Top Sportsbooks
Understanding 'Luck' and Making It Work For You
Two Sundays ago, during the end of the Dallas-Philadelphia game, the Cowboys
were down 31-24 and driving into Eagle territory with about 30 seconds on the
clock. Though Dallas were long shots to score a TD and tie things up, those
"wise guys" sitting on the teaser of Dallas +1.5 up to +7.5 were already
counting their winnings.
With Dallas out of timeouts, a few more incomplete passes, a touchdown, or even
getting tackled in bounds would end the game and give many of these players a
chance to cash at the window with a smile on their face. Then the unthinkable
happened!!! A Dallas pass in the end zone was intercepted and returned 102 yards
for a Philadelphia touchdown. Whether you call this a highly improbable
occurrence, or just plain bad luck, either way all those Dallas teasers bit the
dust and those wise guys lost out.
While a bad-beat like that may leave a horrible taste in the mouth, it's
surprising how quickly players forget about the impact of luck on a team's
performance. Players might later review Dallas's schedule and note the Cowboy's
win/loss record, or the points for/against, but miss the subtle nuances of
outrageous misfortune - like the 102 yard TD interception.
In a very long season - such as the NBA or MLB - the good and bad breaks tend to
balance each other out. After 30-40 games, you get a fair measure of a team's
ability. It's more dangerous to ignore luck's affects on an NFL team due to the
short season; most successful bettors appreciate how luck adds another dimension
to a team's win/loss record.
There are several occasions each NFL season where the better football team does
not win because of a lucky or unlucky play that defies reason. We've all seen
miracles and disasters cause small fortunes to change hands between bookies and
bettors, but still many players ignore this factor when evaluating teams.
In the same way that players perceive a pattern at the roulette table when the
ball lands on black for six consecutive spins (there's actually no pattern, just
a sequence of random events), many players try to deduce a "pattern" in lucky or
unlucky teams. However, the results are better explained by chance.
How do handicappers decide if a result is due solely to luck or from the quality
of a team that finds a way to win? Many professional players do extensive
statistical testing of a single factor that might be heavily affected by luck.
This approach determines whether a team that has shown a certain tendency in the
past (such as throwing many interceptions) is likely to continue in the future.
The "bean counters" do this with correlation tests, trying to validate (or
refute) the idea that "bad teams throw more interceptions" by looking at the
results of teams in prior seasons. Do bad teams throw more interceptions, or are
there simply unlucky events that make a team "look bad"? This question can be
answered even if you have no idea what a "correlation test" is.
First off, identify the factor you want to test for luck - in this case,
interceptions. Look at league-wide statistics for the first eight weeks of the
prior season, to see which teams had the most, and least interceptions. Divide
all NFL teams between "good" and "bad" based on interceptions thrown. If they
threw more than average, put them in the "bad" group, and otherwise in the
"good" stack.
Then utilize the same process while examining the last eight or nine weeks. If
the same 16 teams are bad in both halves, it would suggest a very high
correlation, meaning that a team that throws picks will probably keep doing so.
If exactly half of the "bad teams" from the first half of the season are bad in
the second half (regarding interceptions), it would suggest the event is purely
random (since a "bad" team in the first half of the season is equally likely to
be good or bad in the second).
Practicing this test on NFL teams, will show that interceptions thrown by a team
are almost completely random. In fact, you'd also find similar results for all
forms of giveaways and takeaways - there's less than a 20% correlation. So how
do bettors turn this knowledge into money? Simply recognize when teams are
overrated or underrated because of the impact on perception of good or bad luck.
A team with a high turnover differential looks good in a box-score, but it's
just as likely to be positive as negative on turnovers going forward. Historical
scores do not reflect that, and make these teams (like St. Louis) look much
stronger than they actually are. Similarly, teams with negative turnover
differentials (like Oakland and Cleveland) tend to outperform the public's
expectations (and hence, the spread).
There are other calculations bettors can do that will give a statistical edge in
sports betting. Identifying results that are unusual due to luck will put you in
position to evaluate teams better than the market as a whole. With any statistic
you can isolate and analyze - from the Colts' 3rd-down conversion rate to
anomalies in field goal kicking - if you can recognize luck, you can make money
off it.
Real statistics geeks can go further with the luck factors identified. Advanced
correlation can tell you how much difference a lucky turnover change makes to an
NFL game (it's actually about 3.5 points). Or, a hit batter in baseball, which
accounts for about 0.5 runs. If you can quantify how much these lucky/unlucky
instances change the event outcomes, you can "adjust" your statistics
accordingly and gain a better measure of what truly happened. If your assessment
is more accurate than the market, you will profit.
What are our players betting?
Miami -17.5 -109 at Duke
In Miami's game against FIU last week, a brawl broke out resulting in the
suspension of 13 Miami players. We thought the opener of Miami -20.5 accurately
reflected these suspensions, but the sharps disagreed. The wise guys fired away
on Duke at +20.5 with some buying it up to +21.5. Despite the sharp action on
Duke, we are seeing many more players backing Miami.
Texas -6.5 -105 at Nebraska
The initial opener of Texas -8 drew a flurry of sharp activity, all of which was
on Nebraska. Most of the public is on Texas and there are three times as many
wagers on the Longhorns, including most of our larger Asian players. This game
has been heavily traded, with the sharps refusing to allow the line to drift to
Texas -7.
San Diego Chargers -5 -103 at Kansas City Chiefs
The sharps were split evenly on the opener of Chargers -4.5 with some sharps
buying the Chargers down to -3.5. When we see this type of activity, one of two
things must be true - our push percentages are wrong or their numbers are wrong.
Although there's fairly balanced action in terms of volume, the public clearly
favors the Chargers as we've taken nearly 20 times as many bets on San Diego as
Kansas City.
Carolina Panthers +3 +105 at Cincinnati Bengals
We opened the Panthers at +3.5 (-112) and saw a fairly rare event - the sharps
AND the public backing the same side. With nearly three times as many bets on
Carolina as Cincinnati, this price has drifted down. Similar to the San Diego
game, we're seeing some of our sharper players buy through the "4".
