Ultimate Boredom surfing - comparison of Reid and BB draft success

[Armchair Quarterback]

I am not arguing about what is a hit and what isn't. The point is the article chose their definition and applied it to all the players. It is really hard to come up with an objective evaluation of a large number of football players (much easier with baseball and basketball) and Pro Bowl appearances are at least an objective fact.

Doesn't mean their definition of what a hit is is worth anything.

 

[plk]

"BB was just as skillful but had bad luck."

Just to be clear I am not saying that. I don't know of any way to distinguish between "less skill" and "bad luck" without a very large sample size and a professional statistician.

I am saying that just because of the way probability works you will inevitably get bad draft years and good draft years even if the skill level is always the same. I believe the observation puts the situation in perspective. The draft is a crapshoot, and results are what you expect in a crapshoot.

 

[plk]

Doesn't mean their definition of what a hit is is worth anything.

I would only argue that the analysis in the article is better than endlessly arguing without any objective standards. I suspect that inaccuracies in evaluation even out over the ten-year period.

It is an attempt to assess the team's drafting, not individual players. I would be all in favor of more exhaustive evaluations of drafting.

 

[Deus Irae]

Look Deus, most of the time you're a great poster. Sometimes though, you can be unbelievably dense.



Come on man, "That's exactly how probability works"? Really? Probability means that there's the same outcome every time? That's really interesting, I think they might need to rewrite every statistics book ever written.

That's not the claim I made. Go back and re-read it. In fact, here's the paragraph:

If BB drafted with exactly the same level of skill every year, the results would be essentially the same every year, because the 'skill' variance would be 0, which means that only the probabilities, that you earlier tried to argue were pretty much everything, and the overall talent depth in each draft would vary, and they'd be within obvious ranges.

In a probability-based situation, you'd statistically gain a certain percentage of hits/misses each year, within the variables of number of picks and round chosen. It wouldn't be an exactly even number each time, and it would have the possibility of statistical outliers, but the general trends surrounding 7 draft picks would remain and it would always remain the same, and random. This would essentially mean that PLK's own initial argument is absolute garbage, because he's trying to argue

A.) It's all about probabilities

B.) BB is one of the best in the business

It can't be both A and B. Either it's not all about probabilities, or one GM can be better than the other. That's an either/or situation. Since none of us believe that to everyone drafts equally well, PLK can't possibly mean what he was posting about it being all about probabilities. It can't be a purely random statistical probability and, therefore, he has no real point.

You're probably not wrong about BB not being as skillful as usual. The problem is posters like you and plk always want to see things black and white. You're both right. It doesn't have to be either "BB was less skillful during that period, regardless of luck" or "BB was just as skillful but had bad luck." It's probably both. He probably didn't make decisions quite as well as he usually does, and he probably had some bad luck.

Again, you seem to have misread my posts. BB and the Patriots, overall, draft about as well as any team in the NFL. They also had poor seasons from about 2006-2008 or 2009, Depending upon how 2009 finishes shaking out, and some might argue that 2000 was terrible at the top, as well (I'd say that rounds 6 and 7 made that year's draft an overwhelming success, but that's my personal take). Looking at it that way is not seeing it as just black and white, at all. As for the "skill" aspect, again, War Room gives us some insight on what happened, where BB was overruling his scouting staff.

 

[goheels22002]

Even for a Bleacher Report article, that's a lousy piece of work. The author should be ashamed of having written that.

For Bleacher report, which reminds me of Us or People, it wasn't bad. The guys at Cold Hard Football Facts, who do a much better analysis, agree, in essence, that the Eagles and Pats are about the same over the past decade.

The statistic that I thought was interesting last year during the playoffs was the fact the four semifinalists (NE, BAL, SF and NYG) were all built about the same in terms of where their players came from - roughly 50% were their own draft choices and 50% some other way. It was interesting to see some balance among 1)guys drafted who played out a contract and entered free agency in search of a contract or a shot at winning, 2) undrafted free agents, and 3) trades. Pittsburgh and Philadelphia looked about the same.

The perennial crappy teams were unbalanced - hardly any of their own draft choices and loads of free agents and scrap heap guys from salary dump trades.

It said to me that the trick is to get players who stick in the first two rounds and hope that 30% of your other picks work out.

I also think that branding any player a bust after one or two seasons is a mistake. Teams with quality depth can bring guys along and find where they fit and what they do best. What a guy does in college does not indicate what they'll do well in the pros.

Last, injuries throw off the kind of quick analysis the Bleacher Report guys did. Unless you take a guy with a serious injury history (Ras I Dowling, anyone?) it is impossible to brand a player as a bad pick if they get hurt early in their career. RB Robert Edwards is the saddest story I recall like this. Edwards was incredible and a tremendous selction - the knee injury at the rookie pro-bowl game on the beach was a freak accident. It did not make him a bust or a bad pick. Just sad.

 

[Avenger]

For Bleacher report, which reminds me of Us or People, it wasn't bad. The guys at Cold Hard Football Facts, who do a much better analysis, agree, in essence, that the Eagles and Pats are about the same over the past decade.

The statistic that I thought was interesting last year during the playoffs was the fact the four semifinalists (NE, BAL, SF and NYG) were all built about the same in terms of where their players came from - roughly 50% were their own draft choices and 50% some other way. It was interesting to see some balance among 1)guys drafted who played out a contract and entered free agency in search of a contract or a shot at winning, 2) undrafted free agents, and 3) trades. Pittsburgh and Philadelphia looked about the same.

The perennial crappy teams were unbalanced - hardly any of their own draft choices and loads of free agents and scrap heap guys from salary dump trades.

It said to me that the trick is to get players who stick in the first two rounds and hope that 30% of your other picks work out.

I also think that branding any player a bust after one or two seasons is a mistake. Teams with quality depth can bring guys along and find where they fit and what they do best. What a guy does in college does not indicate what they'll do well in the pros.

Last, injuries throw off the kind of quick analysis the Bleacher Report guys did. Unless you take a guy with a serious injury history (Ras I Dowling, anyone?) it is impossible to brand a player as a bad pick if they get hurt early in their career. RB Robert Edwards is the saddest story I recall like this. Edwards was incredible and a tremendous selction - the knee injury at the rookie pro-bowl game on the beach was a freak accident. It did not make him a bust or a bad pick. Just sad.

Ras-I only had an injury senior year, if not for that he would be a top 10 pick, I think he make a good pick taking him, his injury last year was unrelated to the one he had in college, just bad timing having them back to back years, he will be better this year.

 

[fester]

Fair enough. Good retort. I don't agree with everything but at the end of the day, I still prefer a targeted FA class and /or trades like 07.

The Draft is a crap shoot and top to bottom 1-7 only a 17% chance of a top impact player or long term starter in a period of three years as a judging point.

My point is that he is turning a 51% (using their figures) graded success of that first round pick to a massive experiment of a plethora of lower round players that have a 13% success rate as what this article called "a hit".
DW Toys

Where is your 17% coming from?

Looking at the 2009 draft and using criteria you have established (long term starter --- defined as starting at least 12 games a season by year 3, and excluding players who started a lot as a rookie but then did not start frequently in Year 2 and Year 3), we would expect to see the draft produce a total of 46 +/- "impact players or long term starters" by 2011. Going through the 2009 draft, I got to 46 players meeting this definition by pick 109. Of that 46, 6 made the Pro Bowl and another player (Vollmer) was 2nd Team All-Pro at some point in their career.

Furthermore, going through the data, the 1st round had 21 of 32 players meet criteria for a hit rate of 65%. Slicing and dicing the 1st round data by draft position does not show any meaningful hit rate differences. The 2nd round had 16 of 32 meet criteria for a hit rate of 50%. Slicing and dicing the 2nd round does show a difference -- picks 33 through 42 had a 60% hit rate, but 43 through 64 only had a 45% hit rate, so the talent cliff looks like it started in the mid-40s.

Going onto a slightly different point now with this data, if the hit rate from #1 to #42 is clustered in the mid-60%s pretty much no matter where one is picking, albeit a super-hit rate (Pro-Bowlers/All-Pros) is concentrated in the top-15 of the draft, a strategy of getting as many opportunities to pick in this range makes a lot of sense for the 2009 draft.

And if the cost of getting into the Top-15 is too high (for whatever value of too high you choose), going from the mid-20s to multiple picks in the 30s and 40s again makes sense. Now the problem is the Patriots had the ammo in the sweet spot of the draft and got one definite bust (Butler) and one probable bust (Brace) while only getting one solid player (Chung) from the sweet spot. They made up for Butler/Brace with a home run in Vollmer, but the scouting/evaluation process failed the Patriots despite having good opportunities in the sweet spot of the draft.


https://docs.google.com/spreadsheet/ccc?key=0AkshrhgU2x-pdHJuaWxLMWF6V2ZGVVBtSTgtMjdqZ2c

 

[plk]

A.) It's all about probabilities

B.) BB is one of the best in the business

It can't be both A and B. Either it's not all about probabilities, or one GM can be better than the other. That's an either/or situation. Since none of us believe that to everyone drafts equally well, PLK can't possibly mean what he was posting about it being all about probabilities. It can't be a purely random statistical probability and, therefore, he has no real point.

It's all about getting the probabilities right. It's about correctly selecting the players with the highest probability of turning into the player that you want. Belichick is as good as anyone at that.

There is really only one high probability of success round, the first, and even that is not as high as many people imagine. Maybe 50% success in the second round, if you are good, and it's all downhill after that.

If Belichick selects a player who has a 50% chance of success in the second round, and other general managers are selecting players who have a 40% chance of success, over a period of time Belichick will do better than they do.

But short-term, Belichick's failure rate will be 50%, and he is very likely to have a string of failures at some point. And he will have some strings of successes. This will happen even if he is doing exactly the same thing with exactly the same skill.

The same applies to the 40% chance of success people. They have a 60% chance of failure, but they may well still have some runs of successes.

In this example, Belichick will have an advantage over time, but the short-term results may not show it. Short-term, getting lucky can easily trump skill.

It's the same thing in flipping a coin where the chances are 50-50. Look at a table of a large number of coin flips. You will be surprised by how many long runs there are of heads and tails. A man who teaches statistics says that he can always tell when his students have faked their coin flipping exercise because they don't include enough seemingly improbable runs.

Since I've said all this above, and "God" did not understand it then, I don't expect him to understand it now, but I know that there are people out there who will understand that if they read it.

Let me try it this way since the draft is frequently described as a crapshoot. If you are shooting craps, the results are going to be random, but if you have a more accurate knowledge of the odds, and thus what is a good bet, over a period of time you're going to do better than someone who has a less accurate knowledge of the odds.

 

[NSPF]

A.) It's all about probabilities

B.) BB is one of the best in the business

It can't be both A and B. Either it's not all about probabilities, or one GM can be better than the other. That's an either/or situation. Since none of us believe that to everyone drafts equally well, PLK can't possibly mean what he was posting about it being all about probabilities. It can't be a purely random statistical probability and, therefore, he has no real point.

He didn't mean that it was pure noise, because that would be ridiculous. He meant that every single draft pick has a chance of failure, and that it is impossible to eliminate this fact. This means that you can do everything right and still end up with a poor player, without it necessarily having been a poor decision, given the information you had at the time.

Over time, you will generally see regression towards the mean. Over 7 picks, our theoretical GM who does everything right will do better than most teams, most of the time. Over 12 complete drafts, he would be far ahead of the pack almost always. However, there's still a chance that despite his flawless methods, multiple players will bust, and the any given draft will be a "failure." This still doesn't necessarily mean that the GM was a worse GM that particular season, or that he did anything differently.

This was plk's point, I believe. Results can vary more widely than one might expect despite no change in methodology. It was relevant to the discussion because it was one reasonable explanation to why BB had a run of poor draft results, which is what was being discussed.

I think that you refuted this by saying, no, because BB is not a theoretical GM, it isn't reasonable to assume that he drafts with the exact same skill every year. In fact, in 2006-2008, he drafted less skillfully than usual, as evidenced by his actions in War Room, as well as a unusually poor string of results, which seem beyond what can be attributed to probability.

Yours is a better explanation. However, it doesn't seem unreasonable to me to say that we also had worse luck than usual. I think everyone expected players like Maroney and Meriwether to work out better than they did.

Again, you seem to have misread my posts. BB and the Patriots, overall, draft about as well as any team in the NFL. They also had poor seasons from about 2006-2008 or 2009, Depending upon how 2009 finishes shaking out, and some might argue that 2000 was terrible at the top, as well (I'd say that rounds 6 and 7 made that year's draft an overwhelming success, but that's my personal take). Looking at it that way is not seeing it as just black and white, at all. As for the "skill" aspect, again, War Room gives us some insight on what happened, where BB was overruling his scouting staff.

What part of your post did I misread, and what did it actually say? Everybody in this thread acknowledged that 2006-2008 were poor drafts, nobody disagreed. We are discussing why they were poor drafts.

 

[plk]

This was plk's point, I believe. Results can vary more widely than one might expect despite no change in methodology. It was relevant to the discussion because it was one reasonable explanation to why BB had a run of poor draft results, which is what was being discussed.

Yes. Exactly. Thank you.

I think that you refuted this by saying, no, because BB is not a theoretical GM, it isn't reasonable to assume that he drafts with the exact same skill every year.

I don't assume that he drafts with the same skill every year. My point is that he would still have bad years and good years even if he did.

In fact, in 2006-2008, he drafted less skillfully than usual, as evidenced by his actions in War Room,

My reading of that is that Belichick went with what he regarded as the best information, i.e. the coaches who knew the players. It is his job to override his scouts when he disagrees with them.

as well as a unusually poor string of results, which seem beyond what can be attributed to probability.

It may be. I don't know. If you want an easy way to flip a coin go here, Coin toss probability , and click on flip coin . Do it a number of times. Remember teams start with 10 second round picks in 10 years. I think you will be surprised at the results. I always find it counterintuitive, and I know what is going to happen.

Yours is a better explanation. However, it doesn't seem unreasonable to me to say that we also had worse luck than usual. I think everyone expected players like Maroney and Meriwether to work out better than they did.

I assure you I am not claiming that the less successful drafts were necessarily all a matter of bad luck. I do want to put Belichick's performance in the context of what is possible in the real world by comparing him to the best other general managers rather than some ideal standard that no one meets. That was how this started. I think it is important to understand that runs of failure and of success are inevitable when you are dealing with probabilities.

 

[PYPER]

I hated Maroney so I can't call him a hit, but he did start for, what, four years. It's hard to call that a miss either. He certainly missed on our expectations but his status as a regular starter for us means he can't be called a miss either.

How can Dowling and Vereen be labeled as misses? It's far too early to include 2011 picks.

 

[Armchair Quarterback]

All this probability talk is nonsense. Its about talent evaluation and how well you project that players skill, desire and ability to fit your team.

 

[plk]

I hated Maroney so I can't call him a hit, but he did start for, what, four years. It's hard to call that a miss either. He certainly missed on our expectations but his status as a regular starter for us means he can't be called a miss either.

Yes. How you feel about the drafts in 06 and 07 is largely going to depend on how you see a draft pick that initially seemed good but then went the other way. The human mind is a mysterious thing.

They can be seen as good picks that just did not work out or as bad picks.

That is not an argument that I want to get into.

 

[plk]

Its about talent evaluation and how well you project that players skill, desire and ability to fit your team.

Exactly. I have never said anything else and I agree with you completely.

From among the players available, you draft the ones with the highest probability, based upon talent evaluation etc., of being successful with your team. The point is that all draft picks are gambles, and the trick is to choose picks where the odds are as good as possible. But outside of the first-round, the odds are at best even and usually very long. A lot of failure is inevitable.

 

[BradyBranch39]

I refuse to read BR, so solely based on the first page of replies:

LOoooooolllllll @ BR once again!

 

[AzPatsFan]

The purpose of the draft picks is to stock a football team. Therefor the ultimate measure of draft success is ultimately, team success. By that criterion, 5 Superbowl appearances in 11 years answers who has done the best job restocking a team.

End of Story.