[Pokerjoe]

I suppose this could/should go in the NBA forum, but I wanted this thread not to be about picks, but about the handicapping behind them. I'm pretty much hoping to stimulate some discussion, overtly about NBA handicapping, but really about handicapping generally.

In the NBA forum, I just read though some threads and it's depressing. Unless you're a bookie. But mods, move the thread there if you want.

So I'll lead off. I just going to though a card, make some points, to start things off. No idea if this will fly.

NJ/Atl early line 185', 186 and -6, -6'

ATL is in a b2b situation, and 4/5 (uh, that's back-to-back games situation, and 4 games in 5 days). My work shows that that obv impact seems to be in the line. Joe Johnson's absence, out the last four games, and by my estimate worth 2' pts on the side, also that seems to be in. But the total is way low, to me. Even taking 4 pts off the total for Johnson's X (X means out for injury; we've got to have shorthand, brothers), which is where I rate his total's (T) impact, it's low. ATL has come U the last 3 games (I've been betting on it, fwiw), but I think there's some overcompensation now. UNLESS: there's something wrong with NJ's Devon Harris, and there might be. He missed the 2 before the last, and played ~24min in that, iirc, and not real well. NJ also has 3 consec U, 2 with Harris x, the last with him going about 1/2 distance. I fig Harris' T impact at -2, -2'. But there's nothing showing publicly about him missing the game or being slowed, so I'm going to assume he's in. Understand also that in that last game, his first back from the injury, NJ was getting badly beat, and he might have been held back from his usual minutes because he just wasn't going to be wasted in a hopeless game.

I make this game, with the above noted adjustments, 188'/-6',-7. My standards for betting are 2 pt edges on sides, 4 on totals, and divided by those numbers for units, so an est sides edge of 3 pts would warrant 1' units bet. I don't think I ever have est side edges >5, btw. But meaning in this instance, I'll bite if the total drops to 184.

But it's important to note that this is entails keeping an eye out for late injury info on Harris. Further, this is what I call a seduction bet, in that I'm going to allow myself to be seduced into an O bet by a continuing drop in the total on this game. That isn't always okay, but, to me, it is generally okay when a source for the line change can be identified as mere public reaction to a trend, as in this case, where both teams have come under their last 3.

DEN/CHA, early line 201, 202 and +1',+2
This line is 2' at SIA, which I can't bet at, but which i check for square action. The line opened +1' and has moved up, obv, but if it's moved up most of all at SIA I can't call it sharp action. I make this game 200, CHA -1. CHA -1 would make for ~2 pt edge at +2 (you guys all know you can't count the zero in comparing PR's to lines, right?). I don't see any injuries or travel impacts, nothing, to account for this, so I'm on CHA+2. But if someone's got insight as to why DEN might have an inobv edge, post it. That's what this thread's about.

CLE/PHI, early line 196 and -7
I've got this at 197/-3, so I'm on CLE, for 2 units. The source for this est edge is obv: CLE got rolled by MIA in a big game 3B (3 games back), then rolled bad at MINN or MILW 2B, and well-beat LO (last out) by Detroit. As usual, I'm fading what I think is overreaction. CLE was in a weird spot vs MIA, and got beat down, and was still emotionally spent the following game. But that passes. I'm literally betting on it, anyway. But I'm not betting yet. I'm hoping it'll kick up to 7' somewhere. It opened 6' and moved up a notch from there; I doubt it'll kick up another notch, but I don't mind just watching for a while, because I don't think it'll drop a notch, either, or at least not so quickly that I can't still find a slow 7 somewhere. WHEN to bet is a big question in this biz, no doubt. I have my approaches. Just another discussion starter here, hopefully, but for this game I think the chance I'll nip a 7' just Pre-Game outweighs the chance I'll get shut out of the current 7.

BTW, when I say "I've got this at ..." I mean by my combination of PRs and handicapping adjustments. ATL, for example, I took off 1' points for their b2b and 4/5 situation, and 2' for Johnson's X, off my PR's (I keep my basic PRs as if all players are ready, and deduct for Xs, except for long term Xs, like Bynum for LA, and like, soon, really, Johnson here), and of course HCA, which I have at 3' for this matchup. Now, those are all arguably priced adjustments, no doubt. So, argue them. Again, that's what I'm hoping this thread will be good for. How do you price the various adjustments?

DET/HOU, 200, 200' and -6',-7
At first glance I made this game 204/-6', but I'm not jumping to bet the over yet. "Battier was sent home from practice Monday because of a stomach virus, the Houston Chronicle's Jonathan Feigen reports." I make Battier worth 1 pt in side value and -2 in T impact (btw, don't be mislead by my deductions for T impacts in this thread; I'm just as likely to adjust upward, it's just coincidence that the injury impacts mentioned so far have all been T downward). If I was convinced Battier was in play, I'd take the over here, but I'm not, so I won't. But if he's not, and I can find a +7' somewhere, I'll take Detroit. Something to watch for. Prob I'll be in a card game somewhere at gametime, but I can eyeball the net on my IPOD and make a late bet if need be. Also, Martin supposedly has the bug, too, and his X would further deflate my T est, but he played LO, so he's prob in tonight, too.

GS at DALL, 204', 205 and -9
I don't see anything interesting about this game, ex that by PR I make it 210/-8, so I'm on the over. This line opened at 203', which I didn't get, but I'm okay with 204'. I'm okay with up to 206, obv, but really there's no excuse for not taking this O now.

PHX at POR, 204 and -2', -3
By PR I have this at 202/-3'. I don't care that Robin Lopez is Q; I do care that Miller is X 1 (suspension). Yes, POR has bad-knees-Brandon Roy in action, and so you might think they're deep at the G, or even better off without a Miller and Roy on the court at the same time. But go here (which, btw, is from what I derive injury impacts, with some tweaking) http://basketballvalue.com/teamplaye...-2011&team=POR
Early, I'm making Miller's X worth ~2 on the side, -2 on the T, so I'm close to going U and PHX here, but I have some more work to do on the Miller impact. Actually, I'm on the U at 204; it's PHX I'll think about some more. Here's my concern: a real step up by Roy with Miller out. But, go here: http://basketballvalue.com/teamunits...-2011&team=POR and see that, at the least, the idea that Miller/Roy on the court at the same time is a negative doesn't hold up, and see also that Miller is really key to this team. But I still have to worry about the X stepup, as I call it, where a player's absence is made up for, in the first game, by the rest of the team. Anyway, for now, I'm adj POR -2 for Miller's X, which doesn't bring my PR down enough to warrant action. I'd need to get PHX +4, which isn't going to happen.

WASH at LAL, 207'-208' and -12',-13
LAL has been off 3; Gasol is a tad Q, Bynum a tad prob, but I'm assuming Gasol's a go, Bynum won't be, or won't play much, and isn't key anyway.
I made this game 212/-14'. I'm passing on LAL even if I found the -12 that would get them >2 from my PR, because there's almost no way I'm going to find value on LAL at home as a big fave, so if it looks like value (as it would if I found a -12 somewhere), I'd have to be suspicious. When earlier I was talking about seduction bets, this would be the kind I wouldn't like. If the line drops here it's because someone has private info that Gasol's X or limited or something. But I'm on the O at 207'. It opened at 206'; hope some of you got that.


Okay, enough. I just thought I'd show how I cap; hopefully others will do the same. Could be interesting. Or not, LOL.

[bookie]

Thanks for the link to BasketballValue...I hadn't seen that. My own database keeps pointspread based result numbers on player in-outs, and it will be interesting to compare them.

I'll just comment on the Den-Charl game. Charlotte by numbers is 1-4 ATS with significant margin against teams that play fast. It makes sense as these tend to be teams whose signature is offensive, and Charlotte just can't keep up if it turns into an up and down game. So I like Denver. I've also made an adjustment because Denver by my numbers is a crappy lead-holding team, but they do win. At this price range that's enough. If Denver would happen to be out in front by ten or more1H I'll take the middle equity.

I also like the 1H under. The logic is that Charlotte's game plan will be to slow the game down, and my numbers show they tend to be better at that in the 1H than the 2H. Denver off the plane isn't a big deal, but I do think there's a small effect in the first few minutes as they play the first game of a road trip at 5:00 on their body clocks.

[Pokerjoe]

Thanks for joining in, Bookie. By "pointspread based result numbers on player in-outs," do you mean a judgment on what the market seems to weigh a player's impact? Because I do that, too. I make a math-based assessment, but I adjust that by the market, somewhat, and by how the team actually seems to play in his absence, especially the guys getting new minutes, but without making too much of one game, because one game means so little. Meaning, if I think a player's X should raise the total 2 pts, but the market says 4, especially a few games in a row (because any given game has a lot of noise), I respect that. I don't mind arguing with the market--that's what any sport bet is, basically--I just don't get too stubborn about it.

As to your notes about CHA vs teams that play fast, I have never done any work on style affects, nor 1H bets. We can't all do everything, but I'll have to make a note of it for research.

Okay, now on to the wrap. The wrap, btw, to me, is as important as anything else you do in this biz. There's a lot to look for. Some of it's pretty simple: checking for OTs and injuries, and who ended up playing or not. Some of it's not so simple: figuring out a real score, as opposed to just the actual score, which is the most vanilla stat of all. Some of that is simple: if a team that shoots 70% FTs went 18 of 20 from the line, you'd take 4 pts off their score, right? Some of it's more subtle: resting players in blowouts, winning side or losing. I have some metrics I use. But because I do my PRs by hand, I pretty carefully peruse the box scores and play-by-play, the latter to see, pretty quickly, really, how the game flowed.

I'm not going through these game-by-game here. There doesn't seem much interest in this thread, which is cool, I don't mind letting it die out. But just for example, NJ/ATL, for ex, went O easy, which is important to me because I'd been thinking the market was dropping it low, so that was kind of affirmed. But though ATL covered easy, the lead was only 6 with 4 min left. Then again, NJ shot well from 3's and FTs, which is actually a bad sign. How'd they lose? ATL hit 60% on 2s. If a team hit that well on 2s, especially with 34 assists, then your D was weak or their O was great, but in either case, it's a notch down for NJ, and a notch up for ATL (on PRs, I mean). More of a notch up for ATL, because the high assists and 2pt shooting % makes me think they've really adjusted to life without Johnson, which is the kind of thing you need to look for. I've kind of been waiting for this, because I'm not as fond of Johnson's value as the market. My PRs move like glaciers, but I think I'm taking NJ down a 1/2, and raising ATL 1', which is huge for me, but as I said, I've been looking for some sign that ATL may be ready to play with Johnson, and 34 assists, even if it's against NJ, is a good sign when you lost a high minutes guard. Especially considering this was a game ATL could have been tired (5 games in 7 days), and in such games your depth gets tested, all the moreso if you've lost a high minutes guard.

Another thing I check is line movement, of course. Looking for where it happened, but also why. The only real move tonight was the WAS/LAL total, but that's simply explained. If a game moves toward my PRs, I figure there's probably no mystery to the movement. My PRs aren't magical. I've made money betting NBA for the previous 3 years (hard to believe), doing nothing magical at all. In fact I'd say my approach is anti-magical. Much more fundamental, really, which should be obv to anyone reading this. As I said, I respect the market. I argue with the market, but I respect my opponent. Ex: CLE. Obv, they are in a deep emotional tailspin right now. I was way off the market's eval today, and the market was right, not me. So I have to keep that in mind next out for CLE. I'll be looking for some sign that they're out of their funk, but until then, I have to drop them. I'll write my CLE PR like this: -4' (-2), meaning -4' is their basic rating, but in parentheses I'm putting an additional adjustment which I'll keep on them until they show me I shouldn't. For ATL, for another example, I've dropped the parentheses I've had on them for Johnson's X (he'll be out so long I may as well), and just rate them at +1' (par, btw, is 0 on my side's PRs).

Anyway, just one man's old school approach to the game, fwiw.

[Pokerjoe]

I'm going to do this again. I'm really hoping to get some of the guys who talk a big game in here to join in. Yeah, I'm baiting them a bit. But it seems like a lot of guys here are poseurs who talk high concept theory, but probably don't really know much about handicapping. Also, nuts-and-bolts of actual games handicapping could inspire some good discussion. For ex., Justin's statement a few days back aboiut HCA changing as the length of the home stand increases (iirc). I don't agree, and when I get a chance I'll back it up with stats. But at least it's an actually important, specific handicapping contention, and not the usual "mere" theory we argue about too much.

Also, I wanted to mention http://games.espn.go.com/fba/playerrater as a great injury source, because it includes better info than just injury lists, and presents it, team by team, with, also, a quick glance appraisal of a player's importance. I don't use the numbers--I mentioned above my preferred source for such stats--but I do use it to save myself some time, in that I don't bother crunching impacts of low rated players. It's a great format, to me.

CHI at CLE, 196'/+7, 7'
I've got this 199/+5', except that, as I noted earlier, I've got some hesitancy with CLE right now (B2b and 4in5, btw, for CLE). Most NBA teams play most regular season games without much emotion, but not all, not always. MIA seems to have gotten a monkey off their back beating CLE last week and are really soaring in the market (witness the jump in the overnights in their line vs Utah today, and, I've been told, a jump in their futures odds); CLE is the opposite. Things like this is where strict models fail. DET, earlier this year, funked out, with Stuckey's tantrums and talk of the coach having "lost" the team. CLE is there, now. But how long does the CLE fade last? Eventually it'll end, in that either the market will price it all in, or the team will de-funk. Stock market players, tho, would call it catching a falling knife. That 7' (Bookmaker) has me on the fence. At +8 might get me to bite (in fact it will if it's at -105), but more likely I'll only bet at +8', which I'll be calling a market overreaction. But on the other side? If you're fading CLE, I'd have to say their poor play of late is priced in, especially with the move from the opener of 6'. You missed the boat, to me, if you wanted to fade CLE again.

You think I'm nitpicking, betting at +8 only if I get the -105? To me, that's what this all entails. Setting a decision point, and showing some discipline. It's such a long, grinding season, you'll get beat up bad if you don't have decision points. Fading CLE again is a good ex. If you thought, after last night, they were still a good fade, fine, I'm not saying there aren't, especially considering they aren't getting a chance to catch their breath, here in their 4th game in 5 days. But at what point do you let it go? At what point do you say, ahh, it's priced in? Especially knowing that you missed the 6' open and have, if nothing else, already lost a half point? These 1/2 points are so big. In this thread's first post, I called over on a game at 204' that totaled 205. They just all add up, and a player has to be hugely conscious of them.

So far this year I'm 29.6-24.8, +2.8 units on totals (that's units bet and units lost, the latter including the juice; average bet of 1.43 units) and 33.4-26.5, +4.8 units, avg bet of 1.33 units for sides. That's a grinder's paycheck, but, to me, that's what works. This record is pretty much exactly what I've been doing in the past, this being my 4th year at it. Unexciting, no doubt. But I enjoy it. I'm no different at the poker tables, either, LOL.

DEN at BOS, 204/-6',-7
DEN is in a b2b spot vs a rested BOS team. That affects the total and the side, but trickier still is the current BOS injury mess. Rondo missed LO and missed practice yesterday. But just as important is Baby's missing practice, AND Shaq, and with the other O'Neil out long time, the Celts might be centerless. But because the T opened 201 and the opened -6, I'm reading that movement to mean, at the least, that Rondo's playing (despite listing as Q). I hate to say it, but reading the line is often how you fig who's playing, really. Rondo's X would drop the total, so if it's going up? As to the line: Shaq and Davis aren't much of an impact on the line, separately. But jointly, yeah, I think so. But this is a game for the inside-info guys, which isn't me, so I have to pass. A man's got to know his limitations.

TOR at NY 220'/-7,-6'
I've got this 216'/5', so if I get TOR +8 (not going to happen), I'll bite, and I'm already taking the U, seeing no reason to wait.

DET at NO, 187/-9',-10 from 185' and -8' open.
I really regret not being ready for the overnights, because the -8' was weak. I have it -11', even if I don't deduct a pt for Villenueva's X (he missed last night, a late scratch, and is Q tonight). But I'm really off the total: I have it 193'. So, I'm O here. W/o Villenueva, DET went way under last night (B2b for them, btw, vs a rested N.O. which historically has increased scored a smidgen); I'm not buying his X as being that big to their total.

IND at MILW, 189/-1',-2
I've gotten a little lost on MILW, because Bogut's X's were often combined with Gooden's and Maggettes. Gooden's prob still out, but IND prob loses Collison. I've got this 187/-1', so no action here. I think Collison's X would raise the total a point, bringing my # even closer to market if that's the case.

Boy, this is a lot of work for one bet I'd say is a whopping 55% likely so far, isn't it? I think you need to be okay with that. It's supposed to be work.

OKC at MINNY, 216-217/-5' from 218/-4'
I have it 215'/-5', but that's w/o accounting for Milicic's prob X. Hard to believe Darko's X would ever cost a line adj, but he's been decent this year, and his back up C, Tolliver, got knocked out just a few games back. I Love Love, but ... anyway, I'm not sure if that early line move means Darko's playing. It might.

GS at SA, 213/-10' from 215
I have it 219/-10'. GS b2b vs a rested SA (rested, btw, means off 2+). The line movement is suspicious, going against me. I'm okay with that. I'm over here. My rule is to not add to my bet because of the additional apparent (but not necessarily actual) line value, but to not get backed off, either. So I treat it as if I was still betting from an est T of 219 into an actual line of 215, even tho the actual bet will be at 213. So, what's up with that? Weird? Here's the thing: That extra two points on the total, I'm assuming, is necessary to overcome some inside info somewhere. IOW, I'm assuming that the value I saw vs the opener still exists, but isn't increased by the line change. I'm assuming the line change is intelligent, but equal to the extent of the move. That's controversial, I guess. I'm just saying that's how I deal with such situations. I know from my DB I don't need to be terrified of countervailing line moves.

MIA at UTAH, 190/+1,pk from a -2 open.
I have it 191'/-2, so I'm in a quandary here. It opened -2, so I was smooth on that. I've pumped up my MIA PR some, but I just don't go crazy with PR changes. Est -2 vs an actual +1 is a 2pt proj edge, so it's bet worthy. But I'm really on the fence with it. +1' would get me to bite. Yeah, it's that close. Many decisions are. In fact, if you don't see almost all NBA betting decisions as close, you're doing it wrong. BTW, the Greek has UTAH moneyline at +100, which means I can sell that 1 pt for 10 cents, which is a good sale. If I bite now, that's how I'll do it, but I'm hoping to wait and see more line creep here. Losing the 1 isn't as painful, if it happens, as gaining the 2 would be delightful. I am taking UTA, btw, almost certainly, if I can get at least +100 money line or +1 -105 or +1' -110, and I certainly should be able to (can right now, fwiw).

MEM at PHX, 218/-4
PHX in a b2b. I've got this 212/-3', so I'm U. Just a numbers play, so nothing more to say about it.

WAS at SAC, 200/-4
WAS in a b2b; Blatche X, SAC ok. And almost as I'm typing the total is jumping to 202, 202', from a 199 open, the side 4' from a 3' open. I have Blatche at -1' on the side, -1 on the T, putting the game, inho, at 198'/-4'. I dont' fig the T jump, then, but I'm not fading it. If the market goes bonkers and runs the T up to, say, 204', I'll fade the move. But that's a lot of market to move, still.

LAL/LAC 205/+8
No HCA here; no injuries; LAL B2B. I've got it 208'/+9. I'll be eyeballing the T for a poss late O bet, but that's about the extent of my interest.

Lot of writing, huh? I write as I think, thinking outloud, so to speak, so it's nothing to me.

Good luck.

[jgilmartin]

Great thread, Joe. I have yet to develop a successful NBA model, so I'm going to be totally useless to it. I'll be lurking it to see if I can learn something.

[ManBearPig]

I just stumbled on your thread and just wanted to say that although I really have nothing of value to add, it's interesting to read about some of the logic that goes into some plays. You put a lot of thought and effort into these posts and wanted you to know that someone appreciates it.

I'm sure you want some of the sharper people here to conversate with and discuss theories and reading this makes me feel pretty square even though I like to think of myself somewhere between. I do know that I have a lot more to learn that where I'm at now and although I may not generate much discussion it is interesting to read. I bet if this were in the main NBA thread you would probably get a few more hits.

[CHUBNUT]

Some Years back I set out to develope an NBA system which kept time to a minimum as its a for me a busy time of year with more important sports. Its always been my belief that level games produce a more consistant score and I only look at games where the home team is -2 to -6. After some basic number crunching I come up with a Total for the game ( I did try sides but they were'nt as good for some reason) I then compare that to Pinnys opener and if its 2 or more pts away then I bet it ( my total 200, Pinnys 204 then I go with the books on the over)
My theory here is that all things being equal in what should be a close game the total should be about 200, so by going under or over this figure the books are taking a view and almost certainly know something I dont. It sounds simplistic but it does take into account some solid reasoning and as I said, its not time consuming. Thanks once again Joe for all your posts which actually give out some views as opposed to the usual posts with zero to take from.

[Pokerjoe]

Thanks, guys.

That DEN/BOS line movement was interesting, no? The Q injuries, for BOS, all played (Rondo, Davis and Shaq) and yet DEN's top player, Camelo Antony, of whom there was no injury mentioned early, didn't play? As I said, a man's go to know his limitations, and I'm not privy to inside info like that. I read the lines to gain some, and did fig, from the early movement here, that Rondo was in, at least. Similarly, I figged from the MIL line move that Darko was playing. It's a crappy way to have to get injury info--it would be nice if the league shut down the inside info pipeline that gives some bettors a big advantage--but what are you going to do? Generally, despite my early morning posting in this thread, I actually much prefer to wait until >3pm to bet. In fact, generally, that's when I'm waking up, LOL.

Of MIA: I'm glad to say, that for no particular reason, just instinct, I never did bet UTAH. As I've said, I don't mind arguing with the market, but when the market has a good point to make--in this case, that MIA got a monkey off it's back beating CLE--then I'm much more willing to agree. MIA just really, really seems a different team now, after that CLE game, and may be the team they were projected to be. I'm kicking their PR up more tonight. I don't recall ever raising a team's PR as rapidly as I've been raising MIA's of late, yet I'm STILL behind the market. Well, maybe not: we'll see next game.

Which raises a Q: A, fave of x, beats B by 2x. Assuming the score is fair, do you raise A's PR, lower B's, or somewhere in between? Personally, that's some of the skill, and there's no pat answer. Lately, for ex, I've been dropping CLE's PR, and raising MIA's, but leaving their opponents unchanged. IOW, I'm marking the MIA and CLE outperformance as reflecting changes in those teams only, not their opps.

Bigger question. Really big, actually: Voulgaris is back to betting NBA totals this year. He took last year off to search for a job with an NBA front office, but having failed at that is back betting again. So I have to really reconsider whether I want to bet totals this year, especially considering I haven't made money at it this year. Playing Voulgaris in NBA totals is probably like playing Ivey in poker: if it's the only game in town, you might want to take up tiddly-winks. It's interesting that a single bettor could potentially take so much value out of the market that other players might want to quit, but it's possible in this case. Google his name; he did a good video interview somewhere, maybe 2+2. Guy has a solid, even amazing investment in this sport, and if he's playing, the market's def more efficient, maybe unbeatably so (from our POV, not his, which is why I said, in that thread about market efficiency, that it makes no sense to talk about market efficiency without reference to POV).

[jgilmartin]

Bigger question. Really big, actually: Voulgaris is back to betting NBA totals this year. He took last year off to search for a job with an NBA front office, but having failed at that is back betting again. So I have to really reconsider whether I want to bet totals this year, especially considering I haven't made money at it this year. Playing Voulgaris in NBA totals is probably like playing Ivey in poker: if it's the only game in town, you might want to take up tiddly-winks. It's interesting that a single bettor could potentially take so much value out of the market that other players might want to quit, but it's possible in this case. Google his name; he did a good video interview somewhere, maybe 2+2. Guy has a solid, even amazing investment in this sport, and if he's playing, the market's def more efficient, maybe unbeatably so (from our POV, not his, which is why I said, in that thread about market efficiency, that it makes no sense to talk about market efficiency without reference to POV).

Again, NBA isn't my area, but I would still have to believe sides are more efficient, based solely on the Pinnacle limits. Thoughts?

[Pokerjoe]

Well, Voulgaris has said that he almost only bets totals. That would imply that if he, even with his massive DB work and obsessive game study (he claims to watch EVERY game), can not find value in sides betting, that sides would be, yes, more efficient. That, and the greater interest in them, the greater limits accepted, etc.

I can only say that sides have been considerably easier for me than totals, over the years. Maybe it's because I can pick off how emotions can affect winning, but not how they affect scoring. My totals bets are entirely numbers based. My totals PRs, that is, are entirely program derived, whereas my sides PRs are instinct based, with the note that I have some metrics I use for breaking down boxscores to remove luck and complacency from the score.

But as I've said, market efficiency has an element of POV to it. An exec inside a corporation might be right in knowing that his firm's stock is wrongly priced, even though we mere investors might be just as right in assuming efficiency in the price.

So it's at least possible that, from my POV, with my talents, approach, whatever, there are inefficiencies unobservable to others from their POV, that is, with consideration for their talents, etc. Anyway, when and if the market beats me, I'll quit, I'm not blind to the possibility that the market could pass me by OR that I've just been lucky these past years (and I don't say that mockingly or smugly, either). I have quit football sides, for example, this year, after decent success in past years. Football totals, otoh, it would take a mountain of losing to quit, so deep has been my winning at it over the years. In NBA betting, I'll quit totals betting much more quickly than I'd ever quit sides betting.

[Pokerjoe]

You know, I continually check assumptions, so I didn't mind checking against Justin7's claim about home stand HCA changes.

Data from http://killersports.com/nba.py/query_matrix
First game of a home stand (setting each team to a 1-day rest, to control for the fact that after the first game of a home stand, a team almost never has B2B games, and opponent's frequently do, and we are measuring for HCA here, not rest impact)
season>2000 and site=home and rest=1 and o:rest=1 and p:site=away
ATS: 850-872-32 (-0.2) avg line: -3.3

Second game of a home stand:
ATS: 625-712-28 (-0.7) avg line: -3.2

Third game of a home stand:
ATS: 253-271-13 (0.0) avg line: -3.0

Fourth game of a home stand:
ATS: 133-108-0 (0.7) avg line: -3.2
Now, from this you can see that the average home line isn't varying much at all, nor in a useful pattern. You could, I suppose, argue that there has been an edge ATS in the 4th consecutive home game, 0.7 points, but the sample is small.

Fifth or greater home game:
ATS: 75-63-3 (1.6) avg line: -3.3
The average home line is staying the same, but again we have ATS advantage, even moreso than in the 4th game of the stand.

So Justin might have a point, (well, 1/2 point in the 4th game, 1' in the 5th+ games), but the sample is small. Everyone make their own judgment, obv. I'm thinking I might be willing to add a 1/2 point to the HCA in the 4th game of a home stand, and an additional 1/2 pt for each additional game in a homestand thereafter.

[ThisGuy]

Bigger question. Really big, actually: Voulgaris is back to betting NBA totals this year. He took last year off to search for a job with an NBA front office, but having failed at that is back betting again. So I have to really reconsider whether I want to bet totals this year, especially considering I haven't made money at it this year. Playing Voulgaris in NBA totals is probably like playing Ivey in poker: if it's the only game in town, you might want to take up tiddly-winks. It's interesting that a single bettor could potentially take so much value out of the market that other players might want to quit, but it's possible in this case. Google his name; he did a good video interview somewhere, maybe 2+2. Guy has a solid, even amazing investment in this sport, and if he's playing, the market's def more efficient, maybe unbeatably so (from our POV, not his, which is why I said, in that thread about market efficiency, that it makes no sense to talk about market efficiency without reference to POV).

Well if you want to stay in the game, that's one big reason to bet openers (barring limits) as opposed to 3pm when all the value has been bet out (which seems to be your contention)

[bookie]

Thanks for joining in, Bookie. By "pointspread based result numbers on player in-outs," do you mean a judgment on what the market seems to weigh a player's impact? Because I do that, too. I make a math-based assessment, but I adjust that by the market, somewhat, and by how the team actually seems to play in his absence, especially the guys getting new minutes, but without making too much of one game, because one game means so little. Meaning, if I think a player's X should raise the total 2 pts, but the market says 4, especially a few games in a row (because any given game has a lot of noise), I respect that. I don't mind arguing with the market--that's what any sport bet is, basically--I just don't get too stubborn about it.

As to your notes about CHA vs teams that play fast, I have never done any work on style affects, nor 1H bets. We can't all do everything, but I'll have to make a note of it for research.

Okay, now on to the wrap. The wrap, btw, to me, is as important as anything else you do in this biz. There's a lot to look for. Some of it's pretty simple: checking for OTs and injuries, and who ended up playing or not. Some of it's not so simple: figuring out a real score, as opposed to just the actual score, which is the most vanilla stat of all. Some of that is simple: if a team that shoots 70% FTs went 18 of 20 from the line, you'd take 4 pts off their score, right? Some of it's more subtle: resting players in blowouts, winning side or losing. I have some metrics I use. But because I do my PRs by hand, I pretty carefully peruse the box scores and play-by-play, the latter to see, pretty quickly, really, how the game flowed.

I'm not going through these game-by-game here. There doesn't seem much interest in this thread, which is cool, I don't mind letting it die out. But just for example, NJ/ATL, for ex, went O easy, which is important to me because I'd been thinking the market was dropping it low, so that was kind of affirmed. But though ATL covered easy, the lead was only 6 with 4 min left. Then again, NJ shot well from 3's and FTs, which is actually a bad sign. How'd they lose? ATL hit 60% on 2s. If a team hit that well on 2s, especially with 34 assists, then your D was weak or their O was great, but in either case, it's a notch down for NJ, and a notch up for ATL (on PRs, I mean). More of a notch up for ATL, because the high assists and 2pt shooting % makes me think they've really adjusted to life without Johnson, which is the kind of thing you need to look for. I've kind of been waiting for this, because I'm not as fond of Johnson's value as the market. My PRs move like glaciers, but I think I'm taking NJ down a 1/2, and raising ATL 1', which is huge for me, but as I said, I've been looking for some sign that ATL may be ready to play with Johnson, and 34 assists, even if it's against NJ, is a good sign when you lost a high minutes guard. Especially considering this was a game ATL could have been tired (5 games in 7 days), and in such games your depth gets tested, all the moreso if you've lost a high minutes guard.

Another thing I check is line movement, of course. Looking for where it happened, but also why. The only real move tonight was the WAS/LAL total, but that's simply explained. If a game moves toward my PRs, I figure there's probably no mystery to the movement. My PRs aren't magical. I've made money betting NBA for the previous 3 years (hard to believe), doing nothing magical at all. In fact I'd say my approach is anti-magical. Much more fundamental, really, which should be obv to anyone reading this. As I said, I respect the market. I argue with the market, but I respect my opponent. Ex: CLE. Obv, they are in a deep emotional tailspin right now. I was way off the market's eval today, and the market was right, not me. So I have to keep that in mind next out for CLE. I'll be looking for some sign that they're out of their funk, but until then, I have to drop them. I'll write my CLE PR like this: -4' (-2), meaning -4' is their basic rating, but in parentheses I'm putting an additional adjustment which I'll keep on them until they show me I shouldn't. For ATL, for another example, I've dropped the parentheses I've had on them for Johnson's X (he'll be out so long I may as well), and just rate them at +1' (par, btw, is 0 on my side's PRs).

Anyway, just one man's old school approach to the game, fwiw.

I haven't looked at SBR in a couple of days, so I haven't responded. I'm very interested in your approach, and hope that you will keep up the game by game commentary from time to time. If I see it I will certainly chime in on particular games that we look at differently.

As far as player injury evaluation, I don't pay much attention to how the market evaluates it. That's a hole in my game, but I basically think they usually over estimate the impact of a star player, and that if there is an impact it's delayed a game or two until opponents have a chance to point a game plan at a replacement player and whoever gets the minutes comes down off the rush of getting them. I do have an actual database report that tells me the pointspread results of each team with each player in and out of the lineup, and I especially look for "glue" players (i.e. Nick Collison) who seem to make a pointspread impact or others (like Danny Granger) who's absence the market discounts but whose absence in fact seems to help his team.

I'm especially interested in your old school approach to reading boxscores. I am anticipating an NBA strike next year, which will mean I'll have to bet CBB. I don't like the college game, and haven't followed it for years. This year, though, I'm following a couple of conferences in order to try to get a routine down in case next year CBB is all there is to bet. What I'm finding is that the kind of boxscore reading that you're advocating is the only way I can get some handle on how the teams are playing, but I wish I were a lot better at it and I have to figure out a way to prepare myself to know more what I'm looking for with each team. But anyway the more of that boxscore and play-by-play analysis you include the more I'll like your thread.

[Flying Dutchman]

Nice write-ups Poker. I never get this deep as everything is programmed into the model.

As far as Voulgaris goes, from the video on 2+2 it seems that he used all the video to develop an angle, not a model. And then when he got other folks involved, they also could see the angle so the advantage dissipated due the info getting out.

[Data]

Just a few notes.
1) thank you Pokerjoe for trying to get the discussion going
2) I do not handicap as you do, I have my handicapping automated by defining my rules/angles, this way I can backtest it and see how "good" my handicapping is while you cannot backtest your "feel"-like handicaps
3) the way you tested Justin7's statement on HCA is way, way wrong
4) Voulgaris does not sound sharp

[Data]

if a team that shoots 70% FTs went 18 of 20 from the line, you'd take 4 pts off their score, right?

One more thing, a team does not shoot FT, the players do.

[Pokerjoe]

Well if you want to stay in the game, that's one big reason to bet openers (barring limits) as opposed to 3pm when all the value has been bet out (which seems to be your contention)

I wouldn't say all the value's out by 3. I have no problem betting sides >3. I prefer it, actually. Late money on sides is frequently dumb and adds 1/2 points of pure value, and late injury info levels the playing field against the inside info guys.

But totals are different. I might just bet total's early only, sides early and late. As it is I've never wanted to assume added value because of totals changes; I will assume loss of value.

But injuries have more impact on totals than sides (contrary to popular belief, probably.) Only the top few players on most teams affect the line, but even subs, being in or out of the lineup, affect the total. Players who, in or out, don't have much line impact, drastically change the total. Yao, for example. Much bigger impact on the total than the side, even if he's only playing 20 min. Ridnour, for Minny, for another ex, might not seem important, and isn't on the line. But on the total? Yes. There are lots of 20 minute players whose in/out affects the total by 2+ pts.

So it's a tradeoff, when to bet. My feeling, these days, is that maybe I'd rather only bet totals o/n. Of course that eliminates a lot of games because so many games aren't lined o/n.

[Pokerjoe]

Just a few notes.
1) thank you Pokerjoe for trying to get the discussion going
2) I do not handicap as you do, I have my handicapping automated by defining my rules/angles, this way I can backtest it and see how "good" my handicapping is while you cannot backtest your "feel"-like handicaps
3) the way you tested Justin7's statement on HCA is way, way wrong
4) Voulgaris does not sound sharp

About 2, no, I can not back test my sides PRs, but I can, and have, had years go by and keep records. My totals calculations are not by feel, they're by program alone. Those I can test, but backtesting is hugely overrated, it is ALWAYS a form of data-mining. Only battlefield testing is useful, to me. And the battlefield is always changing.

About 3, Saying it's wrong is not useful, and I kind of detest such snide comments, really, nothing personal. Show us how my test of home-stand HCA is wrong. I've presented evidence that the market doesn't think home-stand is important, and evidence that the scores aren't much reflecting it. So if the market and the scores aren't reflecting it, what else is there? If you don't control for days rested, as I did, you'll end up mistaking HCA with rest-impact. If you want individualized team home stands, you have pretty much no sample size.

But without any ill-will, I genuinely say, if you have a better methodology, show it, we'd all like to see an actual real, specific contribution from you for once. That's kind of what this thread's about. Posting "You're wrong, but I'm not saying how" is just juvenile.

About 4, What does that mean, doesn't "sound" sharp?

[Pokerjoe]

One more thing, a team does not shoot FT, the players do.

Other than "DUH," what response do you expect to that?

Let me counter-DUH: a team is made up of ..... players.

If a team shot over its average, it doesn't much matter which individuals did it, generally. That will all come out in the wash. IF it was mostly a team's high % shooters going to the line that night, fine, that could have some impact, duh. If you put Shaq on the line all night and the team's avg was low because of it, fine, and also duh. That's all in the box score, Data.

I could go on at length about reading boxcores, but I don't have the energy. It's also the kind of thing that lends itself well to databasing, and in case I haven't made it clear, I DB NBA. My totals calcs are entirely DB/program derived, my sides partially so. But I have not as yet been able to DB such a thing as "MIA got a monkey off it's back with it's big win in CLE." Or "CLE was humiliated in what will be it's biggest game of the year (they aren't doing anything post-season) and have been in a funk since."

[ThisGuy]

I wouldn't say all the value's out by 3. I have no problem betting sides >3. I prefer it, actually. Late money on sides is frequently dumb and adds 1/2 points of pure value, and late injury info levels the playing field against the inside info guys.

But totals are different. I might just bet total's early only, sides early and late. As it is I've never wanted to assume added value because of totals changes; I will assume loss of value.

I was only talking about totals, that's why I only quoted where you talked about totals, you made the rest of the stuff clear before. I was trying to help you out (since you were saying it's your BIG question)

[ThisGuy]

"But I have not as yet been able to DB such a thing as MIA got a monkey off it's back with it's big win in CLE." Or "CLE was humiliated in what will be it's biggest game of the year (they aren't doing anything post-season) and have been in a funk since."

That's cause it doesn't exist. They've just hit a easy stretch of schedule and there was obv going to be an adjustment period (despite what the public believed), you can put whatever narrative you want on it though ... Cleveland is a horrible team devoid of talent and the short sample size at beginning of year skewed ur perception

[Pokerjoe]

Nice write-ups Poker. I never get this deep as everything is programmed into the model.

As far as Voulgaris goes, from the video on 2+2 it seems that he used all the video to develop an angle, not a model. And then when he got other folks involved, they also could see the angle so the advantage dissipated due the info getting out.

No, he developed a hugely expensive model. I don't know exactly what video is at 2+2, or what threads people might have read, or, whatever, so forgive me for that. I'm referring mostly to people I know who I trust not to actually have their heads up their butts who tell me A) he's back and B) he's really good and C) he bets big, though disguised. But that's anecdotal only. Sometime soon I'll do some testing of the market this year to see if it's gotten tougher (and I don't mean my w/l results, which are admittedly are trivial). That's all that counts.

[Pokerjoe]

That's cause it doesn't exist. They've just hit a easy stretch of schedule and there was obv going to be an adjustment period (despite what the public believed), you can put whatever narrative you want on it though ... Cleveland is a horrible team devoid of talent and the short sample size at beginning of year skewed ur perception

We can not intelligently revolve this dispute, especially considering your terrible semantics.

To me, dismissing the reality of emotions in sports is ridiculous, and I'm a numbers oriented guy saying it, too, not a numbers-denier. I've learned that there are people who literally can not see the emotions. Like being unable to see the color red, they can't read people's faces. They can win at online poker, not live. And if you can't see emotions, I can't show them to you, can I?

Maybe this thread is about program bettors taking their game to a higher level by getting their head IN the game.

But of your semantics, saying, of any team, that they're "horrible" is not useful. Only "exactly how horrible" is useful. Only are they "10 pt dog" horrible or "12 pt dog horrible" or merely "8 pt dog horrible" is useful. Because that's the question: exactly how horrible. "Horrible," "great!", "gonna kill'em!" is Player's Talk terminology. Were they 10 pts horrible pre-MIA, 12 pts horrible since? That's the question. Not are they "horrible."

And "easy stretch of schedule" is also a useless description, as presented here. The spread takes care of the "easy" part, making it meaningless in that respect. No game is easy, or hard, after the spread's involved.

Unless you're supposing that MIA is disproportionately likely to cover as big faves (like, of pick'em vs Utah?). In which case you should study their record pre-CLE.

Here's the thing: if you don't think or can't see that the market dramatically raised it's valuation on MIA and dropped in on CLE after that game, you aren't reading the market. That's not the same thing as saying the market was right to do so, nor am I saying that MIA's covers and CLE's failures since is proof the market was right. I'm saying your (and miy) judgment of the market's wisdom is the test, and I'm also saying that blindly ignoring the market is foolish. At the same time, I'm saying that every bet is a statement that you think the market is wrong, and obv I'm not saying not to do that. I'm saying respect your opponent, and when he beats you, study why, with a mind open to the possibility that YOU'RE the one who's wrong.

Sometimes, when I lose to the market, I was right and the market just go lucky. Sometimes I was lucky and the market was right. The key thing, to me, when doing non-program handicapping (call it gut-based capping if you want) is to be aware of the nature of the battle.

I think of my PRs as something like prices. I continually study the underlying stock, particularly the incoming data. I try not to overreact (as I've said, I move my PRs' pretty slowly) but I try not to deny the possibility of change of the team OR error in my appraisal of the team.

This is all more iffy than program betting (as I've said, this is for sides only, for me; I don't know how to gut-cap totals). And more fun. And more profitable.

IMO, obv, I'm saying that the difficulty of adding instinct-based appraisal to ratings creates opportunity for edge. I'm saying that that which is easily solved by program (such as MLB) leads to a vastly more efficient market (such as I think MLB is, re any other sport). Solvability by program, in fact, is more important in market efficiency than market size, I'd say. Much more.

NBA totals is closer to MLB than NBA sides is, imo.

[Pokerjoe]

I was only talking about totals, that's why I only quoted where you talked about totals, you made the rest of the stuff clear before. I was trying to help you out (since you were saying it's your BIG question)

No, the BIG question, the interesting one, was as to a single bettor's impact on the market. And it's just a Q, to me. I have no opinion, I'm just curious. To me, the very possibility that a single player's participation in a market could affect efficiency is interesting, yes.

[Flying Dutchman]

No, he developed a hugely expensive model. I don't know exactly what video is at 2+2, or what threads people might have read, or, whatever, so forgive me for that. I'm referring mostly to people I know who I trust not to actually have their heads up their butts who tell me A) he's back and B) he's really good and C) he bets big, though disguised. But that's anecdotal only. Sometime soon I'll do some testing of the market this year to see if it's gotten tougher (and I don't mean my w/l results, which are admittedly are trivial). That's all that counts.

Crap, I can't find it now, it was posted in the NBA forum somewhere. I knew I should have saved that URL.

Anyway, if you watch it you'll see that underneath all that expense, he seems to have developed very simple angle approachs. Else, why could other folks have "stolen" them by simply watching his plays? If not, then he simply was deceptive in the video.

In my modeling, even if you knew the larger part of it, it would be extremely hard to reverse engineer it. Too much math going on under the hood.

[ThisGuy]

But of your semantics, saying, of any team, that they're "horrible" is not useful. Only "exactly how horrible" is useful. Only are they "10 pt dog" horrible or "12 pt dog horrible" or merely "8 pt dog horrible" is useful. Because that's the question: exactly how horrible. "Horrible," "great!", "gonna kill'em!" is Player's Talk terminology. Were they 10 pts horrible pre-MIA, 12 pts horrible since? That's the question. Not are they "horrible."

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

[bookie]

If anybody saved the link to that Voulgaris video please post it.

[luegofuego]

Pokerjoe,

Much kudos for posting in the think tank, your style of winning is one I'm really trying to emulate. I've taken a lot from a ton of your posts even if we're not sharing sports. Looking forward to going through the thread more closely.

[Data]

backtesting is hugely overrated, it is ALWAYS a form of data-mining...

But without any ill-will, I genuinely say, if you have a better methodology, show it, we'd all like to see an actual real, specific contribution from you for once. That's kind of what this thread's about. Posting "You're wrong, but I'm not saying how" is just juvenile.

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

About 4, What does that mean, doesn't "sound" sharp?

Never mind.

[sideloaded]

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



Never mind.

Data things not going so well on the enterprise?

[GELATINOUS CUBE]

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.

[CHUBNUT]

Data's reply is typical of the sort of post on this forum, post plenty say nothing. Its also sad that Austin7 hasnt had some imput considering his supposed knowledge of books mechanics. All in all this thread is an epitaph to what this forum is about, a conduit for a minority of socially inept people to convey a sense of superiority over feckless beginners in betting. I have no doubt that none of them actually make a profit from betting which annoys me more when someone like Pokerjoe goes out of his way to try teaching real skills.

Pokerjoe, you are wasting your time here and need to move away from these mathematical jokers who constantly tell people they are wrong while offering nothing in reply.

[Wrecktangle]

I take Poker's point on backtesting. If you are not very careful, you will over-fit.

Frankly, few folks can really get the modeling right, and I'm not about to tell you I get it right all the time either. And sometimes the market is too good to beat, or a league is too chaotic to make much mathematical sense of it.

But I enjoy it, I like working with sports data, I like modeling and testing new things, and along the way I make a little money on it.

All the huffing and puffing on "being sharp" and "making big money betting sports" is, frankly, mostly (98% of the time) huffing and puffing.

[Pokerjoe]

I take Poker's point on backtesting. If you are not very careful, you will over-fit.

Frankly, few folks can really get the modeling right, and I'm not about to tell you I get it right all the time either. And sometimes the market is too good to beat, or a league is too chaotic to make much mathematical sense of it.

But I enjoy it, I like working with sports data, I like modeling and testing new things, and along the way I make a little money on it.

All the huffing and puffing on "being sharp" and "making big money betting sports" is, frankly, mostly (98% of the time) huffing and puffing.

That's about exactly where I stand as well. I like playing with numbers. I enjoy the challenge. It's my hobby. I've been lucky to have made between ~8k to ~24k every year (this year will be much higher, though) for, probably, a decade now (from whenever I started posting at Majorwager). I still earn my living playing cards, and consider poker much steadier. Well, it IS much steadier. Sports is such a freaking roller-coaster ride and I don't have the temperament for it (in NHL this year, I had a 1-14 stretch followed now by a 10-0 stretch; WTF? Variance is a scary thing, LOL). Compare it to poker where a bad month is breaking even. Not the same. Funny, isn't it, how many of us "gamblers" are risk averse?


But I wouldn't do sportsbetting if I couldn't make something at it. I can not comprehend the "joy of losing" that most sportsbettors must take. And there's this: they're missing the joy of handicapping, of using your brain to figure out a very interesting, complex puzzle, against very tough competitors, within a framework where it's no one's mere opinion that will settle an issue, it's money that will. I like that.

And I'm just about at the point where I think I can really beat basketball. CBB is going fine, and if it goes fine this season, and NBA does, then this will make the 4th straight year I've beaten it, and beaten it pretty well (55%, 800+ bets/year). Crossing my fingers, then, for this season.

So, does waiting for 4 full seasons to pass successfully before I say, yeah, I can beat this sport, I think I'll go pro, make me a nit? You think, LOL? Maybe it's because I beat the HELL out of football sides for 2 seasons, then this year got so buried so fast I quit by November (totals is fine, though, TG). I did that in baseball once, too. Won big over a 400 bet sample one season; had to quit by July the next, so quickly did I dump back all the profits.

But if I couldn't win at it? I can't imagine doing it further. But then again, I can not understand the joy of playing slots, which most people prefer over poker, not despite the lack of thought involved in it, but BECAUSE of the lack of thought. They actually want to NOT think. Oh, well. As long as they're happy.

[Pokerjoe]

You know, here's a sign, of any given market, that things are tougher: if you're number of bets is down.

Mine is down for NBA totals this year some. It's down even more so far for CBB totals, which is scary, because it might mean the books have finally figured out how to make the calculations. Oh, well, gravy trains do come to an end.