3/20: 2019 NCAA Tournament Ramblings

I’m just gonna kinda ramble on stuff, and present a few philosophies you can evaluate to bet something that is very hard to beat. I don’t even totally believe in these things, but the goal here is to consider all sides of something and then do whatever you want.

Here are some things to think about:

—What you’ve seen from most teams already this season is probably irrelevant. The non-conference games are really an interesting group, because on one hand it’s kind of your only means of evaluation for small mid-major teams in terms of what happens when they play major conference teams. On the other hand, that was November, and no one is playing the same way now that they are in November. Sure, use the non-con’s (I absolutely do), but understand that they didn’t happen yesterday, and their ability to be predictive may be really bad. I’d be more concerned about what happened in those specific games, how pace changed, etc. than I would the actual final score-line. Also, be mindful of what the score was thru 30 mins, 35 mins, etc., as some of the blowouts became non-blowouts, or appeared better than they actually were for the mid-major in question. Was the game even close to competitive? Why was it? Why wasn’t it? These are good questions to answer for specific teams.

—What you’ve seen in conference tournaments falls under that category too. That’s 3-4 straight days, playing every day, playing teams you’ve played before, in a different gym than you’ll be in for these games. Injuries/fatigue are completely different dynamics. The amount of time to prepare for your opponent, and especially your opponent to prepare for you, is different. There’s just nothing predictive there. For every team that made a sick run to a conference title and parlayed that to NCAA success (Michigan I guess?) there is another anecdote of a team flopping (generally every Big East team that overachieves in the tourney).

—Once you kind of let go, and allow intuition and general understanding of metrics to take over instead of that stack of game results, it’s actually quite freeing. No, I don’t model this tournament, in that I don’t use my model projections to really make any bet. Most of the time, a basic efficiency calculation pitting the 2 teams overall numbers against each other is going to produce the current market. When it’s not (St. Mary’s), the market corrects pretty fast when limits are low. That doesn’t help me. KenPom is going to be within 1 of the spread in a ton of cases. That doesn’t help me. In fact, most openers throughout the year just use his # and shade it a point or two for injuries or other factors. That doesn’t help me. And it doesn’t help you, so stop using it. Trust me. If you bet every KenPom game that was 3+ points off from the Pinnacle close this year you’re at like 45%. Does that sound like a fun way to go thru the tournament? Same with Sagarin, actually even worse. ESPN’s BPI is better, but you still ain’t winning with it. Massey/Peabody, nope. A waste of time if there ever was one, sifting through publicly available game projections instead of either creating your own or taking a different approach using them merely as a starting point (the absolute starting-est starting point).

—OK so there are a bunch of things that aren’t valuable, like almost everything. Yikes! Is this really just a completely volatile crapshoot? Should we all flip coins? Maybe. And think how much time you’d save. No really, consider the possibility of just letting go and free up your days for just enjoying a very hard-to-predict tournament. Heads, favorite. Tails, underdog. You’d probably enjoy it so much more than you think.

—Also consider the possibility that the spreads might be perfect. Like absolutely, spot-on, NFL late-season game-day without injuries, fucking incredible. If that’s true, and the markets are rock-solid and the pros are already on vacation, what is there to do? Well, if we acknowledge that large favorites really are large favorites in the “truest” since, and the markets are correct, we can bet on some of these favorites to win their games and lay prices that are actually too low even though they don’t seem like it. Which may sound awful, but it depends on which favorites are actually favorites. I shall explain.

—Teams with a KenPom (KP for rest of the article) adjusted defensive efficiency advantage over their opponent of 100 places or more (so like, if i was 1st, and you were 101st, that’s a 100 advantage for me) are 49-1 SU the last 5 years in the tournament. They’re 23-23-4 ATS vs. the Pinnacle close, but remember, the spreads are fucking perfect…sort-of. The one loss is Virginia last year (I will just copy paste that sentence in for some of these, as it’s truly one of the more baffling games in history).

—OK, 49-1, but you can’t even bet 1-seed money-lines, and why on Earth would you want to in a 40-minute game at like -100000? Cool, you’re right, but that’s not the only teams that qualify here. What about 3’s that are really more like 1’s? Or 4’s that are more like 1.5’s? And what if they’re facing teams that are overseeded and should be 16’s because the committee can’t tell the difference? How can you tell? Is there actually value? Maybe!

—Money-line parlays are always completely scoffed at as the tool of the ignorant, but there are situations in the tournament’s first couple rounds where I’ve found them valuable, especially in situations where the underdog is being hammered for no good reason. Remember, the spread is artificial, as in the teams (whether you believe it or not) do not have it as a stated goal in the game to win by a specified amount. The real goal of the team is just to win, by any margin. And the goal of the team losing is, in the NCAA’s more than any other game, to extend the game as long as humanly possible for a chance to get the ball down by 1 possession. These are 19-year-olds, so the latter strategy actually works a lot, almost through sheer attrition sometimes. It’s agonizing to watch, but the goal is to win, not to make the game look good or win by 7.

—Anyway back to a discussion of 49-1. The teams with 100+ ADE mismatches in the RO64 this year: Michigan State, Texas Tech, Gonzaga, Michigan, Virginia, Tennessee, UNC, Kansas, Houston, Kentucky. I actually like Texas Tech and Houston at the current spreads anyway, but a parlay of their 2 money-lines pays about -375, maybe -400 depending on where you’re at. Is that worth it when teams with this big of a defensive disadvantage essentially never lose? Or lose this rarely, to put it more smartly? This isn’t “account dump” time, it’s not even close. Virginia is always possible, at all times. But lets just say that these statistical profiles are indicators of a very, very, very likely win, more likely than even the market maybe thinks. Oh, and Kansas is there too, btw, despite the fact you’ve all talked yourselves into Northeastern. That game’s very weird though, because Northeastern’s shooting is so elite and scares me to death.

—Northeastern qualifies for another dubious trend, though, one that is also predictive. Teams with 200+ raw defensive efficiency don’t do so hot in March. Defense above all other things is a great predictor for a lot of early-tournament situations. I use raw here because sometimes I just honestly disagree with the harsh adjustment that is made between raw and adjusted when it comes to bad major conference teams. SEC teams suck at defense for 15+ games, and they get a 60-spot bump on a list of 353. You play Vanderbilt twice, maybe try stopping them. Villanova, I believe, got even more than that this year. Sorry, when you don’t stop anyone for 2 months, I’m not sympathizing with you by saying that you are now able to stop someone in this 1-game situation because your conference was difficult. You allowed what you allowed per 100 possessions for the purpose of this exercise, and when your raw number is that poor, it’s a problem. That’s the point. Adjusted defense is a great stat, it just sometimes masks really bad defensive teams that are in the middle of these power conferences. The good news is, this entire argument is kinda moot, because no major conference 200+ team made it this year (there were 3 last year, and they all lost their first game).

—Teams with 200+ raw DE are 2-16 SU and 6-11-1 ATS (all ATS stats are vs. the Pinnacle close btw) the last 5 years in the first round of the tournament. In a couple of those situations, they were playing another team with 200+ raw DE so those are tossed. This year, teams with 200+ raw DE are: Fairleigh Dickinson, Gardner-Webb, Iona, and…Northeastern.

—Next up: 200/100 club, which may now not be a club at all. I added a 5th year on the back (basically because I got 20 minutes to plug in 2014 finally) and it kinda blew up the results, which is fine. Like, big-time, especially ATS. At least now we know! Teams with 200+ adjusted tempo, and 100+ raw DE are only 16-16-1 ATS once I plugged 2014 in, so hey that’s kind of a bummer. Noisy Noiserson. Those teams are still 11-22 SU, and a ton of them have been small spread games, so maybe there’s still something there, because the tournament is different than everything when it comes to winning and losing and the stakes. Or maybe not. Who knows.

—Teams in the 200/100 club this year: Villanova, Fairleigh Dickinson, North Dakota State, Baylor, Montana, Colgate, and…Northeastern.

—Next up: A 100+ AOE (adjusted off eff) rank advantage isn’t as awesome as a defensive one (remember, those teams were 49-1 SU), but it’s still pretty damn awesome and it happens all the time. Teams with a 100+ AOE place advantage are 47-7 SU (but only 24-27-3 ATS) the last 5 years.

—What this should show you is that if you are imbalanced in your profile, you are cursed in this tournament. That’s why the “Championship profile” involves high-rankings on both sides of the ball…because…duh. Of course those teams make it to the end. They don’t have warts, or at least, the warts aren’t massive like they are for some teams.

—Teams with a 100+ AOE place advantage in their first-round match-up: UCF, Virginia Tech, Michigan State, Gonzaga, Virginia, Purdue, UNC and Kentucky. North Dakota State had a 173-place advantage in tonight’s play-in game: they won, but didn’t cover…barely. And if that isn’t just a symbol of this whole fucking concept, I don’t know what is.

—So how do you use this info? Well first off, I’d pick your bracket upsets elsewhere. There are plenty of high-seeded teams we didn’t mention at all yet, which at least shows you that maybe they are slightly more vulnerable. But also, I’d consider firing up a few money-line parlays where available and get a reasonable return for teams that have a more massive win-probability advantage than the market thinks, because money-line and spread are tied together in markets.

—Also, don’t JUST use this data. Combine it with all the other things you’re thinking about. Every game should be viewed individually as its own thing, and sometimes stuff’s just gonna be different. What if there’s an injury, or some other factor that cannot possibly be captured here, specific to an individual game? That’s part of why trends, in some ways, are so stupid, but also, in some ways, helpful when combined with an understanding of the individual game.


—Here are some things I think are valuable: UNC to win the title at 9-10/1+ (as I mentioned on the pod)…their path is a huge joke until the Regional Final, and then their Final Four opponent might be a total out-classed one as well. Could be 2 tough games at 10/1, how fun is that?

—Virginia’s pod is probably the most likely to go haywire because it’s filled with the chokey teams of the decade, who all play like 40-possession rockfights. That being said, I’d rather have Wisconsin than Tennessee, Cincinnati or Purdue, and especially Kansas State…and Wisconsin is currently a higher price than all of them. 25/1 isn’t bad there. You could also buy Virginia at +110-+125 and use Wisconsin as a break-even hedge against that since they’re the only other valuable team in the region, in my opinion. You don’t always get there for sure, but you get there a lot of the time with that approach, enough that it works for me. You’ll never convince me Tennessee is valuable, and that’s probably a bad job by me.

—Gonzaga’s pod is the 2nd most likely to go haywire because that’s where the actual good teams are for the purposes of this tournament. Buffalo, Texas Tech, Nevada and Michigan are all GREAT teams in this tournament, and only 1 can make the Elite 8 (or Florida, but I have trouble knowing what to make of them). I’m not sure Buffalo’s that much worse than the others, so 25/1-30/1 is really appealing to win the Region there. Is it weird that I like both Marquette and Murray State’s prices too despite the fact they’re playing each other? 100/1 and you play Leonard Hamilton and then Gonzaga in the 2 games after the 5/12?

—Duke’s region is completely uninteresting, and Carolina’s is barely interesting.

—The upsets (defined as 4+ pts or more in terms of spread) I think are most likely to happen, in order: UC Irvine beats Kansas State, Murray State beats Marquette, Liberty beats Miss State, St. Mary’s beats Villanova, then my boys ODU at +800 over Purdue, or maybe Yale, but I think LSU’s gonna win that game like 101-93. If you like Northeastern, best of luck. Need those 3’s.

—Totals I think are valuable at current numbers, but am not betting much on because I have absolutely grown to fucking hate totals in the NCAA tournament, again in order: Washington-Utah State under, Wofford-Seton Hall under, Louisville-Minnesota over, UNC-Iona over because god help me if I have to watch that game with an under ticket, even on a tiny screen.

Spread bets out on the App tonight at some point after I discuss concepts with a few other people. 4 more days of talking about this stuff, which is really fun. Best of luck to you all.