1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

Mike Wallace being shopped by Miami Dolphins

Discussion in 'Miami Dolphins Forum' started by jim1, Feb 25, 2015.

  1. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    I think the difference between you and Jim1 is that the stats only separate "catchable" from "not catchable"; they don't tell you the degree to which it was not catchable.

    I admit I don't watch Steelers games as much as Dolphins games (and I don't have Steelers game tape), but I also don't remember too many deep balls by Roethlisberger that were so far off mark when the receiver was wide open. Granted, this is an impression based on imperfect memory of a smaller sample size, but still.. there is something different.

    Point is, Tannehill could easily be worse at downfield passing if your stats were based on the degree to which a ball is catchable instead of just a binary number. It goes to show why you have to be a bit careful with stats. Basically, in a nutshell, you like objective analysis even if it's based on poor input data, while others might choose subjective analysis that's based on richer input data. Which is better is objectively speaking hard to say (it can be measured, but who's measured it?).
     
    jim1 likes this.
  2. Fin D

    Fin D Sigh

    72,252
    43,684
    113
    Nov 27, 2007
    Your conclusions are based on self admitted even poorer data based on "an impression based on imperfect memory of a smaller sample size".
     
  3. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    Yeah, I just said that. The impression still remains, and a lot of impressions that were initially not represented in statistics (because the analysis was not good enough, or the input data was not rich enough) were later supported by "objective" analysis. Like I said, which is better is often hard to ascertain, unless you measure it. How? Either by doing a better analysis (in this case with richer input data), or by comparing predictive power (who's more accurate in predicting the future?).
     
  4. jim1

    jim1 New Member

    5,902
    3,054
    0
    Jul 1, 2008
    THANK
    YOU

    There's more to it than catch/no catch. What happened, how it happened, drop, great defensive play, great throw, average throw, lousy throw, good or bad pass pro. These things you can see, analyze and process if you actually WATCH THE PLAYS. If you don't your'e nothing more than a stat monkey with a hollow understanding of what's really going on.
     
    Piston Honda likes this.
  5. Tannephins

    Tannephins Banned

    1,818
    572
    0
    Dec 23, 2014
    Certainly plausible, but that would presumably also be evident simply in the percentage of downfield passes that were catchable. In other words, if Tannehill is so much poorer a downfield thrower to Wallace than Roethlisberger in a similar offense, based on the degree to which he "misses the mark" when the pass isn't catchable, then Tannehill's inferiority downfield would also be evident simply in the percentage of catchable balls he threw in comparison to the percentage thrown by Roethlisberger. But it isn't. Tannehill throws a greater percentage of catchable balls to Wallace when they play in similar offenses. That doesn't jibe with the idea that he's more "off the mark" than Roethlisberger when he misses.
     
  6. Tannephins

    Tannephins Banned

    1,818
    572
    0
    Dec 23, 2014
    And we still don't have the plays. In the absence of that, I'll go with the stats. Certainly you can't say the idea that Tannehill misses the mark to a higher degree when throwing to Wallace downfield than Roethlisberger did in 2012 is supported, when nobody here has watched the video. Just because the stats for that idea don't exist, doesn't mean the alternative explanation (that Tannehill misses the mark by a wider margin) is "proven."
     
  7. jim1

    jim1 New Member

    5,902
    3,054
    0
    Jul 1, 2008
    First of all, that's a bunch of hogwash. Second of all, your insistence that only deep passes thrown by Roethlisberger to Wallace in 2012 are relevant as per RT/Wallace due to similar offensive schemes is flawed and rather irrelevant. The problem that you're having is that if you loosen your death grip on it your argument tumbles like a house of cards in the wind. Deep passes are deep passes, and if only the data from 2012 supports your argument- maybe you should find a better argument.
     
    Piston Honda likes this.
  8. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    The bolded part is possible but not implied. First of all, the distributions could look similar up to the threshold at which the current stats say a thrown ball is "catchable" (so x-axis is a continuous measure of "catchability", and y-axis is counts), but dissimilar beyond that.

    Second of all, the distributions could actually look very different within the region of "catchable", and yet your statistics would show no difference. Because it's a continuous measure, there could be skewness that's not apparent in your stats. For example, if within r meters radius is deemed catchable, but most of QB 1's throws fall within r/2 radius, while QB 2's throws are more evenly distributed within that radius, your stats show no difference, but I think we would all agree you'd rather have QB 1 because QB 1 increases the probability that a lesser WR will catch the ball.

    Hope that's clear.
     
  9. Tannephins

    Tannephins Banned

    1,818
    572
    0
    Dec 23, 2014
    The argument is that Wallace had a precise mix of ingredients (quarterback and offensive scheme) that enabled his best trait -- speed -- to make him succeed. That lasted three years (2009 to 2011). In the subsequent three years, when either the quarterback or the offense (or both) haven't been present, he hasn't succeeded at nearly that level. There isn't anything "house of cards" about any of that when you consider all the data.
     
  10. jim1

    jim1 New Member

    5,902
    3,054
    0
    Jul 1, 2008
    No one said that it is "proven", correct?

    The point is that if we could actually analyze the film, we'd have a much clearer picture of not only what happened but what we're really looking for, why and how it happened. That would be far superior to a Shouright style and rather vacuous argument based on pure statistical analysis.
     
  11. Tannephins

    Tannephins Banned

    1,818
    572
    0
    Dec 23, 2014
    That's clear. Now consider, however, that neither quarterback completed a greater percentage of catchable passes than the other in those two years.
     
  12. jim1

    jim1 New Member

    5,902
    3,054
    0
    Jul 1, 2008
    Dude- Wallace had 10 TDs and almost 900 yards in a season in which his QB was HORRENDOUS at throwing deep passes. Is there a stat technique that would calculate what Wallace's stats would have been if Tannehill didn't straight up botch so many deep balls? How do your stats account for that?

    Wallace should have had a MONSTER year in this offensive system, and if you actually WATCH THE PLAYS IN QUESTION you'll see that RT straight up screwed the pooch on a bunch of throws in which Wallace was WIDE OPEN and upheld his end of the bargain.
    Wallace did what was necessary to have a great year, RT failed him. All that you have to cling on is a faint argument about 2012 and the offensive scheme that year in Pitt with out having any true understanding of what happened. You're flying blind, going on stats alone- that's a fail.

    Also, Wallace is on track for about 12,000 yards and 100 touchdowns in his first dozen years, even though RT has held him back numbers wise the last couple of years. I'm curious- from a statistical perspective, if Wallace can maintain that pace for the next 6 years, are those Hall of Fame numbers?
     
  13. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    Tells us nothing about the distribution of catchable passes if you measure it on a continuous axis (for simplicity, in case someone wants to actually make the measurements, let's just say you measure distance from some point on the body). Since the distributions are unknown, we can't say anything about how good the different QB's are if we had the richer dataset and used a continuous measure of deep throwing ability (some function based on the radius away from the body at the time at which the probability of catching the ball is maximized).
     
  14. Tannephins

    Tannephins Banned

    1,818
    572
    0
    Dec 23, 2014
    Tannehill was actually right at about the league average in deep passing accuracy in 2014. He was at 37.7%, and the league average was 39.8%. The average amount of variation from the average was 8.5%, and so Tannehill was well within the average range.
     
  15. Tannephins

    Tannephins Banned

    1,818
    572
    0
    Dec 23, 2014
    Certainly, if you want to be exact and publish a paper on it. But surely for our purposes here, you must agree that the available data -- comprised of both percentage of catchable balls, and percentage of completed catchable balls -- are inconsistent with the idea that Tannehill throws a deep ball that's more off the mark than Roethlisberger's was in 2012 when both QBs missed.
     
  16. jim1

    jim1 New Member

    5,902
    3,054
    0
    Jul 1, 2008
    How did Tannehill rank in terms of missing wide open receivers on deep passes where the receiver blew past the DB and was open for an easy catch? To use a basketball metaphor here, we're not talking about missing 3's from the corner as much as we're talking about missing layups. The stats that you just provided don't provide insight on how open the receiver was, how good or bad the coverage was, how good or bad the pass pro was, and of course how good, bad or pathetic the throw was. These factors are critical, and the fact that your stats don't address them is a fatal flaw as to their stand alone effectiveness. Cbrad is actually on track here, your basic statistical analysis, as per this discussion, is a fail.
     
  17. Fin D

    Fin D Sigh

    72,252
    43,684
    113
    Nov 27, 2007
    Right, but that means your stance or "impression" carries significantly less weight. Its literally based on nothing.
     
  18. jim1

    jim1 New Member

    5,902
    3,054
    0
    Jul 1, 2008
    You should read that post again and again until it sinks into your head as to why you're wrong.
     
  19. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    Our purpose here is to determine how good/bad Tannehill's deep ball is with respect to other QB's. The specific argument Jim1 put up is about Tannehill missing wide open throws. Clearly, your stats don't distinguish between misses when the WR is wide open vs misses when the WR is tightly covered. Clearly, your stats don't distinguish between a ball that hit the fingertips or even the chest but was dropped vs a throw that was 3 yards off from the fingertips.

    So I'd say that for the purposes of the argument here, the stats you're quoting are useful (they're definitely useful.. imagine the debate if the stats supported the game tape arguments), but insufficient. What would be sufficient would be knowing the distributions of how far the ball was at the time the ball was most catchable, for throws beyond X yards ("deep" balls).
     
    jim1 likes this.
  20. jim1

    jim1 New Member

    5,902
    3,054
    0
    Jul 1, 2008
    No, it's based on his incomplete recollection of what happened, which to his credit he admits, and actually all of us do in this context iirc. He's making plenty of sense so far imo, taking statistical analysis beyond its skeletal framework of just pass/fail analysis into something deeper and better, something that would paint a clearer picture of what actually happened and why it happened.
     
  21. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    No, you could theoretically measure how much weight it carries. I just told you how. Either do the same type of analysis on a richer dataset (sufficiently rich to answer the question at hand), and then compare which view is more accurate, or compare predictive power of the two individuals using different approaches. We don't know which is better or which should carry less weight until you do that.

    Case in point: sabermetrics in baseball. Those pushing sabermetrics were initially like Tannephins: objective is ALWAYS better, even if you can't get all the data you want. Against them were scouts, who (partly because they didn't know stats) said subjective with richer input data is better. After the egos of both sides had enough time to calm down, we see the state things are in today. Essentially no baseball team ignores stats (just look at how they adjust defensively based on stats for tendency to hit towards different directions of the field), AND no baseball team goes so far as the sabermetrics guys wanted. The stat-based approach is not that influential in rating pitchers for example.

    So you can't tell which is better: objective analysis based on poor input data vs. subjective based on rich input data, until you measure their efficacy, and no one's done that for the discussion we're having on Tannehill's deep throws (or for that matter a lot of other cases like this).
     
  22. Tannephins

    Tannephins Banned

    1,818
    572
    0
    Dec 23, 2014
    And what evidence is being offered to provide all that (the bolded portion above) in a systematic manner?

    If none of that is being provided systematically, then you're essentially asking us to trust that a fellow who calls himself "jim1" on a message board -- whose real identity and level of expertise is unknown -- has an inside handle on reality based on his subjective watching of the games, and the fact that Ryan Tannehill's downfield accuracy was measured as no different statistically from the league norm is meaningless.
     
  23. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    You don't have to trust Jim1. This is just a discussion after all. But it's worth pointing out professional scouts rely a lot on game tape. In other words, you can't attack the method as ineffective or less effective just because it's subjective.
     
    jim1 likes this.
  24. Tannephins

    Tannephins Banned

    1,818
    572
    0
    Dec 23, 2014
    The statistics that have been presented are inconsistent with the notion that Tannehill missed a greater percentage of wide open throws than Ben Roethlisberger in 2012, or than the average QB in the league in 2014. The statistics don't measure that notion in the most direct way possible, but they don't support the notion either. You're doing in an intellectual debate about data collection and measurement without recognizing that there is objective information already there that just isn't consistent with the idea.
     
  25. Tannephins

    Tannephins Banned

    1,818
    572
    0
    Dec 23, 2014
    And what scouts do is watch lots and lots of game tape, about lots and lots of teams and players, devoid of any emotional investment in any of it. What we do here is watch lots and lots of one team and its players, heavily emotionally invested in that team's outcomes. You can't possibly deem those two methods of investigation as anywhere near equivalent, in terms of the level of expertise, the breadth of the perspective (i.e., league-wide versus a single team), and the freedom from subjective bias involved.
     
  26. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    The stats are definitely consistent with the notion that Tannehill missed a greater percentage of wide open throws than Ben Roethlisberger in 2012, or the average QB in the league in 2014. Again, consistent just means "logically possible" here. It does NOT mean "evidence for" (evidence for would mean you can show one is more consistent than the other, in a probabilistic sense, or one is consistent and the other is not.. we don't have that here).

    I mean you don't even know which throws were to "wide open receivers" (your stats are not rich enough), so how could you claim it's inconsistent?

     
    jim1 likes this.
  27. jim1

    jim1 New Member

    5,902
    3,054
    0
    Jul 1, 2008
    My contention is that Wallace did an admirable job getting open last year, specifically on deep balls, and that Tannehill was not only statistically ineffective as per deep throws, he was particlularly ineffective when you FACOTR IN HOW OFTEN WALLACE WAS WIDE OPEN DEEP AND THEREFORE GAVE TANNEHILL A FANTASTIC OPPORTUNITY FOR COMPLETING THE DEEP PASS. And then Tannehill, as fans and the media noted en masses over the course of the year and Tannehill admitted himself, blew a lot of the throws. I'm not asking you to trust my opinion, but I would suggest that you actually watch some film so you can have a clearer idea of what actually happened. A statistical review of missed throw/made throw doesn't cut it here, not even close. Cbrad is admirably trying to lead you down a clearly better path.

    Don't trust my eyes, watch the film yourself instead of relying on statistics that hardly tell the whole story.

    And for the record, did you used to post on this site or any other as Shouright?
     
  28. Fin D

    Fin D Sigh

    72,252
    43,684
    113
    Nov 27, 2007
    I don't need to do a theoretical analysis of your impartial and undocumented recollections.

    Yes, Tannephins numbers can be more in depth, but at least there's the pass/fail data to consider. With your stance there is nothing and you're giving them similar weight, while calling for him to give more in depth info. That's......interesting......
     
  29. Tannephins

    Tannephins Banned

    1,818
    572
    0
    Dec 23, 2014
    Presumably the quarterbacks who are missing a greater percentage of wide open downfield throws are also missing a greater percentage of downfield throws in general. There is no logical theoretical basis for the idea that a quarterback who is poor at throwing downfield balances the downfield throws he misses to wide open receivers -- which are easier to complete -- with successful downfield throws to more tightly covered receivers -- which are more difficult to complete.

    An analogy would be someone who can't remember a seven-digit phone number, but has no problem remembering a 10-digit one.
     
  30. jim1

    jim1 New Member

    5,902
    3,054
    0
    Jul 1, 2008
    Why do you think that scouts spend so much time, energy and money going over game film, traveling to games, bowl games and combines if they could just stay at home and look at a stat sheet? Sorry dude, stats have their place but your arguments as to stats being uniquely essential and visual analysis being critically flawed because eyes lie is just weak. A type of lie is a statistic equating a bomb missing a fingertip by a half of an inch with a horrendously overthrown or underthrown pass is a type of lie, or at best an incomplete truth. As is omission of how the OL or RB was in pass pro. As is whether or not the WR dropped an easily catchable pass. As is the omission of whether or not a pass rushing DE or CB made a great play. These are all lies of omission, resulting in incomplete truths. All that you see is pass/fail based on stats, and that's why your argument fails.
     
  31. jim1

    jim1 New Member

    5,902
    3,054
    0
    Jul 1, 2008
    Sure there is, it's called inconsistency. It's part of what separates a good Quarterback from a great one.
     
  32. Fin D

    Fin D Sigh

    72,252
    43,684
    113
    Nov 27, 2007
    I'm not arguing that Tannephins numbers could be more in depth or not. The issue is that you two, who have no numbers, no objective data and only subjective data and incomplete memories of things that may or may not have happened, are dismissing his stance, because he didn't go in depth enough for you.

    That's kinda like never having worked on car, but remember a time a radiator hose went bad on an old car, so you 're asking your mechanic to prove why your new car's engine died by testing the molecular integrity of a radiator hose, when the problem is you threw a rod.
     
  33. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    Never said his stats weren't useful; they are. I'm just saying it's often not a good idea to just assume objective analysis is better than subjective analysis. And when a lot of people who watch tape on Tannehill say he has problems throwing the deep ball, that's something one shouldn't just dismiss because it's subjective. How much weight you or anyone puts on that relative to certain stats is a subjective decision also.

    Oh, and for the record, there is ALWAYS a subjective component in any kind of statistical analysis. I mean, why use the mean instead of the median? For regression (fitting straight lines to data), why minimize mean squared error instead of absolute error. The answers have to do with which is easier to use (easier to do calculations with). So, you never really have pure objective results using statistical analysis (this is a technical point, but worth making).
     
    jim1 likes this.
  34. Tannephins

    Tannephins Banned

    1,818
    572
    0
    Dec 23, 2014
    I'm not comparing the use of statistics to the use of game film of many players and many teams as analyzed by scouts, who have no emotional investment in the outcome. I'm comparing the use of statistics to the use of game film by a group of non-experts, watching a single team and its players, heavily emotionally invested in the outcome.

    In other words, we ain't scouts.
     
  35. Fin D

    Fin D Sigh

    72,252
    43,684
    113
    Nov 27, 2007
    Ugh. Stop. I'm not arguing the value of stats. I'm not arguing the value of film watching.

    I'm saying its absurd for someone who has literally no real info (objective or even valid subjective data) to say the objective data given doesn't mean much and demand more.

    He has provided a starting point, you have provided....nothing.
     
    Tannephins likes this.
  36. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    First of all, I can throw your argument back at you: you have no "objective" analysis to prove that!

    Second of all, taking into account the fact we are talking about Wallace, it seems like Tannehill has trouble estimating where Wallace is going to be X seconds after he makes his move. That's my subjective point of view yes, but that would suggest he misses wide open throws to different degrees when throwing to Wallace vs other receivers, say Hartline. Is that true? I don't know, but the point is there certainly are many possible reasons why a QB could miss wide open throws more than another QB that has similar overall stats of the kind you posted.

    Finally, which number is easier to remember: 350-2987, or 0123456789? Point is, it depends on the underlying mechanism you're using to "remember", or in the football case, estimate where a receiver will be and execute the intended throw.
     
  37. jim1

    jim1 New Member

    5,902
    3,054
    0
    Jul 1, 2008
    Try reading through the posts again- the issue as per your comments is Tannephins assertion that the only years that counts as per comparing Ben to RT as per deep passes to Wallace is 2012 because of scheme similarity. Personally, find that to be bunk and I reject it. What I would like to see, however, is film of Pitt 2012 to see what did happen and the quality of the BR throws, etc.

    I'm curious, but regardless of what the film shows, imo all the years count- you can't just pick the worst year and discard all the rest because of a scheme similarity- a deep ball is a deep ball, and I have no recollection over seeing crap deep balls from BR like I saw from RT last year.

    So basically, as to what actually happened, along with an idea of why it happened- we don't know. It isn;t all that big of a deal because we hsave the other years to consider as well, but that is the gist of this circular argument, which is ironically fairly inconsequential.
     
  38. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    We're only dismissing his dismissal of our stance. Never said stats aren't useful. And your analogy is terrible: it assumes you KNOW what the problem is. This whole discussion exists because we don't know.
     
  39. jim1

    jim1 New Member

    5,902
    3,054
    0
    Jul 1, 2008
    Again, did you ever post anywhere as Shouright?
     
  40. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    First, you need to read what I wrote to make sure you're not misquoting. I never said stats don't mean much. I even explicitly said they're useful. I'm saying they're insufficient to make the claim Tannephins is making.

    Second, the question is whether one should take the view that subjective is always worse than objective, regardless of the richness of the input data. Even statisticians will say no to that.
     
    jim1 likes this.

Share This Page