Statistics tell the whole story. Independently verified?

I don't know if he is working from the same basic "recruiting playbook" that he employed during his first tenure at Kansas State, but I have a feeling that Kansas State may represent a true anomaly or outlier, based on this methodology. Snyder routinely recruited anywhere from one-third to almost half of his signees from the JUCO ranks. Even now, JUCO recruiting receives short shrift, compared to the ranking of high school seniors. I imagine that data was even less reliable for Snyder's earlier, JUCO-heavy years.
 
  • Like
Reactions: 1 person
Very good read. Like some I feel we can finish anywhere from 8-4 to 4-8. I think that the WKU game is vital. If we win that game by 3 TD's we will have great momentum going into the Oregon game. If this becomes a Troy game this team may begin to doubt and our fans will panic. The whole mood of this season will change. I think we win big and the week of Oregon game begins a good week of smack between fans.
 
Somewhere buried on here, there are a few threads from a year or so ago bearing the name "statistics tell the whole story, 1,2 and 3". The evolution of this analysis ends right about where cfbmatrix.com picks up. Go there, he has some great articles and explanations of the fully developed matrix.

http://www.volnation.com/forum/tenn...382-statistics-tell-story-who-should-win.html

http://www.volnation.com/forum/tennessee-vols-football/177540-statistics-tell-story-ii.html

http://www.volnation.com/forum/tennessee-vols-football/185532-statistics-tell-story-iii.html
 
  • Like
Reactions: 1 person
Which is proven by Microsoft vs Apple, right? Oh wait... the company with great marketing holds over 80% of market share while the company with the superior product does not...

And that is hardly a unique example. Statistics tell marketing people what moves consumers... and it works. What you said is (and I do apologize for being blunt) abjectly stupid. If marketing did not work... GM would probably be out of business because they produced inferior vehicles for the better part of 25 years.

Another great example was the stupid move by the former Google exec to change JC Penney's marketing approach. He ignored his demographic and what was proven to move that demographic. He stopped coupons and sales and hired a lesbian to be one of his main spokespeople. While much of the country does not care about that last tidbit... the unique demographic that made JCP money for years does- married with kids and generally thought of as "traditional" families.

I will remember that next time we make money while our competition does not because we have better data and statistical evaluations than any of them. And we use them in a commodity market which has a diverse set of factors.

Past performance does not "guarantee" an outcome but it does help determine the probability of something happening in the future. You are on average going to position yourself better if you have more information and have collated it effectively- statistics. That is especially true when the thing you are trying to predict has to do with human skills or behaviors. People ARE predictable.



If I misunderstood your intent then I apologize. But if you mean what you said... then I did not misunderstand.


Ok, let me address your point, then I will elaborate mine. Marketing, in it's most general form, is a scam to put it simply.

In the past decade, "Marketing" has emerged as a college major; where before a career in marketing would start with a business degree. These Marketing majors have no skill in business; that is important. The only math skills they have are very basic. We are talking basic Algebra skills.

So they will end up doing one of two things:

1) They will be charged with gathering information. Who buys our product? Age groups? Income? Gender? Race? There is nothing wrong with this and it is generally useful information. It is also a job that would be better equipped to a math or business major.

2) They will be charged with coming up with the next big thing in advertising or some other useless waste of money like "Where does the consumer look at our packaging first?"

Marketing in this sense is focused on convincing the business that the consumer is a sheep that is easily manipulated.

But I am pretty sure the type of Marketing you are referring to is the first, and I have no argument there. Using cut and dry resources like demographics to focus on efficient advertising and such is just part of business. Marketing majors generally have no skill in business. It is a catch-22, and it is the reason marketing is a scam.

But that is an opinion, and if you choose to believe that the exact size of the packaging, to the intense study of what cartoon mascot should be on the brand is worthy of the billions of dollars paid to mostly incompetent Marketing majors then I guess we will have to agree to disagree.

Respond to that if you want, but in truth I would like to return to football.

--------------------

Now, the field of statistics is very broad, and covers many bases. But when it comes down to it, the word that more accurately describes what you are doing with these predictions is probability.

Probability= favorable/possible (Atleast I hope it still does, I have been out of the classroom for a while:))

This is easy to demonstrate. favorable/ possible. So if I am flipping a coin and want heads, the probability is 1/2.

That probability is exact. Our two options are mutually exclusive (important), so I can say with certainty that the probability of winning (heads) is 1/2, or 50%.

Now let's switch to a football game. We once again have two options (omitting ties), which are still mutually exclusive. In elementary school if you were asked to find the probability that Team A wins, it would still be 1/2. favorable/possible.

But we know this to be untrue. Here is where it falls apart:

There are an infinite amount of factors at play here. We can play the odds around in a game of poker. After all, there are 52 cards that must be arranged in a finite number of ways. So in a five card hand, there are 2,598,960 possibilities. So even though that is a big number, we can still use our formula. So assuming I want a "pair", that is 1,098,240/2,598,960, or roughly 42%.

Just like with the coin, that is precise.

In football, we have an infinite amount of factors which leads to an immeasurable probability (favorable/infinity). Not only that, most of these factors are not mutually exclusive with each other. Which is just icing on the cake considering we couldn't measure them anyway, but now the probability that some factor A happens, has no effect on factor B.

So what happens is, you are left to sift through these factors and determine which are important and which are not. And we are talking about anything here.

So I could give you a statistical analysis that says APSU will beat UT by 40 points using very real stats and variables. All I have to do is use the stats or factors that favor what I want and ignore the ones that don't, producing incredibly skewed statistics that are mathematically supported.

Well, that's it. My longest post on this forum. I'm sure I rambled in there, but those last few paragraphs were my main point.

That is what I mean when I say that simple predictions from ESPN analysts are no less accurate than a statistical analysis.
 
Ok, let me address your point, then I will elaborate mine. Marketing, in it's most general form, is a scam to put it simply.

In the past decade, "Marketing" has emerged as a college major; where before a career in marketing would start with a business degree. These Marketing majors have no skill in business; that is important. The only math skills they have are very basic. We are talking basic Algebra skills.

So they will end up doing one of two things:

1) They will be charged with gathering information. Who buys our product? Age groups? Income? Gender? Race? There is nothing wrong with this and it is generally useful information. It is also a job that would be better equipped to a math or business major.

2) They will be charged with coming up with the next big thing in advertising or some other useless waste of money like "Where does the consumer look at our packaging first?"

Marketing in this sense is focused on convincing the business that the consumer is a sheep that is easily manipulated.

But I am pretty sure the type of Marketing you are referring to is the first, and I have no argument there. Using cut and dry resources like demographics to focus on efficient advertising and such is just part of business. Marketing majors generally have no skill in business. It is a catch-22, and it is the reason marketing is a scam.

But that is an opinion, and if you choose to believe that the exact size of the packaging, to the intense study of what cartoon mascot should be on the brand is worthy of the billions of dollars paid to mostly incompetent Marketing majors then I guess we will have to agree to disagree.

Respond to that if you want, but in truth I would like to return to football.

--------------------

Now, the field of statistics is very broad, and covers many bases. But when it comes down to it, the word that more accurately describes what you are doing with these predictions is probability.

Probability= favorable/possible (Atleast I hope it still does, I have been out of the classroom for a while:))

This is easy to demonstrate. favorable/ possible. So if I am flipping a coin and want heads, the probability is 1/2.

That probability is exact. Our two options are mutually exclusive (important), so I can say with certainty that the probability of winning (heads) is 1/2, or 50%.

Now let's switch to a football game. We once again have two options (omitting ties), which are still mutually exclusive. In elementary school if you were asked to find the probability that Team A wins, it would still be 1/2. favorable/possible.

But we know this to be untrue. Here is where it falls apart:

There are an infinite amount of factors at play here. We can play the odds around in a game of poker. After all, there are 52 cards that must be arranged in a finite number of ways. So in a five card hand, there are 2,598,960 possibilities. So even though that is a big number, we can still use our formula. So assuming I want a "pair", that is 1,098,240/2,598,960, or roughly 42%.

Just like with the coin, that is precise.

In football, we have an infinite amount of factors which leads to an immeasurable probability (favorable/infinity). Not only that, most of these factors are not mutually exclusive with each other. Which is just icing on the cake considering we couldn't measure them anyway, but now the probability that some factor A happens, has no effect on factor B.

So what happens is, you are left to sift through these factors and determine which are important and which are not. And we are talking about anything here.

So I could give you a statistical analysis that says APSU will beat UT by 40 points using very real stats and variables. All I have to do is use the stats or factors that favor what I want and ignore the ones that don't, producing incredibly skewed statistics that are mathematically supported.

Well, that's it. My longest post on this forum. I'm sure I rambled in there, but those last few paragraphs were my main point.

That is what I mean when I say that simple predictions from ESPN analysts are no less accurate than a statistical analysis.

Believe whatever you want, I am not here to force you to some conclusion. The severe irony in all of this is that you are formulating an analysis with nothing but your own experiences, which is why some data models fail.

In other words, you are extrapolating from the specific to the general. That would be like me saying that I see a dog with fleas, thus all dogs must have fleas. Do you see the weakness in that position? What you are ultimately saying is that "data models fail, so let me give you a model with no data as a way to prove that."
 
Believe whatever you want, I am not here to force you to some conclusion. The severe irony in all of this is that you are formulating an analysis with nothing but your own experiences, which is why some data models fail.

In other words, you are extrapolating from the specific to the general. That would be like me saying that I see a dog with fleas, thus all dogs must have fleas. Do you see the weakness in that position? What you are ultimately saying is that "data models fail, so let me give you a model with no data as a way to prove that."

That isn't what I did at all. The only specific examples I gave was the coin and the cards. In these examples, the data model does not break down. The numbers I gave are exact, I didn't draw cards 2.5 million times. They are not based on my experiences.

It is very simple. The basis of your predictions involve using probability (based on past events) to predict the future.

In real world application with infinite possibilities, these laws of probability are easily manipulated.

You are right about the irony, however.

You fail to see that my 40 pt APSU victory is no less mathematically supported then yours.

That said, how do they differ from an opinion?

They don't.

EDIT:

Re-reading your response, it is clear you did not read mine. If you don't want to read it you don't have to, but do not respond to my conclusion without reading the paragraphs that explain it clearly.
 
Last edited:
I think some stats suspicions arise from a lot of individual stats being used in a misleading way by oxygen-deprived TV pundits. I think those broadcast booths must be way, way up there.

For example, we always hear things from pundits like "when team X runs for X number yards they haven't lost in 12 years". Then we can look at the stats and say, yep, that seems legit.

The problem is that if team X gets a significant lead in the 3rd quarter, then play calling is naturally going to become more conservative and favor the run. Why? You need to protect the football and wind down the clock at this juncture. It also doesn't hurt to wear down the opposing defense a little by letting the big uglies lean on some people.

See the problem. Team X did not win all those games because they rushed for X number of yards. Instead, they rushed for X number of yards because they were winning and needed to wind down the clock. In this case the way the stats are commonly presented is perfectly useless with regards to understanding why team X had a significant lead to begin with.

Not to mention that the broadcasters cherry-pick the time period to make the stats more convincing. In actual math and science, such things are obviously a no-no.

You would never know this from watching TV, but there really isn't a correlation between how well you run the ball and how much you win. You want proof? OK. Army in 2012 had two 1,200 yard rushers and another over 800, averaged 5.5 yards per carry, scored 32 TDs on the ground and churned up nearly 370 yards per game on the ground. What coach would not want those stats? They were the #1 rushing team in all of college football. Must've kicked some serious butt, right? We'll if going 2-10 is kicking butt, they did.

Conversely, Louisville, who averaged 3.4 yards a carry went about the country efficiently cleaning people's clocks. Now, were a lot of their games more exciting than they needed to be because they could not run out the clock? Sure, but they won anyway and that's the goal at the end of the day. Just win.

To win all you do is score more points than the other guys. You don't even need to gain more yards than your opponent if you have an opportunistic defense and an offense that takes advantage of trips to the red zone. Of course turnover margin and red-zone efficiency are much more predictive to winning future games than rush attempts or even yards per carry!
 
Ok, let me address your point, then I will elaborate mine. Marketing, in it's most general form, is a scam to put it simply.

In the past decade, "Marketing" has emerged as a college major; where before a career in marketing would start with a business degree. These Marketing majors have no skill in business; that is important. The only math skills they have are very basic. We are talking basic Algebra skills.
Nowhere did I say anything about marketing as a major or any of that. I said that statistically analyzing a market to present a desired product to the right demographic at the right time in the right forum with the right media.... is effective. By no stretch is it a scam. On the margin, there is an "art" to creating some emotional hook and things like that... but that is not a scam either.

So they will end up doing one of two things:

Marketing in this sense is focused on convincing the business that the consumer is a sheep that is easily manipulated.
No it isn't regardless of who does the research or comes up with the next idea to get the consumer's attention. People are not sheep. They do however exhibit predictable behaviors to the point that if you have enough statistical info you can make an effective appeal to them.

Marketing majors generally have no skill in business. It is a catch-22, and it is the reason marketing is a scam.
I could be wrong but I believe Marketing is still a concentration in the business school at Mizzou. My son attends there and has given some thought to majoring in business marketing.

But that is an opinion, and if you choose to believe that the exact size of the packaging, to the intense study of what cartoon mascot should be on the brand is worthy of the billions of dollars paid to mostly incompetent Marketing majors then I guess we will have to agree to disagree.
When it comes to business I believe 2 things: always be moral/ethical... and money is always green.

If those billions produce an ROI and are neither immoral or unethical then the math works. If they don't... then the math doesn't work.
 
I have read your response(s). All of them. Perhaps the things that stick out to me in your responses aren't necessarily the points that you are trying to make. I will concede that. Remember, with any communication there are two parties: the sender and the receiver. If you wish to be more clear, or less distracting, then leave off lessons about coin tosses and your distaste for marketing degrees and/or bad experiences with numbers.

The beauty of this is that you have a complete explanation of the system, and a feel for the results. Use that data to form a counter argument, test it, and come back with data that proves otherwise.

Bottom line: You can write as many posts as you want about how you think that this SHOULDN'T work but at no point have you proven, or even effectively suggested, that this DOESN'T work. Your skepticism is duly noted and appreciated.

Like you, there is a part of me that wishes that this wasn't real, that football couldn't easily be predicted. But, I can't un-ring a bell. As I mentioned, please review the system, crunch the numbers, test the data, and tell me how it fails.
 
Now, the field of statistics is very broad, and covers many bases. But when it comes down to it, the word that more accurately describes what you are doing with these predictions is probability.

Probability= favorable/possible (Atleast I hope it still does, I have been out of the classroom for a while:))

This is easy to demonstrate. favorable/ possible. So if I am flipping a coin and want heads, the probability is 1/2.
And this is where I think you have a disconnect.

You are demanding for the OP's model to be a good one that for coin flips that the 50/50 ratio be maintained at all times. The truth is that it is very, very unlikely that will happen. What will happen if the model is correct is that over thousands of flips you will get closer to that 50/50 ratio and at infinity you would achieve it. At no other time would you have that guarantee.

And it that particular model, you know to a great degree the variables. There could be some tiny, imperceptible variation in the weight balance of the coin... but that's small. It is pretty nigh impossible with football to discover anything more than the major factors that can get you to 85-95% confidence. That STILL does not invalidate the model for making good predictions.

Now let's switch to a football game. We once again have two options (omitting ties), which are still mutually exclusive. In elementary school if you were asked to find the probability that Team A wins, it would still be 1/2. favorable/possible.

But we know this to be untrue. Here is where it falls apart:
Nope. Here is where you have to include all major factors and continue to refine with other factors.

There are an infinite amount of factors at play here.
Possible. But there are a very finite number that are likely. There may be 100. There may be 1000. But those factors are knowable, measurable, and will determine the outcome more than 85% or so of the time. I am not sure what the posted model's confidence level is... but that is also measurable or either subject to a good estimate.

...In football, we have an infinite amount of factors which leads to an immeasurable probability (favorable/infinity). Not only that, most of these factors are not mutually exclusive with each other. Which is just icing on the cake considering we couldn't measure them anyway, but now the probability that some factor A happens, has no effect on factor B.

So what happens is, you are left to sift through these factors and determine which are important and which are not. And we are talking about anything here.

So I could give you a statistical analysis that says APSU will beat UT by 40 points using very real stats and variables. All I have to do is use the stats or factors that favor what I want and ignore the ones that don't, producing incredibly skewed statistics that are mathematically supported.

Well, that's it. My longest post on this forum. I'm sure I rambled in there, but those last few paragraphs were my main point.

That is what I mean when I say that simple predictions from ESPN analysts are no less accurate than a statistical analysis.

I was going to finish this all up but have some things to take care of. The best thing for you to do is research a guy named W Edwards Deming. He used statistics in manufacturing where the variables can be and usually are greater than those in football. You are ALWAYS going to have points outside of 3 sigma or 6 sigma if you have enough samples. But that does NOT change the fact that the predictions of the model are accurate enough to use.
 
  • Like
Reactions: 1 person
I have read your response(s). All of them. Perhaps the things that stick out to me in your responses aren't necessarily the points that you are trying to make. I will concede that. Remember, with any communication there are two parties: the sender and the receiver. If you wish to be more clear, or less distracting, then leave off lessons about coin tosses and your distaste for marketing degrees and/or bad experiences with numbers.

The beauty of this is that you have a complete explanation of the system, and a feel for the results. Use that data to form a counter argument, test it, and come back with data that proves otherwise.

Bottom line: You can write as many posts as you want about how you think that this SHOULDN'T work but at no point have you proven, or even effectively suggested, that this DOESN'T work. Your skepticism is duly noted and appreciated.

Like you, there is a part of me that wishes that this wasn't real, that football couldn't easily be predicted. But, I can't un-ring a bell. As I mentioned, please review the system, crunch the numbers, test the data, and tell me how it fails.

Well this post has proven to me you don't have the slightest idea of the point I am trying to make.

Not even a minute comprehension of the point I tried to get across.

Oh well. Really all I can say is to go take a pre cal class and get a basic understanding of how probability works.

There are situations with finite possibilities like a coin toss, and situations with infinite possibilities like a football game.

The very definition of probability requires a finite number of possibilities.

I don't have any idea how else to say that.

Go try to get a better grasp on what probability is, then re-read my responses and feel free to discuss.

I am not saying your method "isn't real"; I'm not even telling you it is wrong. I'm simply telling you how your method is inherently flawed.

Anyway, read this, soak it in, and see if you can figure out what I am trying to tell you. Or continue to ignorantly ignore basic(very basic) mathematical principles.

http://www.bmlc.ca/Math12/Principles of Math 12 - Probability Lesson 1.pdf
 
There are situations with finite possibilities like a coin toss, and situations with infinite possibilities like a football game.

Coin Toss Possible Outcomes = Heads or Tails

Football Game Possible Outcomes = Win or Loss

Where's the infinity part?
 
And this is where I think you have a disconnect.

You are demanding for the OP's model to be a good one that for coin flips that the 50/50 ratio be maintained at all times. The truth is that it is very, very unlikely that will happen. What will happen if the model is correct is that over thousands of flips you will get closer to that 50/50 ratio and at infinity you would achieve it. At no other time would you have that guarantee.

And it that particular model, you know to a great degree the variables. There could be some tiny, imperceptible variation in the weight balance of the coin... but that's small. It is pretty nigh impossible with football to discover anything more than the major factors that can get you to 85-95% confidence. That STILL does not invalidate the model for making good predictions.

Nope. Here is where you have to include all major factors and continue to refine with other factors.

Possible. But there are a very finite number that are likely. There may be 100. There may be 1000. But those factors are knowable, measurable, and will determine the outcome more than 85% or so of the time. I am not sure what the posted model's confidence level is... but that is also measurable or either subject to a good estimate.



I was going to finish this all up but have some things to take care of. The best thing for you to do is research a guy named W Edwards Deming. He used statistics in manufacturing where the variables can be and usually are greater than those in football. You are ALWAYS going to have points outside of 3 sigma or 6 sigma if you have enough samples. But that does NOT change the fact that the predictions of the model are accurate enough to use.

Ill start from the bottom up. I have never heard of that man but I will certainly look him up.

I think you mid understood my point a little.

I was not saying a statistical analysis is guess work.

I was simply trying to make two points to the man for whom "the bell has wrung".

1) A statistical anylysis can easily be scewed.

2) Just because you present numbers behind your prediction does not make them inheriently more accurate then the person who lists reasons.

That is really all. As for your points in my post they are all correct. I hope this cleared it up a bit.
 
Ill start from the bottom up. I have never heard of that man but I will certainly look him up.

Let me get this right. You are lecturing people on the validity of statistical models and you don't know who Deming is?

You just lost any shred of credibility with anyone who has real world experience with statistics. Anything else you have to say at this point is moot because you obviously have no clue about the subject being discussed.
 
Coin Toss Possible Outcomes = Heads or Tails

Football Game Possible Outcomes = Win or Loss

Where's the infinity part?

The infinity part comes from all the innumerable inputs you could change that might have an effect on the outcome. With a coin, as long as it isn't a trick coin of some type, there are few things that can be changed to change the 50% chance of heads or tails.

In football there are many, many things you can change prior to the first snap that can change the outcome of the game. The truth is, though, that there are probably only a small handful of key factors that can be used to build a model that would be pretty accurate.

Industrial manufacturing processes have an almost infinite number of things that can effect the finished product. In spite of that, most large manufacturers can predict yield and quality by only tracking a handful of parameters.
 
The infinity part comes from all the innumerable inputs you could change that might have an effect on the outcome. With a coin, as long as it isn't a trick coin of some type, there are few things that can be changed to change the 50% chance of heads or tails.

In football there are many, many things you can change prior to the first snap that can change the outcome of the game. The truth is, though, that there are probably only a small handful of key factors that can be used to build a model that would be pretty accurate.

Industrial manufacturing processes have an almost infinite number of things that can effect the finished product. In spite of that, most large manufacturers can predict yield and quality by only tracking a handful of parameters.

There are more factors leading to a binary outcome in football than there are factors leading to a binary outcome in a coin toss. That I will grant.

There are not infinite factors unless infinite means something different that it has meant for the last few thousand years.

Also consider this, what are the three largest factors that affect the outcome of a coin toss?

Just a guess here, but I'd say thumb strength of the ref, wind velocity, and hardness of the playing surface.

Do any of these things help you predict whether a coin will land on heads or tails? No. That is precisely why they toss a coin at the start of a game. It is a fair way to assign possession because the team that calls heads or tails cannot predict the outcome.

Now what are the three largest factors that affect the outcome of a football game?

I'd say relative talent of the teams, relative quality of the coaching staffs, and any possible home-field advantage for the host team.

Do these things help you predict which team will win and which team will lose?

Yes. These are powerful factors if you want to make a prediction.

Therefore, the number of available factors is not as important as whether some of those factors correlate well with the final outcome.
 
Last edited:
  • Like
Reactions: 2 people
Let me get this right. You are lecturing people on the validity of statistical models and you don't know who Deming is?

You just lost any shred of credibility with anyone who has real world experience with statistics. Anything else you have to say at this point is moot because you obviously have no clue about the subject being discussed.

I never claimed to be an expert in statistics; I would have thought that would be generally clear in my posts. I simply have a solid understanding of probability and understand the difference between a coin toss and a football game.

I'm afraid I don't follow your logic aside from that. I am no expert in football either but I possess enough knowledge to have an intelligent discussion.

Also, the history of a branch of science and that branch itself are two different things. Before attempting to belittle me, I suggest you read what I have posted in this thread and find actual information I have posted that is false.
 
Ill start from the bottom up. I have never heard of that man but I will certainly look him up.

I think you mid understood my point a little.

I was not saying a statistical analysis is guess work.

I was simply trying to make two points to the man for whom "the bell has wrung".

1) A statistical anylysis can easily be scewed.

2) Just because you present numbers behind your prediction does not make them inheriently more accurate then the person who lists reasons.

That is really all. As for your points in my post they are all correct. I hope this cleared it up a bit.

Skewed. For all the statisticians. :hi:
 
Well this post has proven to me you don't have the slightest idea of the point I am trying to make.

Not even a minute comprehension of the point I tried to get across.

Oh well. Really all I can say is to go take a pre cal class and get a basic understanding of how probability works.

There are situations with finite possibilities like a coin toss, and situations with infinite possibilities like a football game.

The very definition of probability requires a finite number of possibilities.

I don't have any idea how else to say that.

Go try to get a better grasp on what probability is, then re-read my responses and feel free to discuss.

I am not saying your method "isn't real"; I'm not even telling you it is wrong. I'm simply telling you how your method is inherently flawed.

Anyway, read this, soak it in, and see if you can figure out what I am trying to tell you. Or continue to ignorantly ignore basic(very basic) mathematical principles.

http://www.bmlc.ca/Math12/Principles of Math 12 - Probability Lesson 1.pdf


If your "point" was to simply say that this system is flawed, you are right. That is inherent with any predictability of less than 100%. Did you ever see me make any contrary point or assertion? In fact, I have consistently used proportions to show how often this works (2/3 is a proportion, less than 1). Anything less than 1, in this case, is illustrating a "flaw."

You could have saved much time and effort by simply stating your key point without regaling us with your brother's credentials, showering us with your apparent lack of understanding of the marketing or advertising curriculum at reasonable institutions, or making wild assertions about your perception of statistics and probability.

As far as my education and background, does it make you feel better to assert some perceived superiority over a stranger online, whose educational and professional background is unknown to you? I guess I shouldn't be surprised, you have shown a predilection for drawing sweeping conclusions without the aid of much data.
 
  • Like
Reactions: 3 people
There are more factors leading a binary outcome in football than there are factors leading to a binary outcome in a coin toss. That I will grant.

There are not infinite factors unless infinite means something different that it has meant for the last few thousand years.

Also consider this, what are the three largest factors that affect the outcome of a coin toss?

Just a guess here, but I'd say thumb strength of the ref, wind velocity, and hardness of the playing surface.

Do any of these things help you predict whether a coin will land on heads or tails? No. That is precisely why they toss a coin at the start of a game. It is a fair way to assign possession because the team that calls heads or tails cannot predict the outcome.

Now what are the three largest factors that affect the outcome of a football game?

I'd say relative talent of the teams, relative quality of the coaching staffs, and any possible home-field advantage for the host team.

Do these things help you predict which team will win and which team will lose?

Yes. These are powerful factors if you want to make a prediction.

Therefore, the number of available factors is not as important as whether some of those factors correlate well with the final outcome.

I agree with all your points here. I didn't make the best choice of words but was trying to say that even though there may be hundreds of factors that have some impact on which team wins, there are only a few important factors. Once you figure out those factors you should be able to build a model that fairly accurately predicts the outcome of games beforehand. The OP has built such a model and his post here points out another, further improved model.

These models may or may not be accurate depending on how well they identify and weight the important factors that influence the outcome of the game. The other poster arguing that they are invalid because they don't take every possible factor into account has a poor understanding of how these types of models work.
 
  • Like
Reactions: 1 person
I agree with all your points here. I didn't make the best choice of words but was trying to say that even though there may be hundreds of factors that have some impact on which team wins, there are only a few important factors. Once you figure out those factors you should be able to build a model that fairly accurately predicts the outcome of games beforehand. The OP has built such a model and his post here points out another, further improved model.

These models may or may not be accurate depending on how well they identify and weight the important factors that influence the outcome of the game. The other poster arguing that they are invalid because they don't take every possible factor into account has a poor understanding of how these types of models work.

This is very true and a great point. However, it is very easy to test and refine the accuracy of such models by using similar data which would have been available prior to past seasons. Well, assuming you have time to compile the data.

If your model predicts the outcomes of say all games for the past 5 years at a 90% clip, then it should be very close to that for the upcoming season. The stronger you can get an unbiased model to test fake predictions past games, the better it will do in actually predicting future games.

If your model stinks, such testing will expose that as well. In fact, once the pain of gathering the data is done, running such simulations and trying to get the best repeatable outcome is a whole lot of fun.
 
Last edited:
Just because an event only has two possible outcomes doesn't mean that the likelihood of each outcome is 1/1.

This isn't "theoretical physics" this is football. Iv'e been watching football for almost 50 years now and have seen countless games where the team that was supposed to win lost and the team that was supposed to lose won. Football is played on the field not in calculations. There probably is something to these theories which is all they are but there are too many variables in the human equation to place stock or even bet on it. Here is my take. Watch a few games being sure to have some Jack Daniels!! Enjoy the games and save physics for the classroom. I kinda like the chaos theories myself. :hi:
 

VN Store



Back
Top