Thursday, March 23, 2017

Sweet 16 Day 1 Picking The Lines

This NCAA Tournament has lacked huge upsets and buzzer beaters, but it has been played at a really high level of basketball. In general, the sport is just getting better and better each year as the talent pool grows deeper, but the lack of upsets also played a role in keeping the big boys in more games.

And in the end, the high quality of basketball has just made this a really entertaining product. We don't need shocking upsets when we can watch powers like Kentucky/Wichita State and Wisconsin/Villanova trading punches at a high level for 40 minutes. And as we head into the penultimate weekend of the season, it's nothing but high quality of basketball ahead. And hey, who knows, maybe we'll still get a buzzer beater, too.

Sunday ATS: 3-5-0
2017 Tournament ATS: 25-25-2 (50%)
2016 Tournament ATS: 36-30-1 (55%)
2010-15 ATS: 220-167-11 (57%)

Oregon (+1.5) over Michigan: This is a fair line, as most computer ratings have a healthy Oregon as a slight favorite (Oregon, of course, is missing Chris Boucher). I don't think the Boucher absence will matter nearly as much as usual against Michigan, however, as the Wolverines are almost exclusively a perimeter shooting team. Oregon's perimeter defense is not great, but acceptable (7th in the Pac-12 in defensive 3PA/FGA ratio). Michigan has been playing remarkably well for the past couple of weeks, and maybe they will keep it up and Derrick Walton will continue his Kemba Walker impression, but if they don't then they don't particularly match up well with Oregon's weaknesses, and the Ducks are probably the better team.

Gonzaga (-3) over West Virginia: This line is small enough, particularly with the fact that Gonzaga shoots 73% at the free throw line if they need to lock up a game late, that I'd just pick whoever you think is going to win this game. West Virginia is a fantastic team as far as 4 seeds go. That said, Gonzaga is a strong 1 seed, and they have a team that is built to withstand Press Virginia. They have depth, experience, and savvy in a backcourt led by Nigel Williams-Goss, Jordan Mathews, and Josh Perkins, and they have the front court length and talent to protect the defensive glass. This West Virginia team is more capable of scoring efficiently in the half court than other recent vintages of Press Virginia have been, but still not good enough to win a controlled-style game against a team as good as the Zags.

Purdue (+5) over Kansas: I chose Purdue outright in my bracket and I'm going to stick with that pick here. Kansas had an incredible game against Michigan State, though the final score was deceptive as to how big the margin was for most of the 40 minutes. And overall, the Jayhawks came into the NCAA Tournament rated as easily the weakest 1 seed. As a match-up, Purdue's biggest problem is going to be staying in front of Frank Mason. But Kansas's front court size is a concern as well, against a Purdue front line that had its way with Iowa State. As good as Josh Jackson is, Caleb Swanigan is better.

Xavier (+7.5) over Arizona: Arizona is the better team, and Xavier's demolition of Florida State was probably a bit of a fluke, but this is an awfully large Vegas line. Xavier has a ton of length and size, and they will be able to match Arizona body-for-body inside, even if they don't have a big man quite as skilled as Lauri Markkanen. Offensively, Xavier passes the ball really well, and they have been surprisingly efficient offensively since losing point guard Edmond Sumner. A Xavier win would be an upset, but I wouldn't be at all surprised to see them play close enough to cover in a defeat.


Trevor Kraus said...

I've always wondered: When the KenPom line for a game differs substantially from the Vegas line, why not bet according to KenPom's prediction? If you did that ever single time, over years and years, wouldn't you have a positive expected value?

Trevor Kraus said...

For example, KP has Xavier losing by 4. Vegas is giving the XMen 7.5. There have to be hundreds and hundreds of similar scenarios every season. If you could find a way to identify and bet on every one...I'd think you'd do well, since the Vegas lines are more tailored to generate action than they are to correctly predicting what will happen, right?

Unknown said...

From what I understand, you can't do that because the sports book lines are more accurate predictors than any computer rating system, because they can account for circumstances not taken into account by computer rating systems. I don't have a lot of sports betting experience, but it would seem to me that a sports book's betting line already has Kenpom's and other similar computer ratings considered when setting lines, otherwise high-bankroll gamblers would just do exactly as you said. I believe you are right in saying that the line is set at what generates the most action, but this does not necessarily mean the greatest number of bets, but the greatest amount of money wagered. If a betting line is inaccurate, it gets get corrected quickly because a ton of money would come in on one side of the bet. People like to cite sports betting as an example of a "wise crowd" as opposed to an "irrational crowd" because it meets these 4 criteria:

Diversity of opinion-
Each person should have private information even if it's just an eccentric interpretation of the known facts.

People's opinions aren't determined by the opinions of those around them.

People are able to specialize and draw on local knowledge.

Some mechanism exists for turning private judgments into a collective decision.

Of course people are people and sometimes get carried away with narratives, which, if you can snuff out what is narrative and what is truth, is where you supposedly can make some money. But it seems to me that this is a lot like picking a winning mutual fund, which is why I don't bet on sports.

Unknown said...

I pulled that info from this wikipedia article:

Trevor Kraus said...

The stuff about the wisdom of crowds is fascinating. Thank you for sharing.

The part of what you said that I don't quite understand is this: "...because they can account for circumstances not taken into account by computer rating systems."

The St. Mary's — Arizona game last weekend was interesting because Vegas had Arizona as 5-point faves, I believe, and KenPom had St. Mary's as 2-point favorites. I would imagine that the main reason for that discrepancy is that Arizona is "long and athletic," whereas SMC was neither of those and hadn't played anybody as athletic as 'Zona. And so the average college basketball fan would, of course, believe Arizona to be the favorite.

But "length and athleticism" also seems to be the kind of eye-test thing a computer SHOULD ignore. That seems more narrative than reality. So is it possible that the circumstances a computer doesn't account for are ignored for good reason?

Anonymous said...

I don't think that KenPom can take into accounts like momentum/ X factor, or injuries ((Oregon, Xavier,), or unofficial home court (Kansas in KC) etc for instance.
And I am sorry Jeff B., I respect the heck out of you, but Josh Jackson will prove tonight that he is much better than Caleb Swanigan. He is a Lamborghini compared to a dump truck with an outside shot (okay, I exaggerate a bit because I'm a Kansas fan! But Jackson is better for real!)

Anonymous said...

Above mentioned Momentum / X factor = Michigan

Anonymous said...

got to Oregon +1.5 and closed the tab. commented here first of course.

Jeff said...

"Momentum" is not something that exists, so it's wise to bet against it. It's an example of a public bias that you want to short.

Unknown said...

Dude! Mass times velocity, why so anti-science? Just kidding, when I spoke of "factors that computer systems don't take into account" I was talking more about injuries/fatigue/travel. It's difficult to have a rational discussion about computer rating systems nowadays without coming off as a Charles Barkley or, conversely, a nerd, but I think it's important to understand how a system works, what it's strengths are and what it's limitations are. KenPom has gone into the impact of the absence of one player (, I still think it's unwise to simply bet games with a discrepancy between computers and "Vegas", unless it's an extremely unique the betting line not a better predictor than the computer systems? I haven't done too much research into it. Like I said in my earlier post...if it were as simple as shorting a team with "momentum" why wouldn't the highest bankrolled gamblers bet that heavy, in effect normalizing the line?

Unknown said...

Is it because once a public line is posted, a "groupthink" mentality takes over, therefore turning the "wise" crowd of bettors who get the early lines into an "irratioal" crowd of the general public who are all going off of the opening line? I don't know what's correct, just find this topic interesting.

Trevor Kraus said...

I find it really interesting too. My instinct is that the real effects of injuries/fatigue/travel are impossible for any bettor to know. Guys might be dealing with a sprained finger that no one knows about. Maybe some point guard didn't like how fluffy his pillow at the hotel was last night (travel). I think basing ANY betting decision on that kind of (mis)information would be foolish.

Would be interested to hear what Jeff thinks, if/when he has a free moment.