Defensive Statistics and Player Attributes - Cont2 - Printable Version +- International Simulation Football League (https://forums.sim-football.com) +-- Forum: Community (https://forums.sim-football.com/forumdisplay.php?fid=5) +--- Forum: Media (https://forums.sim-football.com/forumdisplay.php?fid=37) +---- Forum: Graded Statistical Analysis (https://forums.sim-football.com/forumdisplay.php?fid=153) +---- Thread: Defensive Statistics and Player Attributes - Cont2 (/showthread.php?tid=2535)

Defensive Statistics and Player Attributes - Cont2 - Molarpistols - 07-23-2017 Yesterday, I did some regressions examining the correlation between player attributes and their effect on the players total number of tackles. During the gathering of the data for that one, I also gathered data for most every other defensive statistical category. Today, I'll examine Interceptions/Pass Defense. and in a separate article, Sacks. for Interceptions and Passes Defensed I excluded the Defensive Linemen, because it wouldn't add much to the analysis. For the sacks, I excluded the Defensive Backs, because most DBs didn't record many sacks, and including them would further cloud the already cloudy outputs from my regressions. I did include linebackers in both examinations because LBs spend a fair amount of time rushing the passer and in pass defense, leading to more statistical data points for my use. Interceptions Much like sacks, interceptions are pretty rare (unless you're playing against my QB on the Otters . I give him a lot of shit, but I just think it's funny, I'm not ripping him to start shit). As they are rare, the occurrence of them is lower and lead to statistical regressions that are not statistically significant for the most part, when using the 95% confidence interval that is most common. We'll start right off with the DBs and their interceptions. Speed, and interestingly enough, strength appear to be important for DBs to earn interceptions. The speed one makes sense, but is only significant at the 90% confidence interval. Strength is nearly significant at the 95% interval, which is quite interesting. The only logical explanation I can come up with is that it makes your player better at contested catches and positioning for the ball. Hands seem far less important than I would've thought they'd be for interceptions, but the stat is not significant, meaning we can't take it as law. Tackling, Intelligence and Agility all have negative coefficients, which seems a little odd, but the only one of those I'd expect to actually help with interceptions is intelligence. Linebackers got interceptions as well, and the above is the regression I ran on them. I think we can safely say we could throw out this regression with the bathwater and not feel guilty about it. None of the coefficients, nor the f-stat are very close to being statistically significant. The closest to significant has a negative coefficient, in one of the attributes I would've thought would be most important. Speed should be needed to keep up with RBs, TEs and slot WRs, but supposedly it hurts a LB's chances of getting an interception. Needless to say, we didn't learn much from this one. The output including both DBs and LBs is a little better than our LB one. It's not significant by F-stat, and few of the attributes have a coefficient that is significant, but some are at least close. Finally, hands seems important towards interceptions. The coefficient is significant statistically, and it is positive, even though it's fairly small. Strength is still surprisingly high with a p-value that is nearly significant. Again, my only logical explanation is it helps with positioning for the ball and contested catches. Passes Defensed Good coverage skills don't always translate into interceptions. I examined passed defensed as well for both defensive backs and linebackers. Starting with the defensive backs: I feel like a broken record player, repeating over and over. The regression is not significant, and most (all) of the coefficients are not either. Speed is important for defending passes, showing a very high coefficient and a nearly significant p-value. Keeping up with WR's leads to defending more passes, who would've thunk it? Like with interceptions, strength is surprisingly high again, but is not significant statistically. Hands is helpful, but not significant either. Our playing time is actually positive for one of the few times, which was fun to see. Oh my goodness, we've got a live one! This regression is significant statistically, and a lot of the coefficients are too, I'm astounded. Strength is the highest coefficient (again, weird) and is significant. Either it's important for positioning and contesting, or there is something else going on. Perhaps the strength stat itself is too high or low as a base archetype and could be wrongly giving the impression of importance on our interceptions and passes defensed stats. Speed is important, which our linebacking interceptions regression did not show. Intelligence and agility are also significant and almost as important as speed in defending passes. Hands being negative is a little odd, especially since it's statistically significant. I suppose if you have good hands as a linebacker, you make the interception instead of batting it down? I don't know, I can't logically explain that one better. The final regression, with both DBs and LBs present. The f-stat is nearly significant, but unfortunately is not at a 95% confidence interval. None of our coefficients are significant in this one. Strength, again, seems important as a whole. Speed is more logically important, and has a higher coefficient than strength, but a worse p-value (though they're both quite close on both counts). Intelligence showing as a negative seems a little odd. Thankfully we can explain that away as insignificant as that p-value is absolutely terrible. Overall, statistically speaking, we can't take too much of these readouts as accurate. The significance tests for the most part, failed. We had some weird negative coefficients where we shouldn't have. Strength seems important, which I think may be a fault with the statistical technique, or it's important for ball positioning and contesting. As expected, speed is important across the board, and while the stat here doesn't show as significant often, it makes logical sense as being important, so I think we can believe it for the most part. Again, the disclaimer at the end. I'm not a statistician, just a simple economist that dabbled in some regression analysis in college. We need more data! More seasons, and more teams with more players will certainly help. I hope you all enjoyed, and at the very least found this series interesting! GRADED