Biggest mismatch in Super Bowl according to DVOA

[Foghawk]

There are other more appropriate methods to make predictions regarding the outcome of football games.

Like ball pressure control?

Time of possession! :lol:

Well statistically speaking, if you can pick up on your opponents offensive tendencies from watching their practice film you should have a mathematical advantage. I don't imagine they track that stat though :229031_shrug:

 

[hawknation2015]

These types of threads are generally very embarrassing for people who take them too seriously, and they don't even realize it.

I doubt many of you have taken a stats course beyond intro stats, so I don't expect you to have knowledge of advanced statistics on a theoretical level. Suffice to say, that the application of advanced statistics in fields outside of physical sciences is something you have to refrain from taking too seriously, due to the sample size, lack of ability to replicate the experiment in a controlled manner, etc. Without context, and without taking into account schemes and game plans, these stats mean nothing.

Not to mention that overall metrics that take into account the entire year don't mean much at the moment anyways.

It'd be great if people stopped misusing stats, but clearly that isn't going to happen any time soon.

So you want take the Patriots victory on faith then? Because.. what else is there?

No. There are other more appropriate methods to make predictions regarding the outcome of football games. Statistical analysis helps when you thousands of predictions over the course of the entire year. In the presence of statistical analysis, you're more likely to be correct in most of those games given that wins and offense/defense stats will converge to their expected value over the course of a season. In a one game showdown, they mean nothing, however.

You probably also think that Brownian Motion is good model for the evolution of stock prices.

Like ball pressure control?

Or filming opponents' signals?

 

[hawkfan68]

These types of threads are generally very embarrassing for people who take them too seriously, and they don't even realize it.

I doubt many of you have taken a stats course beyond intro stats, so I don't expect you to have knowledge of advanced statistics on a theoretical level. Suffice to say, that the application of advanced statistics in fields outside of physical sciences is something you have to refrain from taking too seriously, due to the sample size, lack of ability to replicate the experiment in a controlled manner, etc. Without context, and without taking into account schemes and game plans, these stats mean nothing.

Not to mention that overall metrics that take into account the entire year don't mean much at the moment anyways.

It'd be great if people stopped misusing stats, but clearly that isn't going to happen any time soon.

I haven't read so many words that said absolutely nothing since I was forced to read Tess of the D'Urbevilles in high school and I HATED that book. For all your ranting, you made exactly no comment on the actual subject, just criticized the use of statistics by "people" and subjected us to what really amounts to a nerdy tantrum.

Unfortunately for you, statistics is defined as the study, collection, interpretation, presentation and organization of data. Thus, I give you, data as accrued throughout a season of football played by 32 teams each wit

h a 53 man starting roster where all play must fall within controlled parameters as set by the NFL Competition Committee. Any variety in the resultant outcome of the plays would be due to variables in the application of the rules and the players' abilities.

In short, it'd be great if people stopped saying stupid things, but that clearly isn't going to happen any time soon.

You're clearly one of those who has never stepped foot into an advanced statistics course.

Any model, irrespective of how good or realistic, requires quality data to be fed into it. For data to be of high quality, you need multiple samples of sufficient size, and you need to be able to replicated each sample in a controlled environment. You can't take an NFL game and replicate the same outcome and then extrapolate anything meaningful from that.

This is a very interesting point. The sample size of an NFL season is too small to derive any predictive meaning from past results. Seattle has only taken part in 2,400 plays or something. You can't take anything from that.

However, it also means you haven't seen enough plays to make any subjective judgments as to the merits of the 2 teams. Conclusion - we will have to stop talking about the next game as we can't make any informed judgment on it.

I kind of thought this guy was a wanker, but he's only made 89 posts so I can't tell from that sample size.

Really? It was clear to me that he/she is a wanker from his/her second post onwards....

 

[Jacknut16]

These types of threads are generally very embarrassing for people who take them too seriously, and they don't even realize it.

I doubt many of you have taken a stats course beyond intro stats, so I don't expect you to have knowledge of advanced statistics on a theoretical level. Suffice to say, that the application of advanced statistics in fields outside of physical sciences is something you have to refrain from taking too seriously, due to the sample size, lack of ability to replicate the experiment in a controlled manner, etc. Without context, and without taking into account schemes and game plans, these stats mean nothing.

Not to mention that overall metrics that take into account the entire year don't mean much at the moment anyways.

It'd be great if people stopped misusing stats, but clearly that isn't going to happen any time soon.

IF, IF you were legit you would have given an example, maybe even two. You didnt, and I dont think you understand the track record that DVOA has, or even understand what was being said in the first place?

First thing: What do you have an issue with, his using DVOA? Or something he said about it?

Otherwise your post was MORE embarrassing than the Original post.

 

[Jacknut16]

.

What does that even mean? "Believe in randomness."

There's an entire field of mathematics devoted to the topic of randomness called probability and stochastic processes. I'm sure you've heard of it, but wouldn't be surprised if you haven't.

Randomness is just the lack of a pattern or the ability to individually predict events. I'm glad you brought it up, because the outcome of a football game is largely random.

You still havent addressed the argument, completely unacceptable.

What is it you have a problem with?

DVOA has performed really, really well.

As a person who claims to have a mathematics degree you sure do do a terrible job at defining your position, not buying it for a minute.

Until you do and make a logical argument, you'll get classified as a college kid who took a couple statistic classes and is still not quite sure what he was being taught.

Application, Application!