Tuesday, March 10, 2015

Sports Philosophy: Part 2 (Everything Doesn't Happen for a Reason)

“The narrative fallacy addresses our limited ability to look at sequences of facts without weaving an explanation into them, or, equivalently, forcing a logical link, an arrow of relationship upon them. Explanations bind facts together. They make them all the more easily remembered; they help them make more sense. Where this propensity can go wrong is when it increases our impression of understanding.”

-Nassim Nicholas Taleb, The Black Swan

The narrative fallacy, in simpler terms, is the mistaken belief that everything always happens for a reason. Our brains like to link things together and to make stories out of sequences of events that aren’t actually bound by a story. It's a trap we all fall into. The narrative fallacy makes us feel better about things that happen in our lives, good and bad, because if we can give explanations as to why things have occurred we feel a greater sense of control. Accepting that most things are largely influenced by random forces is a difficult thing for most people. To accept randomness is to accept that we are not in control of the world.

Failing to account for randomness is only human nature and it surely makes sense that people would think this way. Our minds are able to function as effectively as they do largely because they employ pattern recognition, which means subconsciously matching external stimuli with internal memories. The past knowledge provides a baseline for what is currently happening in front of us and in many cases this is extremely useful. We have the ability to piece together fragments of sight as clear pictures, even when our eyes miss part of an image. When a blur flashes in front of us too fast to see, sometimes it’s possible for our eyes to figure out what we actually saw. This is because the memory helps to fill in the missing parts of the picture. Our memories help to form images in a pattern-like way, and this makes eyesight more effective.

Another example of pattern recognition is our ability to read. We can remember sequences of words if they are formed into sentences, and it’s more difficult if the words are jumbled and don’t make sense. Flow of a sentence is essentially a pattern and therefore far easier to recollect. This is also why it’s so easy to recite the alphabet but fairly difficult to do it in reverse. The alphabet song is a pattern that we recognize, but the reverse alphabet is not.

So what’s the problem and how does this relate? For one thing, the assistance our mind gives us in remembering things can sometimes be negative. If we are relying on past visual stimuli in order to make sense of something presently in front of us, then we are sure to sometimes think that we saw something that wasn’t really there. A perfect example of this is what happens when you see a person who appears to look like someone you know. For a moment they appear identical to the person in your memory, but they obviously are not. Sometimes we see something that isn’t really there.

This problem of pattern recognition is particularly prevalent in sports, which is an ideal forum for spinning random events into a neat story arc. Momentum, turning points, teams of destiny, etc. have been created to make fans feel more in control while watching games, and these misconceptions are difficult to ignore. When we’re watching something, our pre-programmed brains go out of their way to make us think that what we are seeing is the same as something we have seen before. Sports have far too much complexity and data to remember as a jumbled mess, so the pattern recognition part of our brains comes to the rescue and turns a headache into something that can be enjoyed. To put it bluntly, It’s not your fault that you are misconstruing the events happening in front of your eyes because that’s what your brain is supposed to do.

Sports outcomes are not patterns and they are not stories. Sports are simply just sports. The narrative fallacy is inevitable in all aspects of life though, so random competitions wind up being interpreted as something more than what they really are. Even if you are able to avoid concocting narratives in your own head, ESPN will surely fill that void. Sports documentaries are literally founded on the idea of narratives and that’s why people watch them. Hearing the big speech from the coach to turn the game around raises hairs on everyone’s arms. Again, it’s just human nature.

There’s nothing wrong with believing fallacies if you’re just a fan trying to enjoy a game. You may not have the most realistic expectations for your team if you believe they’re a team of destiny or some other nonsense, but who cares? Fans are allowed to be fans and it certainly makes for greater excitement when rivals clash with each other about things that don’t really matter. If you’re trying to actually make good predictions (and perhaps make money off them), however, it’s imperative that you ditch this way of thinking.

One type of pattern that every sports fan surely knows about is the “hot streak”. A team or player will have short term success which then becomes medium term success and then eventually the successes start mounting and you have a huge run of positive outcomes (this happens to sports bettors too, which I will get to as well). We have so many memories of hot streaks that we think they are predictable. Well here’s some breaking news: hot streaks are not predictable. For every five game winning streak there’s been a four game winning streak that ended with a loss in game five. For every batter that’s recorded a hit in ten straight games there’s another one who hit safely in nine before going 0 for 4 on day ten.

Our brains fall victim to the narrative fallacy in the case of streaks because we only remember the ones that actually materialized. Long streaks are recognizable patterns, e.g. you can easily remember a team’s win/loss record over the past month if every game is a win. If the wins and losses occur in a more ordinary fashion, it’s harder to remember each game result in succession. So every time a streak gets to a noticeable point the natural inclination is to believe that it will continue. Sometimes the streak continues but sometimes it doesn’t, and it’s unbelievable how many sports pundits try to literally guess how it will play out. Winning streaks are simply just more noticeable patterns.

The same thing happens when you are betting. If you’ve won a lot of bets in a row you feel like you can’t lose. If you’ve lost a lot of bets in a row you feel like you can’t win. There isn’t a person alive (at least that I know of) who wins 13 out of 25 bets and then has a gut feeling that they will win exactly 17 out of their next 30 bets. The brain is built for picking up on more recognizable patterns and accepting this will virtually guarantee you to be a more knowledgeable sports fan and sports bettor.

Sunday, March 1, 2015

Sports Philosophy: Part 1

In the 2013 NFC championship game, the 49ers were driving down the field against the Seahawks, needing a touchdown to win the game and reach the Super Bowl. As they completed short pass after short pass, the anxiety of the home Seattle fans was building, and the FOX broadcasters were growing more and more animated with each play.

Any sane person would tell you that the 49ers chances of winning were increasing as they marched down the field. It makes sense: closer to the end zone=closer to scoring=closer to winning the game. The logic is sound. Completely inarguable. Or is it?

A supposedly insane mathematics genius (or just the win probability calculator on advancedfootballanalytics.com) would tell you that the 49ers chances of winning were actually decreasing as they moved down the field. Why? The time. The 49ers were running out of time so fast that 15 yard completions on first down were actually making them more likely to lose.  

The naked eye obviously has a hard time making win probability estimates or calculating the relevance of each play in complex games. It’s hard to tell if a 4-yard run on first down is a good play or if making 1 out of 2 free throws is a good offensive possession. Was the sacrifice bunt a productive play in this situation? How can we know for sure?

You may be asking yourself how this relates to sports betting, or merely to anything at all. Why should we want this information? How can this have any impact on our lives? We want the information so that we can actually know what we’re supposed to root for when we watch sports. And to further the point, this approach helps us make educated decisions if we’re going to make predictions and put our money where our mouths are.

Sports have more mathematics, probability, luck, etc. than just about anyone is willing to believe. But when you finally embrace this philosophy everything becomes clearer. Suddenly the game becomes a simple numbers problem and you have a great understanding of what is actually going on.

If you choose to believe in momentum or clutch genes or whatever else you picked up watching ESPN, you should at least be able to assign some value to it. Asking how much something is worth is absolutely essential to being a knowledgeable sports fan and sports bettor.

I recently had a debate with a friend about a game between the Cavaliers and Knicks in which the Cavs were huge favorites. He argued that the Cavs were a good bet that day because they had All-Stars and the Knicks team was full of scrubs. His analysis was spot on, but his conclusion was baseless. I agreed with him that the Cavs were great and the Knicks were terrible but contested that the difference in talent was worth far less than the 17 point spread implied. The best teams in the NBA outscore their opponents by 7-10 points per game (see standard point differential statistics) and 17-point win margins are only fair to assume in extreme cases.

Subjective reasoning can be useful at times, but only if you’re able to estimate its importance. You cannot simply pick against a team because their best player is injured and expect to be right. The value of that player needs to be properly accounted for and then compared to the game’s odds. This is where most people falter and why most sports predictions fail. They are founded on emotions and subjectivity and not data and facts. Our minds are always getting in the way when we think about sports, a point that I will expand on when this (let’s call it a lecture) continues.

Friday, February 27, 2015

February’s Teams of the Month: Utah Jazz and Philadelphia 76ers

If I had done this post at the end of last month, we’d be talking about the Boston Celtics. I’m not sure if a team can have as good of a month (for covering purposes) as the Celtics did in January. We backed the Celtics for each game of a ridiculous five game road trip the week of January 19th, in which they faced the Clippers, Blazers, Nuggets, Jazz, and Warriors. The Celtics managed a combined point differential of -1, covering in all five games and winning three outright.

This month, we have the Jazz (who went 4-1-1 in the six games we backed them) and the 76ers (who went 4-1 in their five games). My prediction for February’s team of the month was Minnesota, but they ended up going 6-4 ATS for us despite a couple of brutally close losses. The Sixers covering dominance makes sense, considering they are given insanely high point spreads every night, but the Jazz kind of came out of nowhere (we are counting a January 31st 10-point win over Golden State in these records because I’m in charge and it was really awesome).

So yea, Utah’s most notable win was that home thrashing of Golden State, although they managed similar performances against Portland and San Antonio this week. Enes Kanter’s departure to OKC seemed like a bad thing, but Rudy Gobert has really taken advantage of the extra playing time. Gobert, Gordon Hayward, and Derrick Favors form an excellent trio, and they’d probably be in the playoff mix if they had some competent role players (no, Trey Burke and Dante Exum do not fit this category). Linemakers seem to have finally caught on, however, making the Jazz favorites this weekend against the Nuggets and Bucks.

The Sixers have won three games outright this month as underdogs of seven points or more, (these were actually their only three wins, lucky us!) all at home and all by a margin of seven or more. Beating the Nuggets, Hornets, and Wizards was nice, but they also managed a five point home loss against the Warriors. All of this came without any players that anyone has ever heard of (with the exception of Jason Richardson… he’s still good, right?). All in all it was a great month to be a Sixers backer.

I’ll give a couple picks for my March team of the month prediction. We’ll go with the Wolves again, who still seem to be vastly underrated by linemakers and also the public. This is a much improved team since Rubio, Martin, and Pekovic all returned from extended injury absences and Wiggins has gotten much better as the season has progressed.

The Knicks have the worst ATS record in the NBA, but with so much New York bias and a miraculous win in Detroit Friday night, it’s hard to imagine their spreads will continue to be so inflated. That leaves the Nuggets, holders of the second worst ATS record in the NBA this season, who have been getting absolutely slaughtered lately. I expect the lines to become inflated for them over the next few weeks, giving them a good shot to be our March team of the month. Check back in 30 days to see how we did!

Wednesday, February 25, 2015

A Lesson in Probability Theory (and calculating fair moneylines)

Sports betting is math. You may be a dinosaur and refuse to believe that sports are about numbers (although numbers are quite literally how it’s determined which team wins), but there’s not much of an argument to be made that betting isn’t about math. A bet is worth making when the expected value is positive, meaning that the payout outweighs the risk. When a moneyline is +150, that implies a 40% chance of winning. If you were to make 10 bets at a +150 price and win 4 of them, you would be exactly even (4 wins x 1.5 = 6, 6 losses x 1 = -6, hence a net profit of 0). A bet at +150 is worth making if you have a greater than 40% chance of winning it.

A bet against a point spread is no different, but the moneyline odds are usually -110 rather than +150. A price of -110 implies a 52.4% chance of winning, so in order to have positive expected value on a point spread bet you need to have greater than a 52.4% chance of winning.

(Here is a simple tool to convert moneylines into their implied percentage counterparts)

I generally use a larger threshold than 52.4% (typically 55% or higher) to give myself some margin for error, but the point here is that you shouldn’t bet something just because you think you will win. You need to have a positive expected value. This is why you can’t simply bet favorites on the moneyline in all the games and expect to turn a profit. The underdogs win the games a certain percentage of the time due to randomness. Randomness is why a good foul shooter can’t make every single free throw, and why even the best hitters sometimes strike out.

The philosophy that the best sports gamblers employ is this: make as many wagers with positive expected value as you possibly can. The more times you expose yourself to positive expected value, the more likely you are to have long term success. This is due to something called the “law of large numbers” which dictates that the more times a fair coin is flipped, the more likely you are to see the breakdown of heads/tails approach a 50/50 split. This can be applied to sports in general, e.g. the more times a .300 hitter comes to the plate, the more likely you are to see him bat close to .300. In one day or week or even season, anything can happen. But over the course of an entire career, a .300 hitter is very likely to bat close to .300. In betting, this means that the more positive expectation bets we make, the more likely we are to see positive results overall. This methodology isn’t necessarily just for people who bet a lot. Even if you’re a casual gambler you still want to give yourself the best chance to win, right?

Sunday, February 22, 2015

The Introduction: Why (Almost) No one Wins at Sports Betting

(please see the "A Word From the Guy" section for a disclaimer)

Have you ever stood at a roulette table and tried to guess whether red or black was coming up next? Or tried to predict a series of coin flips? Or maybe even tried to predict if a dice roll would come up odd or even? If you have, then you’re basically the average sports bettor.

Let’s say you’re a professional coin flip picker. You’ll have big runs when you’re guessing right and also some huge cold spells when you’re guessing wrong. You’ll question the system you use and try to make adjustments. “Heads-Tails-Heads isn’t working? Time to switch to Heads-Heads-Tails! That ought to work!”

Maybe you’ll make some money and maybe you’ll lose some. You might even win a lot or lose a lot, or maybe just hover around even. And then you’ll try to tell your friends that you’re either a good or bad coin flip prognosticator. But in reality, coin flips are just a 50/50 proposition no matter how you try to spin it.
As ridiculous as trying to make money flipping coins sounds, it’s remarkably similar to how most people go about sports betting. They have their analyses and their breakdowns and their trends, but essentially they are just flipping coins.

Bookmakers generally know how to set proper point spreads, and the reasons people generally come up with for picking teams are almost as useless as the reasons for picking heads or tails. It’s essentially the same thing. If things like streaks, momentum, emotions, etc. actually had a predictable impact on the outcome of a game, don’t you think Vegas would know it too? Most people are too arrogant to believe that the guys who set the lines are exceptionally good at their jobs, and too clueless to realize that these line-setters have a significant advantage over just about anyone.

But sports betting is actually worse than coin flipping. In the coin flipping example, there are no odds or vig. You get your money back when you win, and the house doesn’t get a cut. In sports betting, however, the books set the standard price at -110, so that it costs you $11 to make a $10 bet. This extra dollar may seem insignificant, but would you really bet to win $10 on heads if you’d have to pay $11 every time tails came up?

With all of this knowledge, doesn’t it seem insane that anyone would bet on sports at all? You’re virtually guaranteed to lose in the long run with such a steep hill to climb, so why even try? The reason people do it though, aside from the entertainment factor, is that sports tend to seem predictable. You can always attribute reasons as to why a team won or lost, after the fact, and then try to use that hindsight going forward. In the long run, this doesn’t work. And having to bet $11 to win $10 makes it even harder because it means that you need to win 52.4% of your bets just to get your money back.

So how do you actually win at sports betting without merely getting lucky? It’s not impossible, and there are actually quite a few people out there who are doing it. Beginning with this post, I will be detailing the methods that are necessary to employ in order to win at sports betting in the long run. Stay tuned.

(Oh and of course there will be plenty of picks, because obviously that's what you're all here for!)