On SN1/SN2/E1/E2 and NCAA Basketball

by James

in Organic Chemistry 1, Teaching

That title seems random.  What similarities are there between evaluating reactions in organic chemistry and betting on college basketball? Ridiculous question.  None at all, right?

Bear with me. I’d argue that one thing these two activities have in common is a thought process. In each case, the person evaluating SN1/SN2/E1/E2 and a bettor have to evaluate multiple variables, many of which can act in conflicting ways.

We’ll get to the sports betting in a moment. Let’s look at the SN1/SN2/E1/E2 first.  There’s a lot of things to think about here:

1-betting

Evaluating a reaction can require a lengthy chain of reasoning.  For instance, in this reaction: the substrate is secondary, so SN1/SN2/E1/E2 are all in play. We’ve got a negatively charged nucleophile, which rules out E1/SN1. Is it going to be SN2 or E2? Let’s look at the solvent – it’s DMSO, a polar aprotic solvent, which favors SN2. Furthermore, it’s at room temperature, which also favors SN2.

2-betting

In order to make that decision, we had to analyze at least 3 variables and be able to evaluate the importance of each. How did we know the relative effects of each variable? Good question:  these factors might look obvious in retrospect, but they weren’t something we could have just deduced from first principles.  No – these principles came from doing countless experiments, and analyzing the results. The key trends have been codified and put in your textbook.

If nothing else, life is a process of weighing multiple factors and making decisions.  For the sake of making an analogy, I wanted to find a good example of someone who made these kinds of decisions for a living, but could also *quantify* the results of their decisions.

Which brought me to sports betting.  Predicting the outcome of a match requires a lot of careful thought and analysis. What do sports bettors have to think about? What’s the thought process?

To get more information on this, I asked an aquaintance of mine, Steve (not his real name), a sometime resident of Las Vegas,  to share some insight on how he made decisions on when to lay money down on college basketball games. I wanted to understand how he analyzed teams and broke down the important variables involved.

Tell me about your record in betting on college basketball:

I can tell you I’ve won just under 63% of my bets since I first started doing this in 1991.  In a season in which I give this a lot of attention and I’m living in Las Vegas I might bet 40 games.

Could you explain to my readers why it isn’t as easy as just picking the #1 ranked team to win all the time?

If it were only that easy…it probably wouldn’t be any fun. Betting on the #1 ranked team is like buying at the top of the market.  Let’s say your team has won its first 12 games, and they’re playing great basketball.  How much better than 12-0 are they going to play?  You’re going to be laying a lot of points, and there may not be much value left.  There are other reasons, too.  Amatuer teams aren’t often psychologically ready to play as the #1 team in the country.  As soon as a team becomes #1 they often lose.

How do you make decisions on which games to bet?

One approach is to generate what are usually called “Power Ratings”.  There are all kinds of power ratings available though most of them of worthless.

To generate power ratings you can treat all final scores as simultaneous equations: if Kentucky played Duke (ignoring home court advantage for the moment) and won 88-82, you would generate an equation K – 6 = D.  If Kentucky then played Stanford and lost 74-71 you would have another equation S – 3 = K and so on.  If you solve all these simultaneous equations you’ll wind up with rating that might look like:

Stanford 101.4

Kentucky 98.2

Duke 89.1

Now you consider the home court advantage which is generally 4.5 points. If those are our power ratings and Stanford is playing Kentucky in California the spread would be 101.4 + 4.5 – 98.2 = 7.7.  Rounded to the nearest half point gives us Stanford – 7.5.  If the game were being played in Kentucky, the spread would be Kentucky – 1.5.

Now let’s say you think you know more than the people who make up the line.  (Unlikely.) For example you reason that the home court advantage is not the same for everyone.  Shouldn’t playing in front of 20,000 screaming fans be more of an advantage than playing in front of the 800 fans that might support a losing small conference team?  You might do some research (you look at the final scores of games played under these various scenarios) and decide that the home court in this second scenario is only worth 3.5 points. In this completely hypothetical example (it’s very unlikely the line-makers have not crunched the data to determine this with about as much accuracy as possible) , if your insight were correct, and 3.5 is a better predictor than 4.5, then you could enjoy a 1 point advantage over the house. (I.e., your spread is 1 point more accurate than the house.)

It would still be premature to start celebrating as it turns out you need at least a 1.5 point advantage just to break even because you generally must bet $11 to win $10 when placing bets on sports.

However, if you accumulated enough of these insights you might actually enjoy a sufficient advantage over the house to bet with positive expectation.

Since “power ratings” are based solely on points scored and allowed, how reliable are “power ratings” as a true judgment of a team’s ability?

Basketball ultimately comes down to putting the ball in the other teams’ basket and preventing the other team from putting it in your basket.  It doesn’t matter if you get there with strong rebounding, but poor shooting, or if you’re excel at running the floor, but struggle to score in a half court offense.  The final score encapsulates all of a team’s strength’s and weaknesses. Obviously a team can have an bad night, but an average of a team’s scores (taking into account the strength of the opponent) is a very accurate indicator of strength.  Often times the final score will be very close the spread,  and you’ll often hear bettors crediting the line makers with almost magical powers, but it’s really just a question of generating accurate power ratings.

So you look at the relative strength of the opposing teams, and home court advantage. Are there any other variables you look at?

There are factors which are difficult to quantify.  The most important one is the emotion or motivation of the teams playing.  If one team is especially motivated or unmotivated then you can take all those numbers you’ve crunched and throw them out the window. In my experience, particularly in college basketball, motivation trumps all other factors.  Of course, in theory everyone is playing as hard as they can all the time, but in practice that’s rarely the case. (This isn’t because athletes are consciously loafing. It’s because consistently achieving peak performance is a skill, and it’s not an easy one to master.)

Do you factor in matchups, or winning streaks or playing style?

I don’t factor in match ups or playing style.  In a sense I do factor in winning streaks: I bet against them.  One reason people usually lose is that they tend to think of a team’s performance level as fixed, when it’s not.  A team that has been winning will tend to get overconfident where as a team that has been losing can be highly motivated.  There are some exceptions, though. A team like Duke with a coach who is a great motivator can play at a very high level nearly every game. Occasionally a poorly performing team just has the wheels fall off. (Generally when there is bad chemistry between the coach and the players. This occurs mostly with a first-time head coach who isn’t fit for the job.)

Any other comments about sports betting in general?

I want to be careful not contribute to the hype that surrounds sports gambling.

I’ve had fun with this over the years, and made some money in the process, but I get up and go to work in the morning.

An astute reader will have noticed that winning 63% of 40 games per year doesn’t yield that much of an edge.

It’s very difficult to do better than 63%.  There are people who bet far more games, but you can’t maintain that win rate.

Many years I don’t get involved primarily because, although it can be fun and rewarding,  it’s time-consuming, my interest in watching sports has declined greatly over the years, and I don’t like to get involved in betting significant sums of money through illegal bookmakers or online.  It’s kind of a bachelor activity too. It doesn’t mix well with relationships.

I wouldn’t be surprised if your page views spike with this article, but I hope most or your readers stick to chemistry as a means of making a living.

Thanks so much to Steve for his time in helping me with this.

So if I have it correctly, the important variables are:

  • Each team’s record, including points scored / against, adjusted for the quality of previous opposition
  • The effect of home court advantage
  • Quality of coaching / motivation
  • Whether the “spread” is an accurate representation of these factors

Carefully evaluate each of those factors, and then make a decision.

Chemists are lucky. Unlike Steve, we get to work with near-perfect information. Let me tell you from experience – betting on the outcome of a reaction should be really boring. If things aren’t reproducible within a small margin of  error, there’s some variable you aren’t controlling properly (usually trace air or moisture). Furthermore:

  • molecules aren’t affected by psychology
  • molecules perform at a constant, predictable, and reproducible level of ability
  • chemists don’t have to worry about the spread
  • for every mole of reagents we use for a reaction, we are measuring the outcome of 6.02 times 10^23 “games”. Any fluctuations due to “luck” will cancel out.

But if it makes it more fun to think of a reaction as a competition between various pathways, then go for it.

Related Posts:

{ 0 comments… add one now }

Leave a Comment