I, like many of you, love using stats and numbers to make wagering decisions. Whether we're sports betting or horse race betting we're numbers people.
But, it always strikes me how much nuance and feel we need to place on the numbers to make decisions to properly try and find our edge.
We know trainer "X" is 34% off the claim on a small sample, but how the horse's win is our true trigger. Did they get perfect trips and win? Did they all win by dropping in class? We need to see them improve in our mind's eye to truly know how much we should hammer (or fade) the next one.
In football it's not dissimilar. I listen to Rob Pizzola's podcast and he loves numbers, he's a data guy. But, like in horse racing the fact remains, there's not much edge on pure numbers. Over/under bettors see a number of 47 and will likely have a model spitting out 47 or 48 or 46. Rob will dig deeper and use a bit of the qualitative to make a decision.
What does he think game flow will be? Will possessions be limited this week because of an injury? Two years ago when faced with the same game situation did the coach have a tendency to do something for the edge? What's not in the numbers is very important.
I haven't even touched how some data can just well, lie.
Trainer win percentage from top barns? The top barns get the top horses because a trainer is good. They are jamming nice horses in short fields. How can they not win at these rates? And they're almost impossible to separate. These numbers are self-fulfilling.
Crazy stats like QB wins. Apropos of nothing really, other than the team the QB plays for. Dan Marino's win loss record was an indicator of his prodigious talent about as much as a cheese sandwich he ate last week.
Looking at game winning drives. Aaron Rodgers' historical numbers aren't great. Do we really believe those? You and I would take him in a heartbeat, right? Game winning drives are often a function of team defense.
Kirk Cousins brought Minnesota back with a nice drive to go up 31-24 with 2 minutes left yesterday. His game winning drive was lost in one play, when Aaron threw a beautiful TD pass. Then Minnesota scored and Cousins got a game winning drive stat back, the one he had two minutes earlier (that he might not have gotten if a Green Bay DB could hold on to an INT). The clock ran out, so Rodgers could not get a Game Tieing Drive or a Game Winning Drive stat on this day.
Stats can lie, too, because the underlying stat can be spurious.
Quarterbacks against the blitz stats are often used on telecasts as some sort of massive talent predictor. This is mostly a function of offensive line play (unless you're a Mahomes). With a great OL you can get so many more explosive plays from a blitz when it gets picked up by this solid unit.
The layering of numbers - for those of us who did or do database handicap - is always pretty nasty. If a horse wins we will always find more reasons a horse won, simply because he won! If we're dealing with only 27 data points, we're in real trouble and can make some serious mistakes.
Many nuances of statistics can be modeled, and smart people do that every day. However, that's well-beyond my pay grade - it took me awhile to figure out proper SQL syntax in jcapper, and kids today would lol at me in general. So, for many of us regular people who like numbers to handicap and wager, we have to do our best to understand what they mean. That involves us being smart and leaning on what the numbers tell us, but also being curious and skeptical; analyzing them, and yes, using a bit of feel.
Notes:
Wagering experts employed by the horse racing industry? Many of us think it's a good idea.
I loved seeing this clip from NFL.com about Elinor Penna. Elinor is known in horse racing as the spouse of Angel, but for her football fans she is Elinor the ground breaking, tough as nails and most-wonderful football mad sportswriter and columnist. She's super.
Walner's brother raced this weekend. This wasn't a maiden, but a high level stakes race. Incredible move. Perfect gaited colt that I hope we see over here next year.
No comments:
Post a Comment