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

Offensive Stat Improvement without Ajayi

Discussion in 'Miami Dolphins Forum' started by Galant, Nov 24, 2017.

  1. Galant

    Galant Love - Unity - Sacrifice - Eternity

    19,127
    11,058
    113
    Apr 22, 2014
  2. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    Yeah good observation in that article. From a purely stats point of view there is a significant difference before vs. after Ajayi in total yards per game and passing yards per game but not in rushing yards per game, which actually might make sense.

    I'll just point out that one should look at differences in sets of games instead of just differences in averages because differences in averages without knowing sample size or variance can be very misleading. For example, consider how we've fared on defense since the Ajayi trade (I'll list the offensive stuff too for reference):

    Points scored: 13.14 ppg before Ajayi, 21.67 ppg after Ajayi
    Points allowed: 21.74 ppg before.. 34 ppg after
    So we're scoring 8.5 ppg more, but allowing 12.3 ppg more since the Ajayi trade

    Total offense: 252.43 yards before, 383.33 yards after
    Total defense: 306.29 yards before, 416 after
    So we have 131 yards total offense more after Ajayi, but give up 110 yards more yards on defense

    Passing yards: 176 before, 295.67 after
    Passing yards allowed: 210.86 before, 272.33 after
    So we pass for 120 yards more but give up 62 yards more per game

    Rushing yards: 76.43 yards before, 89.67 yards after
    Rushing yards allowed: 95.43 before, 143.67 after
    So we rush for 13 yards more per game, but give up 48 yards more.

    As you can see, if you just look at differences in averages without knowing sample size or variance, it actually looks like the Ajayi trade affected defense as much as offense. But surprisingly there is no statistically significant difference between games 1-7 and 8-10 in any of those defensive stats.

    There IS however a statistically significant difference in points allowed between games 1-4 vs. 5-10, games 1-5 vs. 6-10, and games 1-6 vs. 7-10. In other words, it's like something changed on defense before the Atlanta game, before the 2nd Jets game and before the Ravens game. But not after the Ravens game (Ajayi trade) or before the Atlanta game. So it's really best to compare sets of games instead of just averages.
     
    Last edited: Nov 24, 2017
    finsfandan likes this.
  3. jdallen1222

    jdallen1222 Well-Known Member

    2,752
    1,373
    113
    May 31, 2013
    Plantation, Fl
    Snatched from the aritcle:

    upload_2017-11-24_13-13-38.png
     
  4. KeyFin

    KeyFin Well-Known Member

    10,488
    12,821
    113
    Nov 1, 2009
    While I do see your point, sometimes the best answer is the simple one. And in this case, it's fairly safe to say that the offense is doing better with running backs that listen to their coach and block for their quarterback.
     
    The Finest and vt_dolfan like this.
  5. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    Hmm.. I hope my point was clear though: the simple answer would suggest that the Ajayi trade hurt the defense as much as it helped the offense because if you just look at differences in averages you could argue that. So I do think you should go further than just looking at averages.

    Oh.. btw.. the other reason you want to go further than just look at averages is because someone could look at how bad the defensive has gotten since Ajayi and argue that similar levels of offensive improvement could be due to random chance because of the small sample size of just 3 games. In both cases, an actual statistical test on sets of games resolves the issue.
     
    Last edited: Nov 24, 2017
  6. djphinfan

    djphinfan Season Ticket Holder Club Member

    111,648
    67,540
    113
    Dec 20, 2007
    no more one dimensional running backs..cant pass protect, cant catch and run, can't run routes, you a tell to the defense.
     
    SocaCasualuk and Rock Sexton like this.
  7. hitman8

    hitman8 Well-Known Member

    3,040
    2,485
    113
    Nov 11, 2016
    Actually, ajayi was the best pass protector among the running backs. I'll give you that he wasn't the best receiver, but pass protection was not his biggest problem. He is actually better than williams and drake in terms of pass protection.
     
    finsfandan likes this.
  8. dirtylandry

    dirtylandry Well-Known Member

    4,214
    1,750
    113
    Aug 2, 2015
  9. jdallen1222

    jdallen1222 Well-Known Member

    2,752
    1,373
    113
    May 31, 2013
    Plantation, Fl

    They weren't that great this past game. Cutler got lit up or threw it away multiple times.
     
  10. Brasfin

    Brasfin Well-Known Member

    2,435
    1,672
    113
    Apr 27, 2013
    Brazil
    Not according to Gase. In fact he basically called out Ajayi that "the backs" never knew who to block... then Ajayi was traded. Newsflash: he was talking specifically about Ajayi.
     
    SocaCasualuk and resnor like this.
  11. KeyFin

    KeyFin Well-Known Member

    10,488
    12,821
    113
    Nov 1, 2009
    But Ajayi didn't play on defense, so that whole premise assumes he could hurt us by changing up time of possession, average field position or even maybe some type of sideline chemistry. I just don't buy it.

    For instance, I KNOW from film study that Ajayi was not blocking up to par. I also KNOW Moore/Cutler have been sacked twice in the past two games total, which you showed was tied for 4th best in the league in that stretch. I don't need additional stats to show that Williams and Drake block a heck of a lot better than Ajayi did...I can see that on film where they're knocking consistently rushers on their butts. And I don't think we have to debate that better blocking creates more offensive opportunities.

    Here's the other problem with your argument though- you're assuming Ajayi was the ONLY variable between the last three weeks and previously. We could list a million variables if we looked hard enough....from the new sod in the stadium to the temperature outside to how the sideline intern mixed the Gatorade. And guess what? An independent study would show the EXACT SAME RESULTS over the past three weeks- new sod made the offense better and weakened the defense. At least the sod directly affects the D, unlike Ajayi.

    I guess what I'm trying to say is that when you're going to point to statistics, you have to eliminate all the variables to make them accurate. And you just can't do that when you're talking about a player who was on the sidelines while our defense was on the field.
     
  12. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    Maybe this is worth a deeper discussion. The entire purpose of statistics is so that you can make inferences when you CANNOT eliminate all other variables. If you could eliminate all except one variable, then you don't need statistics!

    What statistics allows you to do is to calculate the probability of an event occurring or the likelihood of a hypothesis being true given observed data, in the presence of random variation (i.e. other factors making things non-deterministic in practice). Ideally, the distribution of such random variation is known, but if it's not known you can often estimate it from other data, or worst case you use a method that is known to be robust across most such distributions.

    Now.. you're absolutely right about saying we can effectively eliminate trading Ajayi as the reason for the defense getting worse post-Ajayi, but you have no way of responding to my second point: how do we know that small sample size (i.e. random variation) isn't more likely the reason for the improved averages on offense, especially when you have similar such differences on defense without an obvious causal factor? To answer that kind of question you do need to compare different sets of data, not just differences in averages, and run it through a statistical test.
     
  13. KeyFin

    KeyFin Well-Known Member

    10,488
    12,821
    113
    Nov 1, 2009
    We're going around in circles here. As I said twice now, I could clearly see Ajayi taking passing downs off and rarely making contact with the blitzing LB/DE/whatever. He would step in front of them and that's it. I can also see Williams actively scanning around him and bulldozing into the first player to break the pocket...usually knocking them on their asses.

    These observations do not require any statistical breakdowns- your numbers verify what many of us already see and have talked about all year.
     
  14. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    Yes, in this case both approaches agree. But many times they don't, and that alone is reason enough to dig deeper into the data to see which hypotheses are actually supported.
     
  15. hitman8

    hitman8 Well-Known Member

    3,040
    2,485
    113
    Nov 11, 2016
    No, he was referring to williams, his only gripe with ajayi was that he was freelancing and not hitting the hole he was supposed to and also complaining about not getting the ball.

    Pass protection was not one of gase's complaints about ajayi, williams is the one that has been blowing blocking assignments as he just did again this past game.
     
  16. Fin D

    Fin D Sigh

    72,252
    43,684
    113
    Nov 27, 2007
    I honestly think you have this backwards.
     
    SocaCasualuk, resnor and Brasfin like this.
  17. Rock Sexton

    Rock Sexton Anti-Homer

    2,553
    1,793
    113
    Mar 14, 2015
    No doubt.

    The difference this year with and without him is blatant. Also, did you notice the two longest runs of the season (69,66) came after he left? His longest was 21 up until that point.
     
    djphinfan likes this.
  18. Dolphin North

    Dolphin North Well-Known Member

    366
    387
    63
    Apr 30, 2017
    So we are scoring more, but the other team is scoring on us even more than we are improving in points. Our time of possession is dropping. I know it is a small sample size, but I would not be surprised if the trend continues. Give it a few more games, but I would not expect this team to be able to sustain long drives without Ajayi, unless some new version of the bubble screen somehow becomes the new Wildcat. We were not dominating on the ground anyway but it is not improving, so we have taken to the air. Expect more points in a passing league, but also expect more INT's as we pass more and expect the defense to wear down even faster.
    But it is more entertaining to lose 31-24 than 21-16. The NFL has made a lot of rule changes to encourage that.
     
    Last edited: Nov 25, 2017
    finsfandan, SocaCasualuk and KeyFin like this.
  19. vt_dolfan

    vt_dolfan Season Ticket Holder Club Member

    I think it would look that way if you just looked at the numbers...however,

    It was about that time that Timmon's play began to decline, I think as a result of playing injured. Wake sacking the QB has dried up...I think Suh has been playing on a bad leg.

    I think the more likely reason is our Offense was so bad in those Ajayi games that our defense had to stay on the field much much longer. We have a blend of veteran talent, and inexperienced talent playing...I think the vets just got beat up, and the younger players didn't have enough experience to offset.

    For the most part, we all watch every Dolphins game, and I think we can all SEE the difference in our offense before Ajayi and after.
     
    SocaCasualuk and resnor like this.
  20. vt_dolfan

    vt_dolfan Season Ticket Holder Club Member

    He does...or why would we have had 16 sacks with Ajayi..and only 2 since? I might add I think Drake has done a very good job at picking up the blitz and blocking in the backfield.
     
    SocaCasualuk, resnor and Fin D like this.
  21. vt_dolfan

    vt_dolfan Season Ticket Holder Club Member

    Well...dont look at the stats page. Watch the game tapes..what do your eyes tell you?
     
    KeyFin and resnor like this.
  22. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    My eyes tell me the more plausible explanation for the vast improvement in passing stats is improved QB play. It's not hard to see how the Ajayi trade is possibly more coincidental than causal because you have such a small sample of games and Cutler just happened to play by far his best game after that Ravens debacle, and Moore played great against Tampa.

    We'll see which hypothesis holds more water in the coming games. If it's really Ajayi our passing offense should still be ranked high post-Ajayi even after another 3-6 games. But if it's the QB (as I think it is), then you'll see up and down play by the QB's (based on stats history), meaning you're likely to see regression to the mean for these offensive stats.

    Point is.. neither "eyes" nor "stats" can determine causality reliably, as evidenced by all the disagreements among people using their eyes and by statistics on its own not being able to distinguish between correlation and causation (you need a controlled experiment to determine causality). So I'm not giving credence to "just use your eyes" type of arguments in absence of statistical evidence for that hypothesis. Point is.. you NEED the statistical evidence regardless, and for that you need to compare sets of games, not just averages (this last sentence is the only argument I've insisted on in this thread).
     
  23. djphinfan

    djphinfan Season Ticket Holder Club Member

    111,648
    67,540
    113
    Dec 20, 2007
    While Ajayi was playing this year we were the most inept offense I’ve ever seen and pretty much one of the worst in history...without him things look better..

    What’s the common denominator?...... Ajayi..

    When you have a full time player at that position, that becomes such an easy tell..it’s debilitating.

    I mean it seems the most logical no?.. that his one dimensional presence made our offense become so predictable that it became handicapped.
     
    Last edited: Nov 25, 2017
    SocaCasualuk likes this.
  24. djphinfan

    djphinfan Season Ticket Holder Club Member

    111,648
    67,540
    113
    Dec 20, 2007
    But you can see the difference in the passing game with the backs..Ajayi posed no threat to a defense in this regard..you can see the differences in route running, and most importantly quickness getting into the route, getting to the designed spot, catching the ball, getting upfield with quickness, and making moves in space..this has been the easiest element of the game to detect...we didnt need stats at all to prove it, even though they do, we can all see the stark differences..
     
    SocaCasualuk, resnor and cuchulainn like this.
  25. cuchulainn

    cuchulainn Táin Bó Cúailnge Club Member

    23,698
    39,847
    113
    Sep 7, 2012
    Hattiesburg, MS
    I think the main thing that's been impacting the defense has been the offense and special teams. Since the raiders game, the offense and ST has been putting them in terrible positions. The bad calls against them like with the Jones call, some of the bad holding calls, and the safety FU, also takes a toll.

    We also reportedly just lost Hayes for the rest of the year and Wake and Branch haven't been playing well since the Titans and jets games.

    If Burke has anything he's been sandbagging, this game would be a good one to bring it out for as the pats are missing 2 starters on the OL.
     
  26. djphinfan

    djphinfan Season Ticket Holder Club Member

    111,648
    67,540
    113
    Dec 20, 2007
    I disagree completely CB..

    Not only do the eyes tell us in this case when we watch some explosive plays, productive consistent plays with our backs now coming out of the backfield and making catches, but I’m sure the stats back it up..
     
    vt_dolfan likes this.
  27. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    Yeah, good thing is this is testable. We'll see what the stats look like at the end of the year.
     
  28. djphinfan

    djphinfan Season Ticket Holder Club Member

    111,648
    67,540
    113
    Dec 20, 2007
    Your gonna hold on to your theory til the last second huh bud..?... hell, it ain’t easy winning a debate with you bro but I think you should just concede this one... tape don’t lie, regardless of sample size this looks elementary..

    I actually saw it begin last year
     
    Last edited: Nov 25, 2017
  29. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    lol.. nothing wrong with having some patience dj.

    If I'm wrong, the stats will show that and I'll do the analysis for everyone to see. If I'm right though.. then that means you're overweighting the effect of replacing Ajayi relative to say how the QB performs. Remember, the question isn't whether what you're seeing on tape is "real".. I never disputed that part. I'm just in the camp saying other factors are more important for the post-Ajayi trade stats than Ajayi himself.

    Patience dj.. nothing wrong with letting the data decide.
     
  30. Fin D

    Fin D Sigh

    72,252
    43,684
    113
    Nov 27, 2007
    No matter how the numbers end up, it won't discount what DJ saw on tape. You are making the false assumption that the numbers will either show you right or him right. In actuality, if the numbers change, it will likely be for other factors unrelated to what DJ has seen and what he saw did in fact affect, the numbers to this current point.
     
    djphinfan likes this.
  31. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    You should read posts you respond to. I just got through explicitly saying I'm not disputing what he saw on tape!!

    Either way, he's not disputing the testability of these hypotheses, so we'll see what happens. I mean.. either the differences in stats we've seen post-Ajayi are mostly due to him leaving, or they're not. We'll see.
     
  32. resnor

    resnor Derp Sherpa

    16,327
    9,874
    113
    Nov 25, 2007
    New Hampshire
    That's not entirely true, cbrad. The stats at the end of the year could be different than what we see right now, and it could be for other reasons, and what we're seeing right now could be because of the Ajayi trade.
     
    The Finest and Fin D like this.
  33. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    I'm fully aware other things can happen from now to the end of the season. Trust me, I'll do my best to tease apart what can be teased apart once we have the data. Nevertheless, these two hypotheses do generally predict different things.

    If Ajayi leaving was the primary reason for what we've seen in the last few games, then that should by and large continue (accepting random variation), but if it's mostly due to one of our unreliable QB's that can have a great game here and there but otherwise make you feel bad as a Dolphins fan lol.. well then the prediction is regression to the mean.

    Let's just wait till the data comes in and if you have a problem with my analysis then tell me!
     
  34. Fin D

    Fin D Sigh

    72,252
    43,684
    113
    Nov 27, 2007
    I read the post just fine, per usual.

    You are saying that what DJ is saying may prove to be wrong depending on the data by the end of the year....which is what I was referring to.

    And again, you're setting this up to be an "either or" scenario. Yes, i know you keep leaving little outs saying "we'll see" and so forth, but then you also double down and say things like:

    Which reinforces that you're assuming an "either or" scenario.

    I know you think I'm ****ing with you, but I'm not.
     
  35. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    Yes you are ****ing with me. Not interested.
     
  36. Fin D

    Fin D Sigh

    72,252
    43,684
    113
    Nov 27, 2007
    Your arrogance is effing stifling sometimes.

    I don't lie. I wasn't messing with you. I disagree with you cause you're not perfect. You should probably stop reading your own hype, Sport.
     
  37. Fin-O

    Fin-O Initiated Club Member

    11,375
    11,392
    113
    Sep 28, 2015
    Guy's.

    Stop fighting.

    Ohio State just beat Michigan.

    Doesn't that ease your pain?
     
  38. KeyFin

    KeyFin Well-Known Member

    10,488
    12,821
    113
    Nov 1, 2009
    Williams has been blocking AWESOME in the limited time he's stayed in the pocket on passing downs. I'm talking about 10+ plays where he sees the DE come around the corner and Williams blasts the guy full force and knocks him down. Now, given, that's not exactly what pass blocking is supposed to look like, but he's been better than any RB I've seen league-wide this year in that aspect.

    Williams is a massive upgrade to Ajayi is pass protection and releasing/catching out of the backfield...hence our increased passing stats. He's played a very big part of that.
     
    cuchulainn and Dolphin North like this.
  39. KeyFin

    KeyFin Well-Known Member

    10,488
    12,821
    113
    Nov 1, 2009
    Every game is different in hundreds of ways with hundreds of changing variables and they prep as well as they can for each of them. Sometimes, great players make your team look bad and other times, you shut them down. You're talking about statistics like they're absolutes though and define every ounce of what each player is made of.

    For instance, COULD Williams have a bad day blocking tomorrow? Yes! Will I say he's a crappy blocker after he's blocked like a stud for three weeks now? No! For those types of things I need a sample size of one...show me once that you can fire off the ball and level someone, then I'm going to trust you to do it again the next play. And that's what the guy has been doing while Ajayi wasn't.

    Drake has also been fantastic out of the backfield on runs and passes. Again, no additional stats necessary- our two backups have proven that they'll play hard and lay themselves out for Cutler/Moore. This is not a mathematical equation that needs further study.
     
  40. cbrad

    cbrad .

    10,659
    12,657
    113
    Dec 21, 2014
    That's not the question though. The question is what is MOST responsible for a given summary stat such as total passing yards. That's not something you can observe directly. As I said before, I'm not disputing what you see on tape, just disputing how influential it is to a given summary stat.

    Right now the data are consistent with both hypotheses. But suppose after the season we have 7 games with Ajayi and 9 games without him and we want to know the probability stats pre-Ajayi and post-Ajayi are due to random variation. For that you DO need math, in this case a 2-sample t-test for unequal sample sizes.

    And if that test comes out and says the differences pre- vs. post-Ajayi are statistically significant, then my hypothesis is wrong because given the past performances of our QB's, you shouldn't see that kind of change. However, "post-Ajayi" is a condition for the rest of the season, so it's necessary that test supports your hypothesis. Oh.. if the test does say the stats are consistent with random variation, we still need to dig deeper, but we can cover that later.


    btw.. if you really think I've ever said or implied statistics "defines every ounce of what each player is made of", then you haven't paid much attention. In general, statistics throws away tons of information, and a human observer will have much richer information to work with. But in so many cases (certainly in medical research where I help develop math) you make up for that massive loss of information with unbiased inferences, the ability to handle data across many different conditions, and of course no loss of memory or bias in how you remember things (common with humans).
     
    Last edited: Nov 25, 2017

Share This Page