ANALYTICS TWITTER: Won every battle, but was the war a pyrrhic victory?

To name something is to damn it. The problem with "Analytics Twitter" is it arguably exists, whereas the superpower of "Anti-Analytics Twitter" is that it arguably doesn't exist. It gets to fight guerilla warfare whereas analytics is stuck in rank-and-file.

Some random person posts a chart they cooked up with public NFL data and a coding exercise, posts an inflammatory comment to draw attention, and anti-analytics people are quick to characterize this person as "analytics twitter" or "spreadsheets FC" to make a point about the growing use of analytics in the sport and something they are being left out of. The criticisms are generally fair: the sport is played on grass, there is a psychological element, the models are imprecise, there are other factors, the film reveals other truths.

The latter is interesting, because when someone posts a highlight clip, identifies something they don't understand, makes an erroneous judgement, nobody is there to dunk on "anti-analytics twitter" for all of the valid criticisms of film review: information gap, biases and prejudices, tiny sample sizes, selective highlight reels, etc.

Every year, fandoms pick their favorite players and narratives, right at their peak before regression. The experience of an educated football fan is wading through all this noise and seeing the results a mile away. Its pathetically easy for a computer to dominate fantasy leagues and gambling games against the average casual. Its pathetically easy for a "film junkie" or otherwise traditionally-educated sport aficionado to do the same contributing to this perspective of the analytics being unnecessary. The computer and the aficionado dominate in different ways though. They have their different blind spots, just much fewer than the average casual fan.

There is a pernicious fallacy: something that exists is compared to perfection while the contextual background of existence is not semantically categorized and thus incomparable. All of these models and stats and numbers have the same weaknesses as discussed. Those aren't criticism of the objects though (models, stats, charts, analytics, etc), as those objects are factual. Those are criticisms of *interpretation*. The process of churning immense information into insights, predictions, rankings, explanations, etc, all relies on the same facts of what occurs.

This particularly hypocritical attitude is from the fact that every fan loves stats. Winning is a stat. Touchdowns, sacks, yards, missed blocks, dropped passes, all of these are stats. Stats are just collections of facts. We easily derive that a player missing time will have less stats than other players, so we create things like "yards per carry" or "sacks per game". As soon as we employ good stats like "expected points added per play", now the mere existence of statistical uncertainty is a foundational issue. The reality is that the average casual viewer would have a better understanding of the *total state* of the NFL from entirely accepting valid statistical analysis and never watching the sport. Film review has many of its own interpretation obstacles.

Per the latter point, analytics is often really good at processing large sample sizes and building a larger league-wide narrative, lending itself well for things like betting or predictions. Being a fan of a team often includes engaging is week-to-week narratives, tiny sample sizes, and looking at the league through a singular perspective. Analytics generally isn't very good at generating insights into specific teams, games, and situations for sample size concerns. The type of factors that can affect a singular game (missed kicked, blown call, dropped pass, bad playcall, etc) are smoothed out over large sample sizes in unintuitive ways for the average fan. Those being very emotional narratives for people that analytics can outright ignore (for good and unintuitive reasons). Another reason for them to throw the baby out with the bathwater.

The largest problem with the anti-analytics perspective is its unverifiable. You can basically construct any narrative a priori and find evidence to support it. Film review becomes a Rorschach test for the viewer. Its such an incredibly complex and *secret* sport that leaves ambiguity in every play. Should the quarterback have attempted the throw? Should the receiver have ran a different route? What was the playcall, the protection, the reads? These are unknowable. Football education immensely helps untangle reality, but there is an impenetrable wall of lost information. People fill that void with their own, often contradictory perspectives. Every year, the worst quarterbacks you've ever watched has die-hard fans either upholding statistics and highlights to reveal the truth, or admonishing statistics and rejecting lowlights in defense. Its all a silly charade while they take shots at the ontology of numerical knowledge.

This void of information still exists in the statistics, and is often addressed as a damning critique of analytics, despite being an even greater damning critique of film review. A singular play can have things the analytics missed, something the viewer assumes will linearly extrapolate through the collection of plays being analyzed while unaware of the things they are missing without analytics: the unknown unknowns. A human brain cannot remember the nuances and intricacies of tens of thousands of plays necessary to build accurate league-wide narratives, but our brains lie to us. The power of statistics is in this analysis of huge sample sizes past the scope of cognition, revealing much more sensitive and ephemeral effects in the sport often lost to the viewer. These effects are easily dismissed through the paradigm of analytics-rejection.

This is a sad reality for the average viewer. Football is so unreachably complex and nuanced that leave the viewers on the outside looking in, unable to connect to their passion to the scope they would wish to. When an entirely new paradigm appears (analytics) with its own nuance, terms of engagement, education gap, and unintuitive results, it can be ostracizing to the casual fan. Nobody wants to hear "you are wrong for reasons you don't understand", and the mirage of knowability on the TV screen compounds the unintuitive nature.

Those in this situation can see as other casual fans, hobbyists, students, or even seasoned analysts, get things wrong. Its easy, and they are personally incentivized, to encourage this narrative of being wrong *because* of the analytics and not because of the interpretation barrier. Its easy to project this event onto the emergent group of "analytics twitter". Despite the stupidest takes you've ever read coming from the average jackass, there is no group to assign their error to. You can't attribute the systemic issues to the paradigms to a convinced audience if you can't identify *who* employs those paradigms. Thus, people attribute the errors of some undergrad's hobby project to the analysts sitting in NFL booths, whereas the media parrots are just wrong for individual reasons alone (often accused of "not watching the game" LOL).

In the background of this has been a decisive victory for the analytics revolution in the NFL. Teams are attempting much more 4th downs, earlier in games, even on their own side of the field. Teams are going for 2 point conversions down multiple scores. Linebackers and running backs have dropped in their average allocated capital. NFL teams are employing analysts, researchers, and math PhDs, even putting them on headsets in the game. "Analytics twitter" now includes NFL head coaches, but their critics still treat "analytics twitter" as the outsiders. There is still an incredibly common framing about whether NFL coaches "should" adopt analytics as part of their process. They did a decade ago. A question is no longer whether a coach should do what the "analysts" want, its whether a coach should do what *other coaches* are doing. Fans just get to toss tomatoes from their couches.

The whole issue is a result of the semantics of the culture. Since "anti-analytics twitter" doesn't exist, you can't attribute them. You can't identify them and critique their methods. I despise the godawful, actively anti-intellectual, venture-capitalized state of pop analysis. The average fan does too, but they don't see a relation between their anti-intellectual approach and participation in the culture that fuels the pop-analysis. Its self-inflicted, folks. People don't want to accept hard realities of their team and the sport as whole, instead throwing the baby out with the bathwater since they have nothing to lose.