I know this topic has been beaten to death, but most of the previous posts have been done without much reference to to the numbers. If you've read any of my posts before, you know that after I'm done writing a post on this, you will more numbers than you can shake a calculator at.
Unfortunately, after looking in depth at the stats, I'm not quite sure what to believe. I still don't know how good Cassel or Quinn are, and of course nobody knows how good Stanzi really is. So even if you know and believe all the stats, you can still reasonably support starting any of those three. But at least after reading this you should be at least a little bit more knowledgeable about the situation. Personally, I tend to support starting Quinn, with Stanzi as the #2 and Cassel as the #3, and I explain exactly why after the break.
In this post I will be using almost exclusively NY/A (net yards per attempt) and WPA (win probability added). These are somewhat advanced, especially WPA, and from experience most of the criticism in some of my previous posts have been by people who don't quite understand them. So if you don't know what these stats mean, I have included a little glossary at the end of this post explaining them.
How Good is Cassel?
These are Cassel's traditional stats. We'll ignore his first three years since he barely played. His two best years were 2008 with the Patriots and 2010 with us. The biggest differences in those years was that in 2008 he had a better Y/A number, but more sacks than he did in 2010. Some other numbers of note are his completion percentage (much better in 2008) and pass attempts (more in 2008 with the same number of games). His 2009 and 2011, as we all know, were not so good.
For this I think it is better to concentrate on his good years. The talent around him will be much closer to those years than it will be to his bad years, so I feel that his numbers will be closer to those years than it will be to the bad years. Cassel skeptics will probably scoff at that, and I think they have a point, but I feel that the difference in stats is due mainly to the talent rather than any natural fluctuation or change in ability. But with these numbers, we can already conclude that Cassel cannot do well unless there is talent around him. If we have significant injuries to offensive players again, we will be in trouble. But we can also conclude that Cassel is at least good enough to perform fairly well when he has talent around him, which he should in 2012.
The talent we have around the QB is very good, so the question becomes whether we'll see the 2008 or 2010 Cassel. While his NY/A numbers were quite similar, how he got to that point is quite different. Will he have the higher pass attempts, higher completion percentage, shorter throws and more sacks like in 2008? I believe that we'll see the 2010 Cassel, just because in 2008 he took over a pass oriented team, while in Kansas City he has a run oriented team, which lends more to the 2010 stats.
But which version would we prefer? And is either version good enough to win the Super Bowl. I think these questions are better answered by the more sophisticated stats at Advanced NFL Stats
I find these stats for Cassel fascinating. I can't believe that he had a positive WPA for 2011, and a fairly good one at that. Looking at each game individually, pretty much all of that came from the Colts and Chargers games, while the blowouts didn't hurt him that much in that stat, due to the way this stat works. Basically, when he led the comback against the Colts, our chance of winning that game when from about 8% to 100%, and since Cassel did most of the work, he gets most of that win probability. In the Chargers game, since it went to overtime, he took it from about 40% to about 80%, so he got a significant chunk of win probability from that too. In short, that was a weird season, producing weird statistical results.
But, like above, we will concentrate on 2008 and 2010, which shows which strategy we should be trying with Cassel. His 2.91 WPA in 2008 is outstanding. That would have put him in the top 10 in 2008 (although he didn't beat out anybody of note), and in 2008 that number was better than Warner, Rodgers, and Rivers. His 1.28 WPA in 2010 put him around 16th in the league that year, so about average.
Does that mean we should shoot for the 2008 version and shy away from the type of game that gave us the 2010 version? Not necessarily. If you have a QB who can consistently get positive WPA, then more pass attempts will lead to a higher WPA. So it's possible that that difference is due to him having more pass attempts in 2008. Another explanation is the quirkiness that led to his weird 2011 numbers. In 2010, most of our games were blowouts, with either us blowing the other team out or us getting blown out. This is what caused the 2011 numbers, and may lead to lower than expected 2010 numbers.
These stats seem to indicate that Cassel is slightly above average. However, in 2008 he had an team around him that had just produced the highest scoring offense in NFL history, and in 2010 he had Jamaal Charles, whose 6.4 yards per carry made him the biggest non-QB offensive threat in the NFL that year, which almost certainly helped take pressure off of Cassel. When he didn't have that kind of talent around him, he showed his true colors and struggled.
In short, I believe we can be successful and win the Super Bowl with Cassel. However, the fact that he struggled without elite talent around him indicates that we can be a truly great team if we can find a top 10 quarterback. Since we have eliminated Cassel from that conversation, that leaves Stanzi and Quinn.
Why Not Stanzi?
I'm a big Stanzi fan, and have a lot of hop that he can develop into an elite QB and be the next Brady. However, I think that it will be a couple years before he gets to that point, and even if he does turn out to be elite, he is likely not much better than Cassel right now.
I actually came to this conclusion after arguing with some people on here about whether or not elite QBs required development. I argued that most elite QBs were elite almost immediately, citing Manning's 13-3 record his second year and Brady's Super Bowl Championship in his second year. However, that debate led me to look at the data, and it turns out that I was wrong. In fact, elite QBs almost always require 3-4 years of development before they reach their peak. Whether this development occurs in games or on the bench doesn't seem to matter (at least Rodgers makes it look that way). I don't want to spend too much time proving this, but I'll just present Manning and Brady's career WPA to show that it is at least true in their cases.
(Note: Look how dominant Manning is compared to Brady. I've looked at all the elite QB's and believe me when I say they are all about like Brady, if not worse. Manning has been the best QB for the last decade and it's not even close. If he would have signed with the Chiefs, I firmly believe one of the Manning led teams would have been the best team in NFL history. It's a shame he picked the Broncos, where even he will only make them slightly above average.)
Manning's first year, 1999, is not shown, but since it was worse than his 2000, I think it's clear that he didn't get to his consistent level until his 5th year. Counting his year on the bench, Brady didn't become elite until his 4th year. Stanzi will be in his 2nd year in 2012, so unless he turns out to be one of the top 3 QBs ever like Manning, he would almost certainly still be at Cassel's level even if he's destined to become a Hall of Famer.
What Does Quinn Bring To The Table?
Quinn is interesting. Yes, he did terrible in Cleveland, but it seems that everybody does. Besides, he only started 12 games, which is not a very good sample size. Everybody has written him off as a crappy QB, but I think that conclusion is too early. I certainly haven't seen anything to suggest he'll definitely do better with more talent, but he still has potential in my mind, and potential is all we need. But let's start by looking at his traditional stats (I won't mention his advanced stats, they aren't particularly insightful. They're right where you expect them to be.)
Certainly not very good. The most concerning aspect is that the year he played the most games was his worst year. But how do we know if this is due to him being bad, or the talent around him? It's usually very hard to tell. Thankfully, since he never played a full season, we compare him apples to apples to the other QBs that started on the very same team, and see how he compares. Although we don't know how these other QBs stand up to Cassel, we can at least get an idea of how good he may be.
And once you compare him to the other QBs, he actually doesn't look too bad. He had the highest NY/A of the 4 QBs who started in 2008, most notably beating Derek Anderson's 5.1 NY/A. In 2009 Derek Anderson was the only other QB to start games for the Browns, and again Quinn's NY/A was higher than Anderson's 4.2 NY/A.
Derek Anderson may not be the greatest QB in the world, but that's ok for what we want, since we can say that Quinn is probably better, so Anderson present only a lower bound on what Quinn can do. It is also important to remember the fact that all of Quinn's starts occurred in his first 3 years, which means that he had probably not fully developed before he started his last game.
So, while it is probably true that Quinn is underrated, we still don't know much about how good he is. However, I believe it is likely that Anderson and Cassel are about the same talent wise, and if we accept that Quinn is at least as good as Anderson, I find it quite plausible that he is better than Cassel, and with our talent, he should be able to shine. If Quinn's likely floor is Cassel, and his ceiling is potentially very high, then I think he is the obvious choice to be our starter.
NY/A: Net yards per attempt is pretty simple. It is like yards per attempt, except that it includes sacks. A sack in this stat is like passing it for negative yardage. I like it better than yards per attempt because it gives you a better idea of what you're going to get when you call a pass play. I also like it over more conventional stats like total yards and touchdowns because these traditional stats largely depend on the play calling. If your coach calls lots of run plays, even the best QBs won't be able to rack up tons of yardage and points. However, NY/A only depends on how well the QB plays. It's not perfect, but it is the best traditional stat and can be used to make quick and dirty comparisons that have a good chance of being accurate.
WPA: Win probability added is a stat from Advanced NFL Stats, and is a bit more complicated than traditional stats. The idea is that at certain points in the game you have a probability of winning the game. So, before the coin toss, you have a 50% chance. If you're up by 21 in the 4th quarter, you probably have about a 95% chance of winning. Win probability added simply adds how that probability changes after the QB makes a pass (or gets sacked or fumbles). If that play increases the team's chances of winning, they get positive WPA. If it decreases their chances, they get negative WPA. 0 WPA is average.
This is a very nice stat, and it's benefits should be obvious. However, there are some downsides, which I partly explain in the post. One I don't mention is that the win probability is based off a statistical analysis they did a couple years ago, so it may not be completely accurate today. I don't think it will have much effect in terms of comparing players within a year, but it may lead to overvaluing QBs.