I gave up an extra point or two in order to get greater than even money odds on a betting exchange. So 50% is actually an out performance. In fact, tonight I gave 9 1/2 on the Pacer game (an extra point) and lost by 1/2 point because of it. I also lost 1 bet because I made a math error DOH! and 1 because I missed a major injury.

One of the other things I've learned is that a lot of the lines that I disagree with are in games where one of the best teams is playing one of the worst teams. My lines are always a lot larger than the official lines. I assume that's the case because of garbage time and stars getting pulled early causing some of those games to not be as one sided as they could be if the better team went all out. Is that true?

In any event, am I in the ballpark with what most people are doing to create their lines?



Any suggestions?

[evo34]

"You're wrong, but I'm not saying how" because now you pissed me off.



Never mind.

"You're wrong, but I'm not saying how" because now you pissed me off. Never mind.

Go back to the comic book store that employs you... There is generally a reason certain people tirelessly cling to arguing semantics. It's because they have little else to offer.

[Pokerjoe]

I'm a 45 year basketball fan that got more serious about understanding the game, advanced statistics, and valuing players over the last few years. I have a background of playing horses for about 35 years and am a winning player at that sport (modest amounts). So I know a lot about gambling and finding value. I am starting to play basketball now (very small bets to start).

Right now my model is mostly based on point differential adjusted for quality of competition, home/away, days rest or lack thereof, and injuries. I also note whether the team has been playing better or worse lately without actually building it into the line. I use that as a kind + or - to give more or less weight to a potential bet.

I am currently considering adding won/loss record to the model.

Since I'm a beginner, I am curious if the "core" of most people's line is point differential adjusted for quality of competition, record, or other factors that then get adjusted?

So far I am hitting 50% of my bets (21 of 42). That would not be very encouraging except that a few times I gave up an extra point or two in order to get greater than even money odds on a betting exchange. So 50% is actually an out performance. In fact, tonight I gave 9 1/2 on the Pacer game (an extra point) and lost by 1/2 point because of it. I also lost 1 bet because I made a math error DOH! and 1 because I missed a major injury.

One of the other things I've learned is that a lot of the lines that I disagree with are in games where one of the best teams is playing one of the worst teams. My lines are always a lot larger than the official lines. I assume that's the case because of garbage time and stars getting pulled early causing some of those games to not be as one sided as they could be if the better team went all out. Is that true?

In any event, am I in the ballpark with what most people are doing to create their lines?



Any suggestions?

In the ballpark? I suppose, but that's such a vague term. Everyone probably starts with scores, and then realizes that scores only leads toward a variation of Sagarin, which isn't useful. And also the statistical difference between wins/losses and score margin is illusory, in that in the long run you'll end up in the same place. But margins are more accurate in the short run. A team that plays opponents A, B and C and wins by 10 pts each is better than a team playing that schedule and winning by 2 pts each. But scores and wins, I think no matter how you manipulate them, lead only to vanilla models, because they're the most vanilla stats. The edges are found in the boxscore stats, imo.

I'd be bothered by the bias toward large faves you say your model has. No bias is better than bias, but dog bias is probably a little better than fave bias. A bias toward large faves, to me, is a really bad sign.

As to your noting recency, but not directly incorporating it, except for +/-, I have to ask why such a vague inclusion? Recency is a very tricky but also very key aspect of modeling. Ha, I'm using "recency" with a horseplayer. Okay, I mean, how a team has been doing lately.

Good luck, have fun. I imagine you, like many of us, enjoy the work for its own sake.

[Maverick22]

what is the 'sect'?

[ClaudioNaz]

I'm in many years with sports, and totally agree with you that injuries are very very important in basketball, it would change the results of matches !

Yes, there are many factors to look for. I have not "only mentioned injuries." Travel, HCA are important, too. But injuries are more important in basketball than any other sport, and proper analysis of them is a source of edge. Consider the points I made about Atlanta's game against NJ being a sign that they're back in full gear after losing high-minutes guard Joe Johnson. That's a source of edge.

As to most other factors, I sum them in my PRs. Or, try. That's the challenge.

Road record? I've done studies and find no predictive value in it at all. At any given moment, by normal distribution, you would expect some teams to have great road records, some terrible, and the home records as well, and it doesn't mean anything other than that they got a few more or less of their wins or losses at home or on the road, just as a distribution, not as a predictive variable.

The default position in any sports betting observation should be that an outlier is just random distribution. The default mistake of most sports bettors, though, is to do the opposite.

But you're welcome to consider it in your handicapping. I always remember that I might be wrong. Hell, I was wrong to have thought there to be no HCA when LAL plays LAC, when actually there is a bit of one, apparently sports arbitrage

And even better: show us some work in which you establish, not as an old wive's tale, but in fact, that road record has predictive value ATS. We'd all like to see it.

I well remember a debate at a card game once where some guy got hysterically angry. He insisted that the secret to success in NFL betting was in fumbles, he'd seen soooo many teams lose ATS because of fumbles, he'd seen a study which proved that the team that fumbles the least covers the most, etc.

I said I'd much rather bet on the teams that have been fumbling the most, and against those who have fumbled the least. He went crazy. I thought he might short circuit his pace maker. He called me an idiot. I am an idiot sometimes, but not in that instance. It might not any longer be true that betting on the fumbling teams is stone cold profitable in the NFL (I haven't checked the past few years), but for a long time it was.

And I'm sure no one here fails to understand that "teams that fumble the most cover the least" is not contradicted by "teams that have fumbled the most cover the most."