|
Doltos posted:This is absolutely awesome. A super sperg lord on Reddit compiled the stats of most draft eligible QBs of the last 3 years: God bless reddit sometimes I'm so high on Teddy right now
|
# ? Apr 21, 2014 15:48 |
|
|
# ? May 28, 2024 08:50 |
|
Blitz7x posted:God bless reddit sometimes Has this been parsed into any conclusions?
|
# ? Apr 21, 2014 15:55 |
|
Just that Logan Thomas is loving terrible.
|
# ? Apr 21, 2014 15:59 |
|
And Teddy Bridgewater just wins
|
# ? Apr 21, 2014 16:00 |
|
Gumbel2Gumbel posted:Has this been parsed into any conclusions? Some. The first and most obvious is the data gatherer sucks at making readable tables, but there are a few other points we can glean from this. 1. Bortles has never faced top ten competition. 2. Bridgewater and Manziel perform statistically the best against top ten defenses. 3. Bortles is a much better scrambler than people give him credit for. 4. Tahj Boyd is an interception machine. 5. Manziel's accuracy is underrated. 6. McCarron, Teddy, Geno Smith, and especially Russell Wilson like to beat up weak teams.
|
# ? Apr 21, 2014 16:02 |
|
how many of those top defenses Carr faced are just lovely WAC/MWC teams that happen to rank high in defense because they are somewhat decent there and play a bunch of other lovely WAC/MWC teams
|
# ? Apr 21, 2014 16:05 |
|
Chilichimp posted:You're pretty easily swayed if you can take my vague words and not even watch the game tape and make that statement. Murray has been floated as a late round prospect for the Chiefs and you really need to understand how much Parm hates Alex Smith.
|
# ? Apr 21, 2014 16:07 |
|
Groucho Marxist posted:how many of those top defenses Carr faced are just lovely WAC/MWC teams that happen to rank high in defense because they are somewhat decent there and play a bunch of other lovely WAC/MWC teams This is also known as the problem with vague statistics.
|
# ? Apr 21, 2014 16:07 |
|
Doltos posted:This is also known as the problem with vague statistics. You've got to change that loving Avatar, I can't take anything you're saying seriously. It's got nothing to do with the content of the post, I just always glance over and see Skips mouth open and stop listening.
|
# ? Apr 21, 2014 16:18 |
|
No Butt Stuff posted:Murray has been floated as a late round prospect for the Chiefs and you really need to understand how much Parm hates Alex Smith. Tajh Boyd or bust
|
# ? Apr 21, 2014 16:21 |
|
Chilichimp posted:You've got to change that loving Avatar, I can't take anything you're saying seriously. It's got nothing to do with the content of the post, I just always glance over and see Skips mouth open and stop listening. I stopped buying new avatars after someone replaced my awesome Jaraslov Halak with some lovely custom title.
|
# ? Apr 21, 2014 16:24 |
|
It happens bro. It happens.
|
# ? Apr 21, 2014 16:26 |
|
Parmesan Basil posted:Tajh Boyd or bust I'd take either, provided they can sit and learn the offense for at least a year.
|
# ? Apr 21, 2014 16:31 |
|
Doltos posted:Some. The first and most obvious is the data gatherer sucks at making readable tables, but there are a few other points we can glean from this. Russell Wilson's TD:INT ratio is loving absurd.
|
# ? Apr 21, 2014 16:39 |
|
Chichevache posted:Russell Wilson's TD:INT ratio is loving absurd. Dat completion percentage though.
|
# ? Apr 21, 2014 16:44 |
|
I'm sure someone has done a regression analysis on that stuff. What are the metrics that correlate most closely to good NFL player?
|
# ? Apr 21, 2014 16:52 |
|
bewbies posted:I'm sure someone has done a regression analysis on that stuff. What are the metrics that correlate most closely to good NFL player? Tattoos
|
# ? Apr 21, 2014 16:55 |
|
bewbies posted:I'm sure someone has done a regression analysis on that stuff. What are the metrics that correlate most closely to good NFL player? Has anyone though? Even in the reddit thread that supplied those stats it doesn't look like anyone has tried to correlate them to NFL success. I don't even much like doing that, as Sack Seer isn't reliable either.
|
# ? Apr 21, 2014 16:56 |
|
Guys who throw over %60 or a 2:1 TD:INT against top ten defenses seem to have a pretty good track record if you don't count Kellen Moore's one game.
|
# ? Apr 21, 2014 17:02 |
|
Well If others are willing to mine the data for me I'll build a model for it. These are the stats that'd be worth looking at IMO: Games started Wins Passer rating (alternatively, break down PR into Comp%, TD%, INT% and YPA) Pass yards/gm Rush yards/gm Data for both full schedule and vs top 25 ranked teams. Choose maybe 20-30 college QBs at random from the last 15 years. Must have at least 10 games starting. edit - drafted QBs bewbies fucked around with this message at 17:05 on Apr 21, 2014 |
# ? Apr 21, 2014 17:03 |
|
bewbies posted:These are the stats that'd be worth looking at IMO: roll tide
|
# ? Apr 21, 2014 17:04 |
|
bewbies posted:Well Out of all those stats only passer rating directly correlates to the player. Games started and wins are dubious Parcell-like statistics and Pass/Rush yards a game just show a player's usage. The best case against using yards as a statistical measurement for future NFL success begins and ends with Ron Dayne.
|
# ? Apr 21, 2014 17:06 |
|
Doltos posted:Out of all those stats only passer rating directly correlates to the player. Games started and wins are dubious Parcell-like statistics and Pass/Rush yards a game just show a player's usage. The best case against using yards as a statistical measurement for future NFL success begins and ends with Ron Dayne. A large part of the point of regression analyses is to determine what data matters and what data does not and I for one am curious if there is any correlation at all to wins in college. edit - should also include height and weight too bewbies fucked around with this message at 17:12 on Apr 21, 2014 |
# ? Apr 21, 2014 17:08 |
|
Doltos posted:Out of all those stats only passer rating directly correlates to the player. Games started and wins are dubious Parcell-like statistics and Pass/Rush yards a game just show a player's usage. The best case against using yards as a statistical measurement for future NFL success begins and ends with Ron Dayne. I agree with your general point, but you have to consider that a guy who throws 40 times a game is going to be facing a different defense than a guy who throws it 25 times a game. It's a lot easier to put up efficient stats in the latter.
|
# ? Apr 21, 2014 17:10 |
|
Before you even build your model, I have a few issues to raise:
|
# ? Apr 21, 2014 17:14 |
|
bewbies posted:A large part of the point of regression analyses is to determine what data matters and what data does not and I for one am curious if there is any correlation at all to wins in college. Let me correct myself then. In statistics, if you find that NFL QBs that have 30+ wins win more NFL games than those with 20-30 wins, you cannot safely say that correlation of those QBs having 30+ wins implies the causation of them winning on the next level. That's the problem with Parcells' model.
|
# ? Apr 21, 2014 17:14 |
|
Doltos posted:Let me correct myself then. In statistics, if you find that NFL QBs that have 30+ wins win more NFL games than those with 20-30 wins, you cannot safely say that correlation of those QBs having 30+ wins implies the causation of them winning on the next level. Uh, alright. Regression only shows correlation so I guess we're ok. (point being you're probably right that "wins in college" doesn't correlate strongly to success in the NFL but I think it is still worthwhile to take a look at the relationship)
|
# ? Apr 21, 2014 17:19 |
|
I guess I missed FO's Lewin Career Forecast a couple months back but since it's on-topic, Aaron Murray lapped the field. http://www.footballoutsiders.com/nfl-draft/2014/lcf-2014-year-it-means-nothing @FO_ASchatz -I just ran LCF for Tom Savage. Lowest score in this class. And if we don't give him credit for jr-sr year improvement, it's near zero. -Savage strikes me as EXACTLY the kind of QB you want to avoid: low comp rate, workout wonder, rising just a few weeks before draft. -If Tom Savage was ready to be an early-round NFL QB selection, wouldn't it have been seen on the game tape four months ago?? Diva Cupcake fucked around with this message at 17:33 on Apr 21, 2014 |
# ? Apr 21, 2014 17:30 |
|
bewbies posted:Uh, alright. Regression only shows correlation so I guess we're ok. I had more of a problem with the stats you picked instead of the idea to create the model. I believe that most QBs that are even being considered to be drafted can throw a decent enough ball if left alone in a void. I rationalize this by saying most if not all pro days tend to show the draftable QB in a favorable light. What I think really hurts them in the NFL is not their good decision making, but their bad decision making. Therefor I think throwing out TDs and yards gained should be the first step, as both seem to correlate directly to usage over actual tangible skill. An example in this draft would be Derek Carr leading everyone with 41 attempts and 2.9 Tds a game. I'm not saying Carr is a bad QB, but I know from just watching the dudes that he's a step back from Manziel and Bridgewater. So, I think the first step should be looking at completion percentage and INT percentage, and see how that translates when they head to the NFL. The next step should be looking at their passing charts to see where they throw in college vs the NFL. Unfortunately I don't know of anyone who compiles these charts outside of Rotoworld, but I feel like they'd tell us so much about how a player is used in the NFL compared to how he's used in the NCAA.
|
# ? Apr 21, 2014 17:30 |
|
Rap posted:Before you even build your model, I have a few issues to raise: That would be the time to do it, yeah
|
# ? Apr 21, 2014 17:33 |
|
You'd also have to correlate team make up in college versus success in the pros and that team make-up. and what quantifies success. and if an NFL team wants to pay me 6 figures, I will immediately begin modeling the data.
|
# ? Apr 21, 2014 17:44 |
|
As far as vague stats as an NFL indicator, I did the research once a couple of years ago and saw that college QBs with a YPA of 7.5 and above had a higher success rate than QBs with a YPA of less than 7.5. Not enough to make like a bold definite conclusion but. It was a thing. Most importantly if it was their final year and their YPA was that low. Career YPA can of course be skewed - Foles career YPA at Arizona was 7.1 but he had a YPA of 7.5 or above in his 3rd and 4th seasons. Worth mentioning: Ponder, Locker, and Gabbert all had YPA below 7 in their final college season. It's not very useful this year as most of the top QBs in this draft all rocked it out in YPA this year. Some QBs below 7.5 this year: Logan Thomas (7.2), Brett Smith (7.2), Jordan Lynch (7.2), Garrett Gilbert (7.0), Case McCoy (5.9 lol) The 7th Guest fucked around with this message at 17:58 on Apr 21, 2014 |
# ? Apr 21, 2014 17:49 |
|
Declan MacManus posted:That would be the time to do it, yeah Doesn't make sense to me to object to what's going into a formula before you know how it's used
|
# ? Apr 21, 2014 17:52 |
|
Rap posted:Doesn't make sense to me to object to what's going into a formula before you know how it's used Besides, the result should be that certain statistics have no good correlation, not the a priori assumption.
|
# ? Apr 21, 2014 18:07 |
|
OperaMouse posted:Besides, the result should be that certain statistics have no good correlation, not the a priori assumption. It isn't a priori just based on the fact that you're using statistics.
|
# ? Apr 21, 2014 18:08 |
|
I thought we definitively proved that statistics are heretical lies and the eye test is the one true God.
|
# ? Apr 21, 2014 18:12 |
|
Quest For Glory II posted:It's not very useful this year as most of the top QBs in this draft all rocked it out in YPA this year. Some QBs below 7.5 this year: Logan Thomas (7.2), Brett Smith (7.2), Jordan Lynch (7.2), Garrett Gilbert (7.0), Case McCoy (5.9 lol) If anyone drafts Garrett Gilbert they need to be able to face up to the consequences
|
# ? Apr 21, 2014 18:15 |
|
Gumbel2Gumbel posted:I thought we definitively proved that statistics are heretical lies and the eye test is the one true God. You could do a correlation between eye test and actual ability too (though I imagine for everyone but the best/worst it'd be mostly random).
|
# ? Apr 21, 2014 18:18 |
|
Gumbel2Gumbel posted:I thought we definitively proved that statistics are heretical lies and the eye test is the one true God. The eye is our God and Doltos is his one true prophet. Elihu Akbar
|
# ? Apr 21, 2014 18:27 |
|
|
# ? May 28, 2024 08:50 |
|
Gumbel2Gumbel posted:I thought we definitively proved that statistics are heretical lies and the eye test is the one true God. Football statistics are heretical. They become quasi-paganistic with the NBA and are dogmatic doctrines with baseball.
|
# ? Apr 21, 2014 18:42 |