Is N. Suh Worth the Money? | An NFL Analytics Perspective

No its not.

First of all, a player like Suh has astronomically better odds of paying off then winning the lottery.

Secondly, you aren't arguing against the Dolphins, you are arguing against the system that allows the Dolphins to sign Suh for big money.

Lastly, paying a FA big is no different than paying a home grown talent big. The only thing that matters is that you are paying big for players that are worth it from a talent and effective position level. A Suh level DT is the second most important player on your team after QB and the most important player on defense.

 

Wait a second.. I correctly pointed out in that thread that your article there has a major confound: because weak teams pick higher in the draft, they will naturally (through the draft) have more unequal distribution of salaries. Since that article never separated the effect of the draft out of the analysis, you can't use their conclusion to inform us of whether unequal distribution of salaries is better or worse when we're only talking about free agency.

 

If we trade down in the draft to acquire more picks it is playing the lottery. We have an idea what we are getting (winning ticket) but really have no idea until the guy hits the field.

If we are forced to move back then we are picking from less talented people. I am not saying Suh can not make up some of the difference in his play. What happens though when he goes down to injury or a dumb bonehead play. Is it wise to put all our eggs in one basket?? At least with better overall talent we would have a chance at plugging and playing. The guy playing beneath Suh is going to be inexperienced and lower paid then normal. This usually indicates more often then not lower skill.

I am still out on weather this is a wise move or not. Given the status of all the other holes we have to fill and now at lower cost, the overall reassembly is going to be harder, because the replacements will tend to be lower paid less experienced guys. We might be ok. I hope we are.

Seems like this type of decision indicates to me that the coaching staff must think we are very close to competing for a superbowl

 

Check post #28. You basically proved my point with that equation. The whole result is dependent on draft status.

 

And again, this argument you're making isn't about Suh really. Its about the whole system that has uncapped salary levels in a capped salary pool. Plenty of winning teams pay big for players on Suh's level. The system works, the only problem is spending that much on a bad (or on only a "good" player) or a great player at a unimportant position. Neither of those apply to Suh.

 

I am making the assertion that we are going to fill other areas of the team with players of less talent because of money we are paying or going to pay to players like Suh, Tannehill, Wallace, and Clay to name a few. Areas that will not be directly impacted by Suh personally.

I do not dispute the talent of these players. What I am eluding to is we are becoming the NEW YORK JETS and we will have the same salary cap problems for years they had if we are not careful and methodical. Additionally, this type of signing opens us up to liabilities we did not have before such as quality back ups. If we are signing lower paid less talented players because we have 50 million dollars tied up in 4 or 5 players then this team just became a whole lot more vulnerable to injuries and the impact they will have.

Just a thought to demonstrate my point - If Moore is gone now as a result of we can not afford his 5 mill salary and Tannehill goes down ........

 

Its a poor argument because again, its the system not the team. The system clearly works as every year, a team wins. I mean at one time Brady was the highest paid player in the league. His pay numbers were high then and are considered pedestrian now proving the benchmark moves every year.

I mean, the fact that Moore is paid so highly for a back up is counter to your Tannehill point at the end.

This argument is made every year someone is signed for a lot of money. Hell, even on this team, we were told the exact same thing about signing Wallace, yet we went out and signed Albert and heard the same thing again, but now we're on the cusp of signing Suh.Look at the other end, you have Oakland who sucks and they have spent so little money they HAVE to spend more per league rules.

The cap is fluid and raises every year. Contracts can be reworked or dropped. Issues with the cap can be solved. Its not a death sentence. What doesn't happen hardly ever is a player of Suh's ability in his prime become available at a need. You don't pass that up.

 

From the article:

"DRAFT[SUB]i[/SUB] includes the dummy terms Undrafted (1 if the player was not drafted; 0 otherwise) and Round 1 (1 if the player was drafted in the first round; 0 otherwise)."

So, they're only comparing round 1 vs undrafted, as one can see in their Table 1. In other words, they are treating round 1 and undrafted the same (in terms of weight). That's what "correcting for" in their paper means. So yes there is the confound I mentioned with draft status.

 

This is an excellent point. We should have locked up tannehill prior

 

My point was twofold:

1) Draft order is not accounted for. The difference in salary between picking players near the top of any round (say the 1st) vs. near the bottom is huge, and was especially so during most of the time when they did their analysis, which was before the current agreement on rookie contracts.

2) Their analysis doesn't condition on the player being a free agent. So, we don't know how much of the effect is due to salary inequalities before FA vs. after FA.

 

*3. This was an epic misstep IMHO. They had plenty of time to sign Tannehill before this. They are absolutely going to pay more for his services because of this.

 

Shouldn't have to presume with statistical analysis.

 

Why would Tannehill sign?

 

I doubt that. It's actually easy to test (just takes time to get the data). But my intuition says you won't get a uniform distribution for most teams' draft position over those 11 years.

 

Yeah, I should've said that too. Good response.

 

Oh btw.. it's not enough that they are randomly distributed over time (uniform distribution over that time period). It's necessary there's almost no correlation from year to year. Otherwise, the effect of inequality of salaries due to draft order could "cancel itself out" (extreme case: first 5 years team always picks first, next 5 team always picks last). So that makes this even less likely to be true.

 

We already know that the effect of salaries due to draft status does NOT cancel itself out. They demonstrated that in the article. But there are ways where the effect of unequal salary distribution due to draft status won't cancel itself out AND you have a high correlation from year to year (toy example: win 1 game first year, 2 the next, 3 the next etc..). Point is, you need to show both uniform distribution and there not being high correlation from year to year.

As far as whether uniform distribution and/or year-to-year correlation are likely to be true or not, just eyeballing you can see both probably don't hold. Look at the records of different teams the last 15 years let's say (you can get that easily on wikipedia.. for example: http://en.wikipedia.org/wiki/List_of_Miami_Dolphins_seasons ) and you can see it's not a uniform distribution for # of wins/losses and there seems to be good correlation from year to year. That doesn't prove it's not true for draft status because of trades, but it really suggests it.

Oh, and some teams you don't even need to look at the records of. Try the Patriots. They will easily not have a uniform distribution for draft position.

 

Just not sure if turning over players every yr or two will benefit this team, now or at anytime in the future. It sure does seem like we are headed down the same path that Tannenbaum created in NY. I hope I am wrong but things are starting to come into focus a little more each day. As I have stated in other posts I do not dispute Suh is one of, if not the best, DT in football. I am very happy to have him - just not at $19,000,000 a year. It seems as if there are two many downsides. Not from his play or necessarily his contract alone, but from all the contracts together which in my opinion will have a profound impact on the team for several years to come. Inconsistency and turn over will kill a team because there is no chemistry. Sometimes that can be overcome by great coaches. If history has taught us anything the total flop of the team 2 yrs in a row down the stretch with a .500 avg season record would seem to indicate otherwise.

If Suh was the last piece we needed to put us over the top, ok we should try to get to the super-bowl by bringing him in.

I just hope this is not Philbin's ----wave of the flag so to speak saying we are all in. Super-bowl or bust

 

lol.. is this going to require one of us to do some work? Tell you what.. later when I have time (probably this evening), I'll plug in some data and give you the results. But like I said, just eyeballing the records suggests you're not going to get uniform distributions.

 

Tannehill's asking price just went up. If Suh has now set a new precedent, I would expect Tannhill to be in the vicinity of Suh. Since typically teams live and die by QB play. Got a feeling he also will want a good bit guaranteed. Remains to be seen.

http://www.cbssports.com/nfl/writer...uhs-deal-with-dolphins-could-impact-tannehill

It's pretty much impossible to have two players on big-time quarterback deals, and make no mistake, that's what Suh is on. Personally, I would have waited on Ryan Tannehill all of this year no matter what, picked up the fifth-year option and let this thing play out. I want to see more empirical evidence. Mike Tannenbaum officially took over the front office a few months back and might as well take some time.

The problem becomes, what if Tannenhill does have a big year? How much wiggle room will Miami have? Of course, given the state of this team's quarterback position since Dan Marino retired, it's a good problem to have, but there will be constraints from Suh's deal. That's just how it works. I'm not sure Tannehill is going to be worth more than what Alex Smith just got, but if he does surpass that type of production in 2015, the Dolphins could feel a little squeeze. Maybe even a big one.

 

Salary caps aren't static. Mark my words, we're going to see 15-20 million dollar jumps per year for the next couple of years to compensate for all the upper mid level guys signing 100 million dollar deals. Suh isn't one of those guys, so we made out better than the Rams and 49ers.

 

Seems like part of the argument against a deal like this is confidence in a team's ability to find quality players in the draft and through trades. Sure, you may hit on a draft pick here and there. But go check Grantland's piece on the brilliance of Bill Belichick from a few months ago. Even he misses quite a bit in the draft. It's probably easier to find a good player through trades since you have data on the player. But that requires the resources that you need to try and get young, cheap players; draft picks.

So yeah, paying a player a lot in free agency is risky because of the cap. But unlike a trade, you don't have to give up draft picks or other players. And unlike the draft, you have more evidence as to how a player will perform with your team. There's risk everywhere. It's just not as readily apparent with draft picks because it takes longer to get the payout, if there even is any.

 

Assuming the cap is also going to rise is no way to manage a team either. If revenue does not rise neither does the cap. It could even go down which is unlikely.

 

OK.. I finally plugged in some data and did a preliminary analysis. First of all, let me say this entire analysis is done on 27 teams, not all 32, because some teams didn't have a history going back to 1994, and that study looked at 1994-2004 drafts. I don't know if I missed one or the other team but 27 out of 32 should be more than enough to see what the distributions are like.

Here's the deal. The analysis was more difficult than expected because of small sample sizes, meaning you can't really conclude anything with the sample sizes we're talking about. From 1994-2004, you have teams like the Chargers that had only 5 1st round picks (only one 1st rounder from 1994-2000) to teams like the Jets that had 14 (Jets had 4 1st round picks in 2000!!).

Well, for the purpose of statistical analysis, 5-14 is generally way too low to claim the data did NOT come from almost any distribution. That is, statistically speaking, you can claim there's no evidence it didn't come from a uniform distribution, or didn't come from a normal distribution, or didn't come from practically any other distribution. In other words, stats are useless if we just use the raw data in this case.

What we need to do is to increase the sample size (artificially) until it is past a threshold that allows us to distinguish among different distributions. Why artificially increase it instead of using a larger range? Because the number of 1st rounders varies so much by team that to get similar sample sizes, you'd have different ranges. I also wanted to use just the data the paper used. Anyway, by simulation, I found that 20 samples is at the extreme low end of acceptability there, so we'll go with 20 (don't want to artificially increase the # of samples too much because that'll start biasing the results).

So, how do you artificially increase the samples to 20? By bootstrapping, which is a technique that makes the fewest assumptions about the distribution the sample came from. Bootstrapping just means to assume the sample is the best estimate of your distribution, and you just sample from that (so you sample from the sample!). I did that for all teams, 20 times for each.

Now that you have sufficient samples, the question is what test to use. For small sample sizes, something called the Kolgomorov-Smirnov test is best (also makes no assumptions about the distribution). After applying that, I found that 6 of the 27 teams had distributions of 1st round draft picks from 1994-2004 where you cannot reject it coming from a uniform distribution. However, for 21 teams you can reject the hypothesis.

Which 6 teams were these? Vikings, Giants, Lions, Raiders, Seahawks, Rams. All others had distributions of 1st round picks from 1994-2004 that statistically speaking did not come from a uniform distribution.

 

Analytics is great but you can learn a lot by just watching what the guy does. Put on the tape, call up the NFL rewind or whatever. Watch the guy. He's a D-line monster. If it were my money, I'd have thrown that much at him too. If he plays for us like he played for the Lions, we got a good deal.

 

Nice post on the topic: http://www.advancedfootballanalytic...yroll-personnel/237-analyzing-the-suh-signing

 

when you sign a player like Suh that comes with that salary, its a given that you better have confidence in your scouting staff to fill holes in the draft and low level free agents who have upside.

 

lets just call that the Kong effect...Pouncey now knows he better get as strong and quick as possible or else he's gonna get run and embarrassed..this team improvement comes from just the simple presence of Kong.

 

With Suh, our OL will have the benefit of practicing against one of the best DTs in the league, this will definitely help them to get better and increase their toughness and aggressiveness, Lord knows this team needs it.

If Dallas Thomas is still a starter on the OL, though, Suh might need to take a few plays off during practice, because the offense won't be able to run anything with him there...