Quantify QB selection process.

[restless]

In light of the recent discussion on selecting future QB, I have an idea to offer. I am starting a new thread because I think this topic deserves some in-depth analysis and a position of its own. Feel free to merge it with others if you see fit.

First of all, why the need? JZ was chosen over TS for while and now I think most of us probably agree on which one is a better QB for this team. Krenzel was chosen over Scott a few years back. Craig led the team to a NC. It's hard to argue against that. But Scott did show that he has the ability to lead the team to victory in Happy Valley and might have done that in Ann Arbor. Given the importance of the position, how to select the best QB for the team got to be on coaches' top priority list. You might say that there are too many intangibles such as leadership, heart, will, etc, involved that made it mission impossible to quantify this process. I am a researcher in the field of social science and believes in quantifying intangible stuff. For the least, it won't hurt to explore. Here are my 2-cents:

step 1: identify key factors.
This can be done by surveying Subject Matter Experts (SMEs) such as coaches, QBs, other experts, and/or players. I list some that I think are important. The list include physical skills (arm strength, quickness of release, accuracy, etc), mental skills (memorizing play book, reading defense, decision making both in terms of quickness and ability to make the right decisions), and leadership.

step 2: quantify these factors.
Physical skills should be relatively easy to assess. Coaches probably already have them.
Mental skills can be relatively easily assessed as well. A fairly easy way would be simply ask coaches to offer their observations.
Leadership may be harder to obtain, but not impossible. A simple way to get it to survey the team on who they would like to be the QB.

step 3: develop these scales, collect data, determine weights on each scale, and validate these scales.
Apparently this will take some time.

To stay ahead in this competitive game, you have to be innovative. Quantifying the QB selection process should be worth a try. Thoughts?

 

[osugrad21]

I like that...but I don't agree with it.

Over the years, I've seen kids who are absolute weightroom monsters, have blazing 40 times, etc. However, they were horrible football players when the lights came on...meanwhile, the slower, weaker kids were simply "gamers."

Imho, there is no guaranteed way to measure or assess a player's ability...combines are great, 40s are nice, and the "look test" is simple. However, you never know what a guy can do until he is thrown into the fire.

The problem is that coaches who want to save their job must rely mostly on practice performances to make a judgment call. The process you describe is usually completed by a coach...obviously not as detailed as you describe, but in the same idea I guess. However, certain aspects of a great player cannot be measured or predicted.

Chris Spielman is about the best example for this discussion. Using your "formula" method, Spiels is about 3rd string at best...

 

[LordJeffBuck]

...but on the other hand, coaches always think that they can turn the raw athletes into great football players. You see it at every level, even at the NFL, where guys who were jack squat in college put on a show at the combine and get drafted in the first round - happens every year, and most of those guys don't pan out.

In addition, it is much easier to quantify "speed", "strength", "agility", and "quickness" than it is to judge "heart", "timing", "intelligence", "instincts", and "vision". Thus, coaches will often look to the "measurables" when making personnel decisions - "Hey, don't blame me - my RB has 4.3 speed, my QB can throw the ball 80 yards, my OL all bench 500 lbs.... I put together the talent, now they have to produce on the field." When in reality, the "talent" is just numbers on paper - no real football skills.

Finally, "plus" measurables often lead to ego problems with players - "I'm a world-class athlete - why should I lift/practice/read the playbook/study film? I can get by on my natural ability." The "million dollar body, ten cent brain" scenario....

IMO, Coach Tressel does a great job of identifying athletes with an excellent combination of "measurables" and "skills", with the right "attitude" to succeed.

 

[restless]

I like that...but I don't agree with it.
Imho, there is no guaranteed way to measure or assess a player's ability...combines are great, 40s are nice, and the "look test" is simple. However, you never know what a guy can do until he is thrown into the fire.

The problem is that coaches who want to save their job must rely mostly on practice performances to make a judgment call. The process you describe is usually completed by a coach...obviously not as detailed as you describe, but in the same idea I guess. However, certain aspects of a great player cannot be measured or predicted.

Chris Spielman is about the best example for this discussion. Using your "formula" method, Spiels is about 3rd string at best...

Thanks for your response.

The foremost thing in predicting is identify the most influential predictors. Your point is that 40 times or performance in practice are not good predictors. This could be true. But there got to be some factors that are more important than others. The question is how to identify them. This is why it should be evidence-based and why the validation process is so key.

Let me try to spell it out to see if it makes any sense. Assuming we identify 15 factors and collect data on them. Then we will have a chance to see which ones perform better. Maybe none will. Then that means we are looking at the wrong place. The other scenario, which I think is more likely to happen, is that after a few rounds of research, we could identify a set of predictors that are consistent and effective.

Using numbers alone to say who is better may be arbitrary. But it could be effective tools if appropriately understood and applied. It will also make the decision process more transparent, which I think should help reduce biases.

 

[restless]

I have a world of respect for JT, who is actually responsible of turning me into a football fan. Don't blame me. I came from another country and never watched football until a few years ago.

I certainly agree that he is great in identifying talent and put them in a place to succeed. But to say that he doesn't have biases like the rest of us is at least a stretch. That's why I think such a selection system may help. A lot of research evidence in medical decision making and decision making in general has shown that this type of system almost always outperform clinical judgment.

 

[osugrad21]

Thanks for your response.

The foremost thing in predicting is identify the most influential predictors. Your point is that 40 times or performance in practice are not good predictors. This could be true. But there got to be some factors that are more important than others. The question is how to identify them. This is why it should be evidence-based and why the validation process is so key.

Let me try to spell it out to see if it makes any sense. Assuming we identify 15 factors and collect data on them. Then we will have a chance to see which ones perform better. Maybe none will. Then that means we are looking at the wrong place. The other scenario, which I think is more likely to happen, is that after a few rounds of research, we could identify a set of predictors that are consistent and effective.

Using numbers alone to say who is better may be arbitrary. But it could be effective tools if appropriately understood and applied. It will also make the decision process more transparent, which I think should help reduce biases.

I can respect that approach, but again, not too much faith in it. However, the way the business is constantly morphing, it is interesting to explore.

To me, the only real test lies when its one-on-one in limited space...heart cannot be measured nor coached.

However, I am definitely interested in any attempt to do so...

 

[restless]

However, I am definitely interested in any attempt to do so...

Agree. This is definitely an empirical question. Untill we have data and evidence, it's anyone's guess.

My observation, no matter how raw and limited it might be, was that Troy always had a better understanding of the game than Justin. Especially in 04, Justin seemed lack the ability to keep up with the game. He forced too many throws. Santonio and Tony had to make a dive to catch or catch it in traffic. Troy certainly had his dues. He probably tucked and ran more than he should at the beginning. But I think forcing the ball is much more dangerous. JZ's int number proves it. Throwing in the team's trust in Troy. I think JT's choice of JZ was biased, unintentionally would be my guess.

Now talking about Craig Krenzel, he was not the best passer in the world. But his understanding of the game was superb. He knows when to fight and when to save for next try.

So if I had to guess, I think decision-making and team's trust are more important than passing percentage in practice. A simple regression would confirm or disconfirm my theory. If anyone could get any data.