So....forgetting the draft, are we a better team than that one exiting the playoffs?

Don't forget, this is also a team that almost beat Seattle week 1, and would have without the Kenny Stills drop.

I see our struggles with the "easy" portion of the schedule mainly due to injuries, and lack of depth. Assuming we were healthy, and playing well within our new system (like we were down the stretch when we ripped off that winning streak) those margins of victory likely would have been bigger.

I think both sides of this argument have valid points and are "right", but nobody wants to admit it and "lose" the argument at the end of the day.

 

The thing is, Brandon, this thread is about whether or not we're better now than at the end of last year. I'd argue we are. Everyone agrees that we over performed last season...with the caveat that we did over perform because we were riddled with injuries, and had a first time first year head coach. Phins is arguing that somehow the team on the field with backups on both sides is somehow representative of the team when healthy.

 

I havent said anything in regards to those games. Losing your star QB is an injury excuse I will actually listen to as opposed to other positions simply because of the impact a QB has on the game and the lack of actual starting caliber guys in the league, even though we do have one of the better backups in the league.

Also I have literally no clue what you are talking about in regards to how the Pats beating us in week 17 when Tannehill was injured at all counts as a caveat to the pats super bowl win. I have never said we have to blow every team out that we are better than. But to not beat any of them and to play that close to all of them is a red flag for how good this team actually is.

And enough of the it came down to the last pass against new england. they embarrassed the **** out of us. They were up 31-3 with 5 minutes left in the third quarter with their 2nd string QB and 3rd string QB playing the entire game.

 

The easy part of the schedule was just about the only time all year we were healthy, what are you talking about? The O-line was only healthy in weeks 5-10 when pouncey played and we still played close at home to the Bills and Jets, and the Chargers on the road and played absolutely terribly against the Rams, although i dont think Pouncey played in that game.

But again, you guys continue to act like every team doesnt have injuries every year. Albert, Pouncey, Misi, Maxwell are guys that get hurt literally every year. And there will always be tough holes to fill through injury over the course of the year. Has our depth gotten any better? We re-signed the worst graded guard in the league last year, have the same backups along the OL, got worse at the other S position across from Jones and havent signed anyone else. The only position it can really be argued that we have improved our depth at is LB. We added Hayes and that improves our starters at DE and I think he is actually our best addition of the offseason thus far, but Branch is still one injury away from starting again and he was absolutely terrible last year and I have no idea why we gave him a big contract.

So we can continue to blame it on injuries, but every team has injuries, and we really havent improved our depth again, granted we have plenty of time left in the offseason.

 

That's not what im arguing at all. Obviously every team is better when they are fully healthy. The only problem is no team is ever fully healthy. If the Dolphins are fully healthy and dont suffer any significant injuries that expose their continual lack of depth on both sides of the ball, sure, we'll be looking really promising. No team goes through a season like that and if they do it is more of an aberration than any skill.

 

Phins18 can do his own thing, I'm not making his points or arguments.

My point is that you keep bringing league avgs, and they aren't relevant when speaking on a particular team, how often good teams, how often bad teams and so forth, doesn't have much bearing on a particular team.

Reality of the Dolphins is that they won a lot of close games vs subpar teams and lost big to good teams, that is not an indication of skill, it's an indication that you need to improve.

 

League averages are very useful in statistical analysis, especially when talking about individual teams because you can compare to what's expected. They're even more useful once you have the standard deviations, or better the actual distribution around the mean.

In this specific case, let's just for simplicity sake say that you generally have close games when you are more evenly matched with your opponent on game day. So it doesn't matter that Seattle ended up being a top team, we were close to their strength when we played them, and the same is true for those "weak" teams you're speaking of.

Others have already pointed out good reasons why this was true for us for many of our close game wins: injuries at key positions, but let's ignore the "why" and just ask how we can use "averages" and "distributions" to calculate the probability any kind of skill was involved.

OK.. how about assuming the Dolphins were an "average" team that outperformed their expected level of success, either because of weak opposition, luck or whatnot. That is, assume 8-8 is expected for the Dolphins. When you're in the middle of the distribution (NOTE: this is where "average" becomes useful in the analysis), then the probability a team you face on game day is weaker or stronger is 50%, which simplifies the analysis (I'd have to bring in extra data if we did this assuming we were really a 9-7 team).

Fine.. if the probability we'd face a stronger or weaker team is 50% by assumption, then we can just calculate the probability of any such "average" team winning at least 8 out of 10 close games, where "close" means the game ended with 0-7 point differential. That calculation is easy because you can use something called the binomial theorem and get the result immediately without simulation.

The result is 5.5% of the time that you'd expect an average team to win at least 8 out of 10 "close" games (so either 8, 9 or 10 wins). That's precisely on the cusp of where a statistician would say your initial assumption that the team was actually "average" is unjustifiable.

So yes "averages" can be useful in determining how unexpected a result was (should be obvious really), and in this case it's untenable from a statistical point of view to say the Dolphins were really "average" but just got lucky with their 8 wins. The level of luck we're talking about is 1 out of 20 times luck. You want to believe in that fine, but I'd rather accept that "average" doesn't fit the data, statistically speaking.

Now if you assume the Dolphins were really a 9-7 team that outperformed you wouldn't get that kind of rejection of the hypothesis, so I can't say that's too unlikely to be rejected but the burden of proof falls on you not on me to say 9-7 was the most likely given our opponents (actually might be true given that we did have more weaker opponents than stronger ones for close games, but who knows until you do the calculations?... calculations that use "averages"! )

 

I'll use the same analogy I used in my earlier post, league win avg is around 8, so if we use that as a basis then we can assume that all teams will be .500?

You see how faulty it can be using league avgs on particular teams, it's like finding out the avg depth of an ocean, it's pretty useless if you're fishing 5 miles off the coast of Cape May, because you need to know the depth where you are.

 

Faulty logic assuming there is any source of noise/randomness and one team has to win and there are no ties.

In the case that all teams are absolutely identical (down to the coach, players and scheme) but we assume there is some source of variability in every action, the binomial distribution once again tells you the probabilities.

You'd expect 12.2% of all teams in the league to end up with 10 wins. The most likely outcome of course is 8 wins, but that only occurs 19.6% of the time.

One technical question people could ask is how do we know what the variability looks like? There's a powerful result in statistics called the Central Limit Theorem (CLT) that tells you how the distributions of SUMS of random variables look like, no matter where the individual sources of variability come from. So as long as you can identify multiples sources of variability on independent trials (this is easy in football because you have separate plays), the CLT tells you the distribution of the noise will be Gaussian and the binomial theorem calculates the probabilities for discrete numbers of trials.

 

You are helping me prove my point here.

 

Not sure what point you're referring to. I do know I'm helping to disprove the two points you made in post #326.

Specifically, 1) I'm showing that averages can be useful when talking about specific teams, and 2) that winning 8 out of 10 close games is much more likely to be an indication of skill than luck.

Is there a separate point you were making I'm not thinking of?

 

But you keep saying that the only good teams we faced blew us out. We were one play from sweeping Seattle and NE to start the year 2-0. Then our line collapsed and we were the worst offense in football for three weeks...I don't think anyone's even counting the Browns, Bengals and Titans. All three were a joke with our QB getting slaughtered.

The only other game we lost THE ENTIRE SEASON with our starting QB was the Ravens...and that one was ugly. That's the only "powerhouse" team blew us out when Tannehill was on the field- just that once in 2017.

Stills dropped the winning TD pass in week one. Parker dropped the game-tying pass in week 2. We had no line week's 3-5. If we're talking QB's, none of that was Tannehill's fault. Then he won six straight, lost one and was leading in his last game before getting injured. If it wasn't for the line and player mistakes, Tannehill would have started the year 11-1 before the injury.

So yeah, that was obviously a crushing loss on our roster (even though I still love Moore). RT17 played phenomenal.

By the way, that "blowout loss" on Moore in week 17- it was a seven point game midway through the fourth and we had the ball in the Pat's redzone. A receiver basically handed the ball off to give the Pats a long score, which put us in last-minute desperation mode. You can easily say that game hinged on one play as well, even though we lost by 21. One stupid mistake was a 14 point swing and killed all momentum. Since you keep bringing up context, there's some for you to think about. It was a tossup game for 56 minutes.

At to the caveat comment, I was quoting you- it was a caveat that we blew out Pittsburg with Big Ben hurt. So don't we get the same thing against the Pats and the Steelers without Tannehill? Again, you can't have it both ways...either you don't fully count the win since one team was hindered without their QB or you don't.

Either way, going 10-6 and missing 12-4 by two dropped passes is sensational for a first-year coach, a brand new offense and a defense that may almost entirely second-string. If we got the line sorted out in per-season instead of week 5, then maybe we end up 14-2. That's darn near legendary for a first year coach and we're starting next year with all our important first stringers. That's all anyone here is saying...why continue to argue this?

 

Dont have time to respond to this right now, but holy homer post. WE were down 31-3 in the 3rd quarter to the Pats 2nd and 3rd string QB, coming back in that game is nice resilience, but we were down 31-3 to their 2nd and 3rd string QB. You say we are so close to going 14-2, yet completely ignore my entire point, every game we played against those bad teams we could have and in a lot of those cases should have lost. So your 14-2 in your rationale could have easily been 2-14.

Again, I havent brought up those 2 games against the Pats and the playoffs game against the steelers. my entire point was that we played so many bad teams and played them all close. Good teams dont play that many games against terrible teams and not win by more than 7 against anyone.

 

Yeah, you're ridiculous if you truly believe that most teams are blowing out good teams with a first year head coach, new offense, and all the injuries.

You have extremely unrealistic expectations. Come on, just tell us your previous username(s).

 

Enough with the ****ing injury excuse. Jesus christ we are not the only team that had significant injuries. We had an absurdly low strength of victory, and the quality of our team showed in the closeness of the games against good teams as compared to bad and our -17 point differential. You can continue to use this first year head coach and new offense hope trafficking, but my point is that we significantly outperformed our talent level, and our pythagorean record cements that.

You can say we had a new offense but we also had a QB in his 5th straight year with the team, the same LT, RT, and C from the last 3 years, the same WR's from last year, and a RB who has been here for 2 years. New offensive system, but we had huge continuity on that side of the ball.

And its not unrealistic expectations. I have no expectations for this team because they have been exceedingly average for about 15-20 years at this point. You are just a homer who cant acknowledge that the team took advantage of a cupcake filled schedule and showed how good they truly were by playing that many terrible teams that closely and getting destroyed in just about every other challenge they actually had to face. There is no problem taking advantage of a schedule, but if the Browns played nothing but pee wee team and went 16-0 and struggled and never won any game handily against that many teams that they are supposedly better than, would you brag about how good they are or look at how close those games are against bad teams?

 

Lets just look at a few of the key players...

Mike Pouncey. Out weeks 1-4, Questionable week 5, off the injury report weeks 6-9, IR the rest of the year. Weeks 1-4 we were 1 and 3. From what I can see we won each game he played in. 5-2 record after he went on IR, and of course the playoff loss.

Reshad Jones missed basically our entire winning streak. His last game was our win against Pittsburgh. So, 8-2 without him.

Xavien Howard... missed weeks 2-5. 1-3 record during that stretch. Howard then was out for the win against the Jets. Doubtfull for the SF win and Baltimore loss and I dont think he played in either. He was also limited in the Arizona win. All of which were games during our "winning streak"

Add Jelani Jenkins to that list of missing lots of time during the win streak. Sure, he's not the best LB in the league, but he was a starter for us. Meaning that, when he's out, someone likely worse than him is playing in his place. That... makes it more difficult to win games, even games you "should" win.

I think you're being a little selective with your memory here about guys who missed time during the win streak. That's just four players I looked at quickly, and I can't be bothered to look more in depth because I think this whole thing is silly. You can go look yourselves here for the full list... http://www.pro-football-reference.com/teams/mia/2016_injuries.htm Sure, the list is a bit limited in the sense that you dont know when they left games, and if they were questionable you don't know if they played or not without lookin at the actual game book, and I just can't be bothered to do that.

Yes... all teams have injuries. Nobody can dispute that. However, it's important I think to remember that while all teams have injuries, I think it's fair to say we sustained more to starters, and we all knew we weren't a particularly deep team. So the fact that we had put together the successful streak we did, regardless of when players missed time is pretty impressive, especially given the fact we weren't a deep team, and guys missed significant amounts of time during that win streak. That means you're playing guys in quality minutes who shouldnt be playing key minutes. Like, yes, we should have beaten the teams we did, and we probably should have beaten them better than we did, but ultimately you should easily be able to see why.

Again, I still stand by what I said earlier.

To stay on the topic here though, are we a better team? Yes. Simply because we're getting our guys back, healthy. That's huge. That's 3 players above that I just listed. Three key players that are starters for us. That's a huge improvement alone. Add in Timmons, Hayes, and an upgrade at TE, so I'd say yes, we are better. That doesn't mean we're ready to compete with a team like NE though, not yet.

I just find this argument a bit... odd in general. I get what you're saying with point differential etc, etc. I think you just need to look beyond that. This is focused on the injury front, and I get it, everyone deals with them, but there was far more at work here. New coach, new systems, new players, still growing QB, etc, etc. The list goes on. I'd say, if we see the same thing this year, then there's an argument for sure that we're not good enough yet. However, I dont think anyone thinks we are either.

 

Mike Pouncey misses games every year and there is legitimate concern that his hip will impact him the rest of his career. Jones was missed, but again, every team has injuries.

Did you seriously just name Jelani Jenkins and Howard? No **** if he's out someone worse than him is playing because he is a starter. That doesnt make him good. That just further strengthens my point that part of the reason this team wasnt better was because we consistently have no depth.

We arent the only team that suffered severe injuries, and you have no proof that our injuries were more to starters than any other team. Our bad depth being exposed doesnt make our season more impressive, it furthers my point that our team as a whole did not have the talent of a 10 win team and took advantage of an easy schedule.

When i was referring to that health over that time i was referring to the health along the OL. That period after the Titans game through the next 5 or 6 games was when they were healthy.

 

OK, so you're bringing up Pouncey here again... in this sense to say we're not better this year with him, than last year without him, simply because he may get hurt again? I get that... sure, but he's going to be here and healthy to start the season, so that makes us better. If he goes down... then yes, different story potentially.

Why wouldn't we mention Jenkins and Howard? Both were essentially starters, and missed time. Howard should be fine this year, and Jenkins has been replaced with a better player. So, that's an improvement. I'm not saying Jenkins being a starter makes him good. I'm saying that when he's out, and his replacement is in we're not AS good as we'd be with him. So, the fact we won games during our win streak without him, and playing someone like Vigil/Paysinger/Hull, is an accomplishment. Especially when combined with the other guys missing.

I'm saying, and I think others here are trying to say that what this team accomplished given their situation was nothing to be upset about. I also think that it goes hand in hand with the statistical things that you're saying were bad with respect to our team. I agree with that, it's not sustainable, or practical to think we could keep winning like that consistently. However, I think the point is very fair that those numbers for us could have been better with the guys we were missing. That's all.

I hear you though, we should have beaten those teams by more, I also think there's reasons why it didn't go that way. Justifiable reasons. I'm also of the belief that depth is lacking, and we have a long ways to go still. I don't think anyone should dispute that either.

I just think that ultimately, you have to look at and understand the reasons WHY we played those "bad" teams so close. I think that's all anyone is trying to say here. You can't dismiss those reasons.

 

OMG....so now we can't count the second half of games either? Let me get this straight-

- Injured QB1's matter
- Other injuries don't matter
- Wins don't matter
- 2nd halfs don't matter
- Comebacks don't matter
- Beating bad teams don't matter
- Beating average teams don't matter
- Beating good teams only matter if their QB1 plays every snap
- Making the playoffs doesn't matter
- Being a 6th seed makes you a loser
- Stats don't matter...especially when they prove you're wrong


That leaves me with a handful of questions- why do you even pretend to like football? Why bother even posting here? Is there anyone on this site you haven't called ridiculous or other names? Because if you can't have any respect for anyone and the team is really as God-awful as you make it out to be (from your perception), I don't see why you'd even bother logging in or turning on a football game. It really seems like you hate the Miami Dolphins and anyone who would dare say a positive thing about them.

So I'm getting smart like the others and placing you on ignore.

 

You seem to be talking about this stuff as if it were hypotheticals, it's not, it's empirical data, last season happened, we don't need to talk about it as if it's a hypothetical.

"8 out of 10 close games is much more likely" Speaking on it as if it were theory, but we have the facts, theories are for things we don't know, and you speak on these 8-10 games as if they were in a bubble.

You are omitting the fact that good teams beat us handily, if you can't beat good teams that is a sign of lack of skill.

 

Do you know of a way to measure how much of an outcome in a football game was due to luck vs. skill?

If you can't measure it directly, then there is no other option but to bring theory into the discussion.

Now to be clear, you could estimate the amount of luck vs. skill in a single game if you had the ability to take the same teams in the same situation (or almost the same situation) and have them play each other tons of times. Then you could see how much of a particular outcome was due to just natural variance, but we can't do that.

So while no one is disputing our record, nor the records of the teams we beat or lost to in close games, that alone does not tell you how likely it is for 8 out of 10 close games to be won. You can't just assume you know which of those games should have been won or lost.

 

There's no estimating luck, lol, it happens when it happens, good or bad, like when most of the teams we played last year underperformed based on the previous year, that's luck, bad luck for those teams, good luck for AFC east teams.

However, luck isn't the point, I'm just looking at the facts, we played close vs bad teams, we lost vs good teams, which is pretty much a text book example of needing to get better, we were pretty much a paper tiger heading into the playoffs last year, carrying a record that was not indicative of our talent level.

Very reminiscent of 2008, enter the playoffs with a good record and get pushed around by a real playoff team.

 

"There's no estimating luck" .. Dude, estimating the amount of "luck" or events due to chance is what maybe half of all statistical analysis does!! That you don't understand how this is done doesn't mean it can't be done.

As for the rest.. no one here (that I know of) is arguing we don't need to improve. You can even use the close games vs. teams with bad records as an argument. That's fine. But don't act like our close games record wasn't a sign of skill. It's precisely the type of skill good teams have and shows this team is improving (IMO mostly due to Gase).

 

Exactly. Those close games are the types of games this franchise has been losing for years!

 

That's what I was going to say as well- don't make it out like it doesn't matter if you win by 3 or lose by 3...there's a massive difference there. To me, it's what defines the team's character, their desire to overcome mistakes, etc. You can't measure that stuff by stats either, yet we all know those things improved in 2016.

Besides, a season is defined by two stats- wins and losses (and ties if you want to get technical). NOTHING ELSE matters in the grand scheme of things; everything a team does is built around those two stats. Talking about how close a game is doesn't even weigh in as a factor because you're ignoring all the other variables.

Heck, the first game Moore played this season, he didn't throw a single pass in the 3rd quarter. Were we a worse team because he didn't unleash a series of bombs in that game like he did the following week? Of course not. We played conservative until the opposing team made us play differently....because that's all that mattered. And when the time came, there was Moore to drive us for the win.

The only reason it was close to begin with though was because Gase took his foot off the gas and dialed back the offense to burn the clock. That makes him smart- not inferior.

 

It is an interesting question though as to how important W/L vs. score differential is in people's minds when you ask people to rank teams. Not that I have an answer but here's a starting point:

The 1972 Dolphins went 17-0 with a score differential of +13.6. The 1985 Bears went 18-1 with a score differential of +17.8. Most NFL commentators put those two teams as the two best ever. Does that mean that IF you win the SB, a +4 score differential across a season ~ 1 win in people's minds?

Also consider that the 2007 NE Patriots went 18-1 with basically the same score differential as the '85 Bears: +17.5. But they didn't win the SB and no one puts them on a "top X" list (unless you're a Pats fan? haha). So that kind of shows that score differential is relatively meaningless compared to a SB win but it does seem to enter into people's minds otherwise.

It would be interesting to see how people generally value these things, but we need a larger list of team ratings to find out.

 

No, people try to estimate luck, there's a difference, luck is completely unpredictable except in the sense that it is going to happen, but where, when, how and how much is as big a mystery as anything, it is quite literally easier to predict where lightning will strike.

Does luck play a part in every game? Absolutely, can anyone predict how it will manifest itself? Absolutely not.

You are trying to say that reasonably, luck will only play X amount in a season, which is impossible to predict, because luck doesn't follow reason, you can give me avgs, but that's all they are, avgs, and they have no bearing on specific events.

The basic variables in just the simple design of the ball are astronomical, and a single bounce of that ball can determine the winner and loser of a game, and there is no predicting which way it will bounce, that's all pure luck.

You say that it's a sign of a good team to win close games, and in a bubble that is basically true, but since we couldn't beat good teams, that fact cancels the other out, you can't look at one and not the other, which is what you are doing.

 

I'm trying to figure out how to explain this to you but I'm slowly running out of ideas. You know.. not everyone understands this stuff.

First of all, I think you're confusing this idea that one cannot precisely predict the time, location and magnitude of a particular event with the false claim that one cannot predict its probability of occurrence. One way of interpreting the probability of occurrence is to say that over time under the same conditions, your predicted probability will match the observed frequency of the event occurring.

This means you can test claims about the probability of something occurring. It's not just some mathematical mumbo-jumbo that has no relation to the real world. That Central Limit Theorem I talked about can be rigorously tested to see if the observed variances when doing physical experiments match the predicted ones, etc..

So while it's true that you can't perfectly predict the time, location and magnitude of any event (since quantum mechanics science has come to accept the idea that there is an inherent probability in the ability of observers to make such predictions), you definitely CAN predict the probabilities with extraordinary accuracy (quantum mechanics being the best example).

That's all I'm doing. I'm not trying to say in a particular situation X% was due to chance. I'm saying that the best estimate of how much was due to chance is to X%. When a ball bounces, you can't predict precisely what will happen, but you can predict the distribution of what will happen in many cases extremely well.

Anyway.. not sure you'll ever understand this. I've seen many other people never get it. But I will say this. Your logic is wrong.

 

Well, the formula would be different when comparing the best teams ever- stuff like total yards, turnover ratio, margin of victory, etc would come into play. You can definitely learn from stuff like that so don't take what I said earlier the wrong way.

I just feel like it's incomprehensible for any sports fan to not understand the basic principle of competition- someone wins and someone loses. And yes, sometimes the losing team may have played "better" and have been the "better overall team", but that's not what the history books remember. We only remember the winners unless our team really sucks....then we remember the "what could have beens."

To put it simply, we had a string of 7-9 to 8-8 seasons over the past couple of years and the teams we saw had zero identity. Tough in the red zone- that's about the only thing we had going from week to week. Suddenly we have a 10-6 year though after a complete team collapse early on, and our identity is a whole lot clearer.

I think that's what the OP meant by this whole thread- where we've grown and what we will become in 2017. Why we're debating if winning is a signal of a good team or not is a true head-scratcher though- we may as well argue about how long it takes for paint to dry.

 

The NFL is truly a game of inches...

Thats why you saw our team play well enough to win on the road vs the Seahawks but lose, and our team play bad enough to lose vs the Browns and win.

We won a lot of games by those "inches" last year, but I believe thats what good teams with a good culture does. You could make a case NE could've lost EVERY SB game they played in if it wasn't for those inches they seem to always be on the right side of...unless of course they were facing a non scared well coached team like the NY Giants with a QB with ice water in his veins.

Good teams find a way, we are a good team (finally) so I believe we will keep finding ways to win and not ways to lose.

 

Are you kidding with quantum mechanics, lol, I'm talking about football, you're off again on theoretical hypotheses on an unrelated subject.

Predicting that luck will happen is not what we are talking about, we are talking about precisely how much luck was involved last year, which cannot be quantified.

"I'm saying that the best estimate of how much was due to chance is to X%."

You realize that your estimate has nothing to do with the actual facts right? Because any single event can vary from 0-100% luck, you do realize that correct, so any estimation you come up with is a pure guesstimation, but you want to treat it like fact.

So my logic isn't wrong, lol, you are having your own discussion about unrelated theories, probabilities and hypotheses that aren't pertinent to the discussion.

You're sitting here trying to scientifically prove your opinion, lol, which as we all know, can't be done, you can't prove an opinion.

 

I have to laugh that Finster thinks cbrad is trying to "prove his opinion". I thought he was confused, but after that statement he's not confused, he's literally blind and deaf. lol

 

Yeah.. this sentence is proof you don't understand how statisticians estimate the probability of an event occurring by chance, or the likelihood of a hypothesis being true given the data.

The analysis I did is based on the data whether you can ever understand that or not. Anyway.. no use trying to explain this to you anymore.

 

Lol, well let me put this in a real simple way for you.

How much luck, exactly, was there involved with the 2016 Miami Dolphins season?

 

This is actually a good question. I can't give you the answer right now because it would take me a long time to get the relevant data, but I can explain the process you'd need to go through to solve the problem.

1) You need to first start with some hypothesis about what the probability of team X beating team Y is in any given matchup. Let's for shorthand label that p(X>Y). Now, there is no unique way of doing this (which is why you have to start with a hypothesis) but whatever method you choose you have to be able to defend well.

One highly defensible approach using only W/L records would be to look at historical data on the actual p(X>Y). That is, look at all cases where a team that is let's say 5-3 beats (or loses to) a team that is 4-4. Granted, this approach would ignore differences across eras, but you can test whether distributions in different eras are statistically speaking different or not, and if they're not you can use the entire data set. If they're different, then you'll have to restrict your analysis to a particular era.

You also might not have enough samples of a 5-3 team playing a 4-4 team (I think you do though..), but if you don't you can just fit a mathematical function not only to all cases where a 5-3 team plays a 4-4 team, but to every possible combination and then use the function you fit as your prediction for any possible matchup.

2) Once you have p(X>Y) for every X,Y that you need (the matchups in the season), you can calculate the probability of any set or sequence of outcomes (technically there's a difference if you care or don't care about the order in which something occurred). Regardless, the probability of a set of such outcomes is just the product of the individual probabilities.

OK.. say you observed 10-6 as the final record. There are a ton of ways of getting 10 wins from the individual matchups. It's tedious but you'd have to sum the probabilities of every possible set/sequence of outcomes that gives you 10 wins and that tells you the probability of having exactly 10 wins (good exercise for computers!).

In statistics you usually ask how likely is it to get 10 or MORE wins as opposed to precisely 10 wins because getting precisely K wins in N possible games becomes astronomically small the larger N gets so it's not very informative. But getting at least K wins in N games remains meaningful both in theory and practice so you'd calculate the probability of getting at least 10 wins using the method just described.

That's how you calculate the probability of the observed number of wins being due to luck, answering your question.

Oh and note that because the teams you play have different records than the teams another 10-6 team may have played, the method automatically adjusts for "difficulty of opponent". Further note that this only used W/L record in step 1. You can add as many conditions as you want of course.

 

My bet is Finster is curled up on the floor crying after reading this...lol

 

I'd go a different route entirely- I'd get out game film and go through all sixteen games, assigning a +/- value to any "lucky" play that led to an outcome change. For instance, Stills dropping the bomb in week one. We'd be at -1. The two missed field goals from a typically makeable range. Now we're at -3. You'd have to go through every single play though for 16 weeks to see where mistakes happened that created advantages or setbacks.

For instance, if a pass is tipped and our offense gets a turnover, +1. Tannehill trips and falls down on a rollout, -1. Landry drops an easy pass, -1. An opposing RB fumbles while making a cut, +1. Yet if the running back gets lit up and fumbles, that doesn't count because it was skill instead of luck. A starter gets hurt, it's a + or - 1 depending on the team. Someone misses an easy tackle, that counts too. Then add it all up and you have a "luck factor" to score the year on.

 

Yes, but you also have to factor in weather, field conditions, injured players, players getting injured, nicked up players, players that are sick, a CB that is hidden behind a lineman, the sun getting in a players eye, the refs, things that may have happened in players personal lives prior to the game, the coin toss, lol, and a bunch of other stuff.

Also, each individual piece of luck has it's own factor, as in they are not all equal, a tipped pass that ends up as an INT has a higher factor than an avg dropped pass, but each dropped pass can be different, a dropped pass early in a drive that ends up as a TD is a low factor, but a pass that is dropped in the EZ that leads to no points is much higher, and if it's at the end of a game on 4th down it's a monster factor.

IOW, there is no way to ascertain exactly how much luck is involved in a game, or even the skill/luck ratio of any given play, there are factors that we'll never know about, and even in the example Key has where the RB gets blown up it's skill, that's only partly true, because luck plays a role in whether or not the defender is in a position to get to the ball to knock it out, because the defender can't control how the RB runs, and what if another player picks it up and gains yds, or scores, then the FF is all of a sudden, bad luck, but it was good luck just a second ago.

Trying to guess how much luck is involved in a game, because guessing is all you can do, is like trying to guess how many jelly beans are in a jar you can only partially see, it's an exercise in futility, and anyone who says different is trying to sell something.

 

Actually, there are a number of ways to figure out how many jellybeans are in a jar. I'm sure cbrad would help you with that, too. lol

 

I keep reading this. And with all due respect, there is no way to know that. It happened in the second quarter, not in the closing moments. Anything could have happened if they scored a TD there. Maybe Miami goes on to blow Seattle out. Maybe Seattle adjusts and they blow us out. Maybe it's a close game.

But so much football was left after the drop, you cannot say "if we score 7, 10+7=17, so the score would have been 17-12. We win."

That's not how this works. I think this forum needs to re-read Ray Bradbury's The Sound of Thunder one more time.

Remember when Marty let Biff steal his Sports Almanac? You see what happened to Hill Valley?