oldmanfan

Fair appraisal. I don’t give grades myself.

Why give a big contract to a guy the PFF has ranked as like the 135th best CB in the league? (sarcasm off now) Get it done; you say you want to use the draft to get your team put together then you need to keep the All Pros you draft.

And some of that is him, some of it is drops, some of it is offensive philosophies where he throws downfield more in college and the pros. I have no problem saying Allen needs to continue to improve. Why can’t some just say they understand that statistics can be misleading? No. There is no difference period. Given the number of passes he threw there is no difference between 58.8% and 60%.

Got it, see above Please quit obsessing over this. There is no statistical difference between his percentage last year and 60%, yet you treat it like it’s the Holy Grail.

Can you clarify the data set? Is this completion percentage? Never mind, I see that it is now. I think we’d all like to see a higher completion percentage because it would basically just look better. But when I said above that just focusing on a number can be short sighted, this is an example. You have to take all the confounding variables into account that could affect completion percentage. The throwing ability of a QB us one of those of course. But drops are another one. Throwaways another. Play calls - how many throws are more downfield vs. quick checkdowns. Allen to me needs to improve on reading defenses. That would let him throw more often to the hot open guy, make the quicker checkdown throw. That will improve his percentage more than anything. i see the 60% thing has reared its ugly head again. I’ve done this before but the difference between 58.6% and 60% adds up to less than one completion per game

I’m afraid I’m not much of an expert on breaking down O line play. My uneducated mind agrees we tend to not do well with stunts. I’m not sure how to reconcile the blitz vs. hurried data; you’d have to know how many hurries come from non-blitz rushes vs. blitz. As for the median of 317 passes this season you’d want to know the range for that data.

As far as standard deviation you would have to give me the specific variable, but as a general rule the more measurements you take the more normal the data distribution and the smaller the deviation. As far as measuring accuracy, truly the only way to measure that is to throw at a fixed target. I have mentioned the dartboard analogy here before, and it has generated a lot of comment! Accuracy is how close you come to a specific point, and precision is how repeatable to hit the same point. So if you surrounded a bulls eye, but you make a ring an inch away from the bulls eye, you’re accurate but not precise. Conversely, you can hit the same exact point time after time, but if that point is two inches to the right of the bullseye you’re precise but not accurate. So, taking that to your comment on the on target percentage, I would first ask how you define your target. I have seen pff define it as right between the numbers. That to me requires not only high accuracy, but really high precision. The great QBs (and the great dart throwers) are both highly accurate AND highly precise. Allen is pretty accurate; I have graded several of his games and defined accuracy by a throw in the receiver’s catch radius. But he can improve precision, throwing it to a tighter spot where say a receiver makes a catch more in stride. He is both when he has time to set up and use good mechanics like most QBs. And to me his ability to process and make quicker decisions where to go with the ball is the key for that. Protection can help there as well, but overall I want to see him make quicker decisions. His bad throw stuff I think is affected by him tending to throw downfield farther. He has to learn to take the check down (funny how many want to see that now we’re the same folks screaming about Trent Edwards). Take the dart board guy again. If he’s throwing at a dart board 40 feet away vs. 10 feet he’ll probably have more bad throws at 40. I think Allen has it physically. His challenge is the decision making. He was better this past year than his rookie year. He should continue that trend next year; I see no reason to think he won’t.

You have gone off the reservation now. What I said very directly is that trends are not definitive proof. Anyone with a single functioning neuron in their skull would rather see positive vs. negative trends in data. For example, Allen's completion percentage went up several percentage points year one to year two. That's a nice trend, but hardly conclusive. I'd like to see it continue to go up next year and not fall back. His interception rate went down, that's always a nice thing to see. But not definitive. Where you continually miss the point is that you and others take a single data point and try and make it into something meaningful, when it is not. The 60% completion rate is a perfect example. People go on and on about that as if it the Holy Grail of QB play. But there is little data to prove that, and Allen is not statistically different from that number anyway. And here is what you and others who don't seem to like the guy do: you just then arbitrarily move the goalpost. So if he hits 60%? it will become that 62% is the Holy Grail, or 65% when he hits 62%, and so on. Let's take the quote from you on this magical 62% number; I'll copy it again here: "Because the baseline for accuracy in those days was 60%. If you were over 60% you were accurate in the pros. Today that's creeped up to 62-65% coming out of college" So there's your quote, and you are saying now that you did not say you need to have 62% accuracy to be successful? Excuse my bluntness, but then what the hell do you mean with this quote? Do you have to be accurate to be successful? Or not? Or what? As for Allen and that he cannot improve, your rather strong implication based on your erroneous understanding of neuroplasticity is that he won't. Can't was perhaps too strong an interpretation on my behalf. But, again, that is where your background matters. I have a background in anatomy and physiology. If you have a background in medicine or psychology or physiology where you have studied this we could have an very insightful discussion. Do you? I am an admitted skeptic on most football ratings for QB play (or other positions for that matter) because, by and large, they do not take into the multivariate nature of the game. Take a single stat like completion percentage. The number of variables that go into whether a single pass is completed (effect of the pass rush, defense, QB's ability to throw, WRs ability to catch ) are many, then multiply those over X passes per game, and then X passes per season, and I contend the variables are not well accounted for. QBR tries to take variables into account, but their approach is proprietary and even QB's in the league (I have seen quotes from guys like Rivers) do not understand what the QBR is calculated and what it really means. Let's take a QB go 20/25 for 2 TDs and no picks and a win vs. a guy that goes 30/62 for more yards, 1 TD, 2 picks, and a loss. The former probably has a better QBR, but does it mean he's a better QB? The former implies the QB took what was there, had a stronger running game, and a better defense. The latter probably had a worse running game, and a D that gave up points and forced him into throwing more to try and catch up. The latter may be a HOFer, and just hit the wrong team on the wrong day. I prefer the Potter Stewart method of evaluating QBs myself. Potter Stewart was a Supreme Court justice who in a famous trial on pornography said: "I know it when I see it". I've been watching football since the Bills started in 1960. Haven't seen enough to know it yet with Allen. But I've seen enough to know I want to see more.

Accuracy is not completion percentage. That theory has been debunked over and over again. Changing the goalposts? Hardly. I have been as consistent as you can be around ehre that statisitics are used very poorly in these conversations because people by and large do not understand statistics. Statistics is a science. There is no such thing as "football" statistics vs. statistics in general. That is foolish and why people get misled into thinking some numbers matter in football. But you have actually said in this thread that #'s are generally true because they generally are. You said that. And not to be harsh, but that statement really means you should exclude yourself from discussions involving any form of statistics. It shows you really don't get it. Did I celebrate Allen's performance after Dallas? Sure, because I thought he played pretty good and we won the game. Did I celebrate after the playoff loss? No, and there were things Allen did wrong in that game that he needs to improve upon. I think if one looks objectively at Allen his performance in year 2 was better than year one. His numbers trended up for most passing stats (note I said trending; doesn't mean is was statistically valid but I would rather see a positive vs. negative data trend). He still needs work on recognition especially pre-snap so he can get set quicker and get the ball out quicker. needs to read the blitz quicker. In general he needs what most young QBs need; to get the game to slow down more. He could benefit from some more talent around him, sure, but every QB can say that. He could also benefit from a more consistent offensive philosophy, but again that's consistent with many other QBs. What I do know is that his neuroplasticity is certainly not stopped at this point in his career, and that there is no reason why he can't continue improvement. Whether he does so or not is on him. we'll see starting in July.

Not just gooder, but more gooder.

My guess, and it's just a guess of course, is that the use of the Latin phrase there means the poster is a gladiator.

"Because the baseline for accuracy in those days was 60%. If you were over 60% you were accurate in the pros. Today that's creeped up to 62-65% coming out of college" The above is a direct quote from you from page 109 of this thread. So there you go on the 62% thing. Your answer confirms what I suspected; you don't know how statistics work. 58.6% is not different than 60% statistically, when looking at Allen's work this past season, because the sample size is not nearly enough to say that. And that matters. Just looking at a number and believing it means something is a trap way too many people fall into. So let's move to the following. If Allen improved from 58.6% to 72%, would I accept that? Well, let's do the analysis. A quick power analysis says that such a difference would have meaning based on a sample size of about 250, or 16 throws a game. I suspect he would meet that. So yes, that would be a meaningful improvement. Not then, let's move to this quote: "#'s are generally true because they generally are" . I have to tell you, that is just statistically silly. Laughable really. Do you have any idea how easy it is to either support or deny a given position just by playing with numbers? As I pointed out above, I review a lot of scientific manuscripts and reject many precisely because they play with numbers to try and make things seem what they are not. Your background is relevant because you claim to use stats to support your arguments. Thus, your understanding of the meaning of stats either lend credence to your arguments, or detract from them. Since you clearly do not understand stats (as so dramatically pointed out by saying numbers are generally true because they generally are), you have no more credence over anyone else around here expressing an opinion. As I said above, you're certainly entitled to your opinion. But let's not think your opinion is more factually based than others, since you don't really understand what you're talking about.

I don't believe you understand the concept of neuroplasticity, and I say that as one who has taught neuroanatomy and physiology for decades. Neuroplasticity refers to changes in synaptic connections that occur in the brain in response to altered environments, such as occurs with brain injury, or learning, or such. New synaptic connections are formed, and other lost, as the brain responds. You are implying that Allen cannot develop much because at his age he has behaviors locked in that cannot be easily changed. That is simply not true. Plasticity used to be thought of as a childhood phenomenon, but more recent research and understanding indicates that plasticity is not limited to any age. This is seen most dramatically in stroke victims, where different areas of the brain can reprogram and take over functions that were associated with the stroke-affected area. Allen certainly is not so old that he cannot learn more or be coached to perform different tasks differently. To suggest that would imply that learning ability is lost in one's early 20's, which is of course ridiculous. I will politely also point out that your concept of statistics is faulty. Just quoting stats like a certain completion rate is inadequate. It is whether such stats have any meaning that is critical, and you fall into the trap many fall into by quoting stats with no real understanding of how the stats were developed and whether their derivation is appropriate. You threw out that Allen is a worse passer that a guy with a 60% completion rate because 58.6% is less that 60%. Not true. Are your familiar with the concept of the power of an analysis? Power refers to how many observations that are needed to validate statistical differences between samples. With the number of passes Allen threw last year, there is no statistical difference between 58.6% and 60%. He would have to throw thousands of passes for the difference to have validity. You also just threw out that, for college passers coming out of college, that 62% is the number that needs to be shown for success in the NFL, with absolutely no justification for doing so. Having an opinion is one thing; you're welcome to that one for all it means, which is really nothing. But if you're going to stand up and say that you cite stats as part of the debate, then you need to show folks why the stats you cite are meaningful in any way. I review dozens of scientific manuscripts a year for professional journals, and I reject most. And the reason why is generally one of two: inappropriate statistical analysis, or poor study design that more tries to meet an author's per-formed conclusion vs. a true analysis of data. You seeem guilty of both. I have my doctorate and as pointed out above have taught neuroanatomy and physiology, and have graduate training is statistics and study design. I tend to be a nerd about that stuff around here because I think folks chronically misuse and abuse stats in looking at NFL play. If you'd like to share your background so we can accurately assess your knowledge base on things like stats and neuroplasticity, I and I am sure others would welcome that.

For the old folks, two words: Harvey Johnson

I’ve covered this before. There is no statistical difference between 58.6% and 60% given the sample size of passes thrown. You have no understanding of statistics it would appear. Similar to your throwing around neuroplasticity and intimating folks can’t learn after a certain age (and as I said above I teach neuroanatomy and physiology). Throw in your ignoring actual written material about the example you gave (Brees) and your ignorance in equating accuracy and completion percentage even though that has been debunked time and time again and let’s just say your opinions don’t carry much weight. Oh, and your arbitrarily setting new standards for completion percentages? Get over yourself.