How does Vegas do it?

coldbrew -

None of the responses provided to your UCSB/UCLA total inquiry are incorrect, however, I think that I can provide a more direct solution.

I only handicap CBB, so cannot speak to other sports, but can tell you that the oddsmakers (whether it be offshore or Vegas) have become extremely reliant on the projected outcomes produced by Ken Pomeroy on his KenPom website.  The site projects outcomes for each and ever Division I game.  Yesterday, KenPom projected a final outcome of 95-69 (164 points); hence, the 164 released total. 

If you were to examine the opening totals vs. the KenPom projections on any given day, you will find that the posted total mirrors the KenPom projection in most instances.  The oddsmakers also use the KenPom projections for setting the point spread, but are a bit more willing to move from the KenPom number based on public perception of the teams.  For instance, That UCLA side opened at around -27.5 to -28.0 despite the -26 KenPom projection.  The oddsmakers knew that the whole world had just witnessed the Bruins put on an offensive clinic vs. Michigan on national television a couple of days prior, so slightly adjusted the line to reflect the anticipated public response.

The KenPom projections are actually fairly simplistic despite all of the algorithms.  Basically, it consists of projecting the number of possessions each team will have using current Adjusted Tempo ratings (AdjT) in relation to the national average and then computing the number of points each team will score based on current "tempo-free" points per possessions data.  All of this information is available on the KenPom website, and you can compute your own game projection in a matter of moments using a simple Excel formula. 

Of course, creating a line as good as the Pomeroy line does absolutely no good for handicapping purposes.  The trick is to create a better line than KenPom.  That is the task that I have undertaken for the last couple of years.  Find out what the KenPom line doesn't account for and you that omission to beat it.  Definitely easier said than done.  KenPom doesn't account for injuries or player absences, but I have found that with few exceptions (missing PGs perhaps being one) that absent personnel isn't as important as most believe because the oddsmakers will make any necessary adjustments. 

Also, KenPom improved his HCA ratings this year to more accurately reflect actual home court advantage (rather than a fixed percentage) - making it tougher for the player to realize value in that respect.

In my research, I have concluded that there are two primary facets in which the KenPom line does not accurately account.

(1)  It's slow to recognize teams that are playing substantially better than or worse than expected.  To be sure, Pomeroy's system recognize teams outperforming and underperforming expectations, but is slow to accept those deviations (from initial beginning year projections).  Accordingly, if you can identify teams significantly out or under performing vs. expectations, you can make some headway in this regard.  This year, Arkansas State is an example of a team which is over performing in relation to the KenPom projections, and for which value can be realized.

(2) The second methodology for beating the KenPom line is by properly identifying teams that are likely to be hyper-focused or lack focus.  Really good teams coming off SU losses tend to perform extremely well in their immediately following game.  For example, Kentucky was an auto-play for me (despite the big number) following its home loss to UCLA. As a -20 home favorite vs. a decent Valpo squad, UK won by 24, but the game wasn't even that close.  The Cats were up 34 midway through the 2-H before putting it into cruise control down the stretch.  Can you imagine how motivated those UK players were after enduring some intense Calipari practices following an upset home loss?  I think that the Celtics would have been in trouble on that day (ok, perhaps that's going too far) 

PapaShango - I would concur that the books don't care about getting it right and that their primary goal is to split the action, but with the proliferation of KenPom in the media and especially the betting market, books would be remiss if they didn't account for his projections in their lines.  The books use the KenPom projections precisely to attain 50/50 action because they know so many bettors will be examining the KenPom projections as part of their wagers. 

This is not conjecture.  There is empirical proof supporting this.  Let's examine just a couple of today's opening lines:

UCF/ at GW - Pomeroy projects GW 65-64 (129)

     Opening Line:  GW -1.5 (128.5)...The public bets GW up because of the absence of BJ Taylor while betting the total down for the same reason.

CofC at ECU - Pomeroy projects CofC 59-58 (117) 

     Opening Line:  ECU -1 (119.5)...Books know that CofC not a Public favorite (presumably because of its offensive struggles), so adjusts line accordingly.  Proves to be correct in this assertion, as the Public has bet ECU up to -2.  Public eschews such low totals so

Ga So at FIU - Pomeroy projects GA So 70-68 (138) 

     Opening Line:  GA So -2 (142.5)...Point spread in perfect alignment with KenPom.  Total is set higher for reasons unknown.

I'm not claiming that the point spread and total will always be in exact alignment with the KenPom projections or that the books don't incorporate other resources in constructing lines.  I opened my post by stating that the prior comments concerning line origination were not wrong.  I'm merely addressing the original inquiry as to how oddsmakers so accurately set the lines.  They most definitely examine the KenPom lines before releasing.  To not do so would be naïve, as so many in the marketplace rely upon KenPom numbers in wagering.

There is definitely fixed games too, I've already seen a couple this year, when its a 3point game with 8 seconds left and the team down is literally slapping the garbage out the player with the ball for 8 full seconds with no call I'm sorry but the NCAA needs to investigate garbage like that IMO.