The correlation analyses are very, very weak evidence. Extremely weak. They do not rule out what people seem to think they rule out, scientifically/mathematically/logically speaking.
OMG stat people, no one saying the 40 is useless. Some of us are saying it is just made more important than it actually is. For every guy you bring up, there's a Ted Ginn or a Lorenzo Booker.
Care to elaborate? That's not a very detailed argument on your part. If the argument is that 40 times on their own are not important in predicting success for a WR in the NFL, those stats are pretty strong evidence of that. When I pointed out the problem with Approximate Value, let's be clear that the primary problem is using AV for what it was intended for: finding the relative value of players across positions. There, the arbitrary weights are a real concern. But compare within a position and the only problem w.r.t 40 times is whether they include a variable similar to "40 times" and weight that arbitrarily. They don't, so the distribution you see is likely similar to what you'll get if you could actually measure WR value perfectly. So yeah, if the argument is that 40 times on their own are important for predicting NFL success of WR's, then that correlation analysis is pretty strong. And it seems to be counter to the intuition of some (like that of ToddPhin when he said there would be a strong correlation if you compare 4.55+ to 4.45-). If the argument is that 40 times in conjunction with other stats provide a meaningful predictor of NFL success, then that correlation analysis doesn't rule it out, but you gotta show there is a way to use 40 times in conjunction with other stats to: 1) predict NFL success well, and 2) where the "40 times" variable has a huge influence on the prediction. Haven't seen that so far.
Count me in the crowd who had no idea Gilyard was a 4.6 guy. I was thinking he was in the 4.4's and would've made the mistake of drafting him in the top 50 if not for his eye-opening forty time. People here aren't even in the right ball park with the foundation they're using to establish a correlation or lack thereof. Who gives a crap about a superficial general comparison between 4.2, 4.3, 4.4, and 4.5 players that involves not a shred of context? Do you? I know I don't. Of course it won't measure anything in and of itself. That would only apply in a world where every player at their respective positions were of the same size, shared the same skill set, and played the same sub-position in the same scheme. The proper correlation, however, involves functional speed vs nonfunctional speed. Does the player have the requisite speed to match his perceived NFL skill set and ability level? That's the question the 40 helps to answer. Did any of us ding Mike Evans b/c he didn't run in the 4.3's? Of course not, that'd be idiotic. Is he any less of a player than Brandin Cooks b/c he didn't match Cooks' 4.33? Of course not, his game isn't built around speed like Cooks' is. That doesn't mean his forty time doesn't need to be measured though, as he still must display sufficient speed to match both his skill set and perceived caliber at the NFL level. Evans' 4.53 paired with his ominous 6'5 230 size offered enough confidence to take him 7th overall. Had he instead been a slower 4.65 guy, the caution flag would've been raised and his draft stock would've deservedly slipped. According to Brad's silly theory, the 5'10 190 Cooks would still be the same player at 4.55 as he is at 4.33, when the reality is a 4.55 version of Cooks is basically Davone Bess who is obviously NOT a 1st round caliber player. A 4.5 version of Mike Wallace commands ZERO safety help over the top, so if you think you're drafting a Mike Wallace for his speed, you obviously need to KNOW that he indeed does possess the level of speed you're looking for. Obviously all these bigger receivers and corners don't need to be as fast as their smaller counterparts, so that alone discredits anything Brad is attempting to show. He needs to break it down rather than use such a meaningless, broad generalization. He needs to demonstrate there's no correlation between success and big receivers running sub 4.6 compared to those running 4.6+. Ditto for average-sized receivers running sub 4.5 compared to 4.5+, and for undersized receivers running sub 4.4 compared to 4.4+. Ditto again for Will LBs who run sub 4.65 compared to 4.65+, Mike LBs running sub 4.75 compared to 4.75+ [especially if they're meant to be sideline to sideline or play in a Tampa-2], pass rushers at sub 4.8 compared to 4.8+, safeties at sub 4.6 compared to 4.6+, TEs at sub 4.8 compared to 4.8+, and H-backs at sub 4.7 compared to 4.7+. I'm just ballparking these numbers but they should be close enough to the averages to support what we mean.
Now you're moving the goalposts. The original argument was that measured 40 times even with NO other context (this was implicit) were important and that arguing otherwise had little merit. I showed that wasn't the case. You yourself claimed there would be a correlation for the exact type of stats I was presenting (so no other context) if one looked at 4.55+ vs. 4.45-. There's nothing wrong with being wrong about that. Just admit it and move on to the bigger question of how 40 times could be useful when combined with other stats. Oh, and please note I qualified what those stats actually show.
Baalke went on to discuss how important the 40 time is later on in that article http://www.49ers.com/news/article-2...spective/2263a098-32f5-4926-8768-94ca0606c4e5
The problem with the study is that it is an incomplete sample. It only measured WRs who make it onto on NFL team. What percentage of 4.3 college WRs make it into the NFL compared to the percentage of 4.5 college WRs? If 40 time has no correlation to NFL performance why arent there any 5.0 WRs in the NFL? To find the correlation properly you would need to get the 40 times of all WRs eligible to enter the NFL draft, not just those who made it onto on NFL team.
Yeah, I pointed that problem out in my first post. However, it's not that much of a problem in practice. Why? Because scouts and others that do think 40 times are important are arguing it's important in determining relative draft position, meaning all that can be made conditional on being drafted in the first place. No one seems to argue that 40 times are important only up to some threshold (= slowest 40 time for a drafted NFL player at some position), but not important beyond that.
By stating that there is a cut off speed, you are acknowledging that straight line speed, as measured by the 40 times is a desirable aspect for a WR to have. As a thought experiment if you have 2 WRs identical in every way except one has 4.2 speed and the other has 4.5, who is going to be the more valuable. «NB The only way this can be done is in a video game environment. If the hypothesis is that both are equally valuable (40 times are not relevant) then both will have similar careers. But if speed is important then the faster player will produce more, and be drafted higher. I haven`t done a data dive, but I expect that almost all 4.1 and 4.2 WRs get drafted, most of the 4.3s, up to half of the 4.4s and a few of the 4.5s and almost none of the 4.6s. Another way to look at it is the common phenomenon of veteran WRs drastically reducing production after injury or age reduces their straight line speed. If we accept that higher speed is more desirable then why does it not show in correlation, which comes back to the incomplete sample, and why this is important. I don`t think anyone is arguing that pure staright line speed is the only component of a WR. There`s also agility; catching ability; vertical ability (combination of height and vertical leap); mental discipline (route running); intelligence (ability to understand the playbook, diagnose what to do on broken plays); and work ethic. maybe a few more, such as physicality (Look at Ted Ginn Jr.) If teams are properly evaluating that mix of talent then any one component should have a zero correlation. They weigh up the competing factors and make decisions based on the mix of skills available. A positive correlation would mean that NFL teams undervalue a characteristics. A negative correlation would indicate that NFL teams overvalue a characteristic. What the study shows to me isn`t that 40 times are irrelevant. More correctly it means that NFL teams as a group are good at evaluating how important speed is for WRs compared to the other skills they have. I remember reading studies back in the 90s when analytics was basically unknown in the NFL where some positions had a negative correlation to 40 time, indicating that before more modern analysis teams would overvalue combine numbers.
Yeah, let's address this one at a time. So, yes the existence of a cut-off speed is evidence that scouts value 40 times. That however is irrelevant to whether their valuation is accurate, which is the entire point of this debate. Yes, I'd assume that if ALL other things are equal, then faster speed increases the value of a WR and there should be some positive correlation. But not all other things are equal. This alone is why you would not necessarily expect there to be a correlation between 40 times (on its own) and WR success. You make some key logical errors after those statements: 1) Even if teams are properly evaluating the mix of talent, that doesn't mean you should expect the correlation of any individual component to be zero. It's actually very rarely the case in nature that this occurs (usually there is at least one, and often multiple, factors that on their own show strong correlation). So this stuff about positive correlation for that one component => undervaluing etc.. is complete nonsense. 2) This study says nothing more than that there is no correlation between 40 times and NFL success of WR's when you use Approximate Value to measure success. That's all. It has no implications whatsoever about whether NFL teams "as a group are good at evaluating how important speed is for WR's". To show that, you'd have to know the optimal way of combining the information available and determine where on the scale NFL teams are relative to optimum. 3) Decreased production by veteran WR's is not necessarily due to reduction in 40 times. It could be due to all those other factors you listed, so there's no implication from that observation. Finally, and I'll say this again, the incomplete sample issue is irrelevant as long as the hypothesis in question conditions on the player being drafted, and since scouts that value 40 times think it's a way of determining relative value among potential draftees, the sample is perfectly fine.
This... The 40 yard dash is an accessory, which can help put film into context. It can help put into perspective whether a guy has enough speed to compete at his position in the NFL. The problem I have with the 40 yard dash is it often becomes much more than a contextual tool. The 40 yard dash becomes the primary selling point. As you mentioned, someone like Ted Ginn can be elevated to top 10 pick despite the fact he doesn't have the skill set required to ever live up to that draft position. It happens far more often than not. See the 2015 NFL draft with Dorsett and Perriman, both elevated to first round picks because of their 40, both of whom are going to have to overcome a lack of discipline in their craft and skill set to justify the picks. I'm not saying they can't, but it's a coin flip at best. And as for the examples of players running poor 40's and not living up to their college game in the NFL, how can anyone say it's because they ran 4.6 and not 4.4? What were the other times in the other drills? What were their shuttle? Cone? Splits? ****, Jarvis Landry ran a 4.6 at his pro day, and his team mate Odell Beckham Jr. ran 4.43, but their 20 yard split was damn near identical. OK, maybe Jarvis isn't built to run a 9 route, but does that mean he can't have a successful career? Is the 9 route the be all end all? The 40 is absolutely overrated. Not meaningless, but overrated. As with every measurable, the specific measurable itself should be viewed in combination with other measurables. What were the 10 yard, 20 yard, and 30 yard splits? What were the cone and shuttle times? All equally important. Player skill set, discipline in mechanics/craft, and measurables are meant to be viewed together, not separately. All of them little puzzle pieces. p.s. The most useless combine event? The Bench Press... Absolutely one of the worst lifts to measure pure strength. Any lift measuring pure strength has to include lower body strength too. Even as a measure of upper body strength, the bench press is still ****. I have no idea why the NFL hasn't gotten rid of it and replaced it with something more reliable like the overhead press. And besides that, benching a stagnant 225 is more a measure of muscular endurance than strength once you're beyond 12 reps. If you want to measure pure strength? Make the players deadlift and overhead press a weight of their choosing for 5 reps. ??? x 5. Much more accurate.
Great post Clark,Maybe they should also have them do leg presses to test their leg strength which is also important for the offense, and defensive linemen. Also, what is most important to me is their play speed in pads, as was stated earlier, Rice' 40 time was not great but damn if he did'nt always get open, and as stated in an earlier post, we hardly ever saw him caught from behind. I can't wait to watch Landry this year with his improved speed, which was in another thread about his training right now....I know it is very early in his career, but if he keeps improving his game, when his career is over, maybe he gets in the same sentence as Rice, somebody eventually will and it will be the very skilled receiver with all of the intangibles that does it. Just think, who actually thought that most of Marino's records would have fallen so quickly..I mean it was a long time before he came along and broke Tarkenton's records..so yes...give me the guys who play speed in pads is quick and fast over the track stars.
