Okay, blame it on an early Xmas break for me to start this topics and get a break from Steelers game talk ad nauseam... "The Pats were the best 1-15 team in the NFL" (Seriously, I vaguely remember reading something like this about the Rod Rust Pats... The honor now should go to the 0-12 Dolphins) "The Pats were the worst Superbowl champion in the history" (2001 Pats) "Sometime, the better team doesn't win" (you-know-who) Premise A: You are what your record says (Bill Parcells) - TRUE Premise B: "The record doesn't say it all" or "The number doesn't tell all" (whoever) - TRUE If we accept Premise A, why bother with statements like those? Are all current 6-6 teams in the NFL equal? In our life, we encounter failures, but do we accept them at face values, or do we think "that is not the real me and I'm really better than that"? There is a real scientific concept behind this and forgive me, those experts out there, it has to do with "scientific inference". Premise A is about evidence, fact, and data. Premise B is about what lies underneath that generates that evidence, fact and data. If a team is say, 8-4, what else is there to talk? It is what it is. 8-4 is 8-4. But is it "SB-bound" 8-4 or "fraudulent, one and done" 8-4? It is scientifically valid to ask the second question, which is about the underlying capability. Too often, people counter an argument by saying, you are 2-4... or you just win a 4-7 team by 3 lousy points... The other side say, yeah but that's an aberration. On and on waste of time and saliva. Is there a way to resolve this? Yes. Definitely. When you make a phone call these days, chances are you get a computer answer "OK, let me try to help you... Please say... ". Automated call center like this is a multi-$billion business. Some cell phones have voice/speech recognition (Motorola...), say the name and it makes the call. More than that. Audio analytics and video analytics are trying to determine your emotion. Not only do these call centers computers recognize what you say, they can determine the state of your emotion, especially annoyance and anger. Try to get angry with a computer and you'll get a live person right away: it is programmed to detect anger and alert a human manager so that they don't lose a customer. For the same script "What the hell is this?" you can say it with intense anger, with light-hearted humor, or with boredom... A person can be offended or know right away that you are facetious. If reading the script and literal meaning is all we humans care, actors and actresses would not be professional. Why pays Eddie Murphy to be the voice of Donkey, or Mike Myers the voice of Shrek... Any Joe's on street can say those lines. This technology (voice recognition and audio analytics) is made feasible by the same desire that football fans try to see if their 8-4 team is SB-caliber or one-and-done type. It's about determining what lies underneath the evidence and record; the true capability of a team. One such scientific method is called Hidden Markov Model (HMM). It's about how best to infer the true or innate capability, given the evidence. It allows determining the spoken words and emotion (underlying true value) from speech and voice (evidence & record). The multi-$B business of call-center audio analytics software and the wide-ranging speech recognition application is a testament to this scientific success. In a "softer" science that is very popular, it's called "profiling" that is familiar to afficionados of CSI and forensics stuffs. In more serious applications, it's about the inference of terroristic intents or potential threats. Humans and animals are evolution-wise hardwired with some HMM inference software as well. It's about subconscious inferring the innate capability of a person based on evidence that we previously encounter with that group of people, which sometime we call "prejudice". To some extend, although much more difficult, my colleagues apply this method to determine team capability. This is different from those so-called hard fact and quasi-statistic analysis. So the answer is... No, you are not exactly what the record says you are. But the record does say a lot -although not all, about you. It's fun and fair to try determining what a team capability really is. It has true predictive value. So-called hard fact and quasi-statistic analysis is just a description of facts, they have no predictive basis and value. Take any poster on this board and collect all his/her post. They do say a lot about that poster. But not all about that poster, Do they?