Mad Dog":3079wdfk said:
austinslater25":3079wdfk said:
Because the two greatest coaches in the game haven't done it historically isn't a good argument. They can be(and often are)wrong to do so. We just have much more information available now that we know it probably makes more sense, in most instances(not all) to go for it inside the 30's than to punt. We punted from the 37 at one point on 4th and 1 that netted us 17 yards. That doesn't make any sense at all. None. I could go more into this but people are camped on both sides and no one is going to change their opinion on this so it's probably best to agree to disagree.....
And the analytics people can also be wrong. Because the lack of data points in football necessarily means the analytics are flawed.
In medicine, if i want to prove a treatment is 5% better than an old treatment, I need to run a study that randomizes about 1000 patients to statistically prove it. And even that runs a 5% error rate of being wrong and that the differences seen were due to chance. With Football and all its chaos and variables and low number of data points, the statistics are just trash and highly variable.
The analytics guys never give you their confidence limits. You can say that a certain action on 4th down offers you 3.2 more points per game than a different action but what they don't tell you is that its 3.2 points +/- 4.7 points. Which means any action could gain you as many as 9 points or cost you as many as 2 points and such wide variability means you absolutely should use more info than just a number to make a decision.
I can't disagree with the one point by austinslater25, that an appeal to authority, even the authority of almighty Bill Belichick, is just that, an appeal to authority. I raised it half in fun and half because you will rarely find an analytics follower (certainly not a Seahawks analytics follower) who will critique Belichick on football decisions the way they hack on Pete Carroll. But that's not the meat of the argument. It's only a warning/indicator that the very best coaches and organizations in an ultra competitive $20 billion industry, with access to sophisticated professional analytics and to vastly more comprehensive data, are not obeying the "laws" of popular football analytics. Just maybe they are doing it for very good reasons rather than because they are "old fashioned".
Mad Dog lays out a really serious and persuasive case for why. Others with quantitative backgrounds have also written deeply challenging things, scattered here and there on forums and blog sites, about the limits of football analytics generally, the expected points framework, diminishing returns of this or that "efficiency", and so on. But one of the most disturbing things about popular football analytics is that its followers have shown basically zero interest in any of this methodlogical debate and more or less ignore all challengers. Post up a critique and you will just get silence from the likes of Baldwin. In the meantime we are left wondering what is the validity of any of these claims to football knowledge? More like modern medical science or more like nineteenth century leeching?
I'm all for what a few others have said here -- up the middle so to speak -- in terms of taking up whatever statistical research there is on specific, isolated football situation if it is at least somewhat amenable to quantitative study. But even there, this provides nothing more than a loose guide. Don't surrender the decision to a machine or worse yet an economist! Coaches simply have more relevant and complete information of the given situation on a given day. Coaches also have biases, yes, but hopefully data and collective input can help to counter the worst elements of these. Still, until I hear a valid football explanation, beyond "luck" and "Russ is just so good", for how Carroll could defy all odds with his 55-0 record nursing a lead then I personally will give him quite a bit of slack for erring on the side of caution, at least when the game is firmly in hand.
When it comes to the more grandiose, sweeping claims coming from some of these guys telling coaches to work from some cookie cutter template based on "optimal" run-pass ratios and so on -- at that point, sorry,, they have exceeded any and all authority. Many of these guys don't even understand the basics of football, and in the Seahawks Twitter world, when film study guys like Matty Brown take the time to correct faulty notions using concrete in game evidence, typically again the response is a deafening silence. Any historian of football should know there is no single way to win, no template of the time. Everything has a counter and football trends come in cycles.