Draftmetrics

[adam]

So I was reading some articles on NFP, that linked me to Draftmetrics, here are the articles: http://www.nationalfootballpost.com/Late-r...to-success.html and http://www.nationalfootballpost.com/Studyi...-NFL-Teams.html

 

Some great info and analysis, checkout their site here: http://www.draftmetrics.com/

 

Some of the take-aways, based on Big school vs Little school drafting, the Bears are almost dead on league average, where teams like the Seahawks and Saints almost exclusively draft from a BCS eligible school.

 

The Bears drafted over the league average from the Big-12, and less than the league average from the ACC.

 

For the entire league, here is the percentage of 5-year starters from their draft position (1993-2006):

Picks 1-13 - 74.7%

Picks 14-40 - 55.3%

Picks 41-66 - 39.8%

Picks 67-86 - 22.9%

Picks 87-149 - 16.0%

Picks 150-189 - 8.9%

Picks 190+ - 6.8%

 

The drop-offs are to be expected, though I didn't necessarily think they would be at those picks. Once you are past pick 66, you have less than a 25% chance of drafting a regular starter. It drops off quickly after that.

 

From 1993-2006, the Packers, Steelers, Rams had the most total 5-yr starters, drafting 31 or more. The Rams were surprising to me there.

 

The Bears are actually in the 2nd group with 28. However, when you get to the efficiency part of the draft, that is where we fail. For efficiency, we were 17th at -.033, meaning we drafted less starters than what was the norm for our pick positions. The Packers, Steelers, 49ers, Giants, and Rams were the most efficient. When I think about this, the Rams have had a lot of top 5 picks, so I can see where they were almost in the can't miss category boosting their efficiency. The Lions, Eagles, Raiders, Saints, and Vikings were the worst. Now this makes sense too, Matt Millen and Al Davis selecting bust after bust, year after year.

 

To me there was not much groundbreaking info there, but the success rates of picks was pretty interesting. For us, since we are not in the top 13, it almost makes sense to move back and pick up another pick to give us a better chance at drafting a starter.

 

 

[scs787]

Not sure how to respond to this but I agree on the last stanza, we should still look to move down to pick up more picks.

[adam]

Not sure how to respond to this but I agree on the last stanza, we should still look to move down to pick up more picks.

Yeah, I was just pointing out another analysis site and some of the stuff they posted. I had never heard of the site before, but it backs up and refutes some of the info that others have posted lately.

[AZ54]

Be a little careful with their groupings of the picks (i.e. picks 14-40) as even they admit it was somewhat arbitrary where they drew the lines. I do like their Pro Bowl analysis as it clearly showed your safest top 10 picks is LT. In fact that analysis yields more insight into who/what to draft than simply looking at 5 yr starters which, if they are on the Cardinals Oline, doesn't say much about the quality of the pick. Any team picking in the top 10 that does not get a Pro Bowl player missed IMO.

[jason]

Good thread. Wonder where the replies are?

[adam]

I was thinking about the trade down scenario and how that would apply to the starter percentages. So if the #20 pick gives us a 55.3% chance at landing a starter, trading down with the 49ers for their 31st, 74th, and 131st pick would give us these percentages: (55.3% + 22.9% + 16.0% = 94.2%). I know it would not exactly play out like this, but I really like those odds better than the one shot one kill scenario at 20.

 

I was also thinking that if we could trade down in the first with someone like the 49ers, then we could do it again and move back into the early 2nd. Moving down twice, once to the end of the 1st, then to the early 2nd, could gain us a 3rd, an extra 4th, 5th, and 6th rounder for moving into the high 30's from 20. To me this is decent, 4 picks for 15-18 slots, but it would not give us an extra 2nd (other than the one we get for losing our first).

 

I would love a scenario where we could pick 3 times before the start of the 3rd round, even if all three were in the 2nd.

[AZ54]

I was thinking about the trade down scenario and how that would apply to the starter percentages. So if the #20 pick gives us a 55.3% chance at landing a starter, trading down with the 49ers for their 31st, 74th, and 131st pick would give us these percentages: (55.3% + 22.9% + 16.0% = 94.2%). I know it would not exactly play out like this, but I really like those odds better than the one shot one kill scenario at 20.

 

I was also thinking that if we could trade down in the first with someone like the 49ers, then we could do it again and move back into the early 2nd. Moving down twice, once to the end of the 1st, then to the early 2nd, could gain us a 3rd, an extra 4th, 5th, and 6th rounder for moving into the high 30's from 20. To me this is decent, 4 picks for 15-18 slots, but it would not give us an extra 2nd (other than the one we get for losing our first).

 

I would love a scenario where we could pick 3 times before the start of the 3rd round, even if all three were in the 2nd.

 

All I can say to that is: Jacksonville has 22 starters on their roster.

[adam]

All I can say to that is: Jacksonville has 22 starters on their roster.

Sure, but they were talking about 5-yr starters, which narrows it down considerably, and yes every team has 22 starters, but for the most part someone who sticks around 5 years is more than likely going to be better than the guys stuck on the bench behind an All-Pro.

[jason]

I was thinking about the trade down scenario and how that would apply to the starter percentages. So if the #20 pick gives us a 55.3% chance at landing a starter, trading down with the 49ers for their 31st, 74th, and 131st pick would give us these percentages: (55.3% + 22.9% + 16.0% = 94.2%). I know it would not exactly play out like this, but I really like those odds better than the one shot one kill scenario at 20.

 

I was also thinking that if we could trade down in the first with someone like the 49ers, then we could do it again and move back into the early 2nd. Moving down twice, once to the end of the 1st, then to the early 2nd, could gain us a 3rd, an extra 4th, 5th, and 6th rounder for moving into the high 30's from 20. To me this is decent, 4 picks for 15-18 slots, but it would not give us an extra 2nd (other than the one we get for losing our first).

 

I would love a scenario where we could pick 3 times before the start of the 3rd round, even if all three were in the 2nd.

 

You're right about your follow-up sentence. That's not how statistics work. Each roll of the dice has it's own percentage. If you don't believe me, go to Vegas, watch the roulette wheel hit red 5 times in a row, then bet everything on black.

 

Having said that, I agree with you on the trade down, however unlikely it might be. It seems this year in particular has a pretty deep value, and getting some mid-rounders in the 2nd would be similar to late 1st rounders. I'd say the late 1st rounders this year would be slightly lower (let's say 45-50%) and the early/mid 2nd rounders would be higher (let's say 35-40%). If this is something the F.O. believes, then the trade down makes a lot of sense.

[adam]

You're right about your follow-up sentence. That's not how statistics work. Each roll of the dice has it's own percentage. If you don't believe me, go to Vegas, watch the roulette wheel hit red 5 times in a row, then bet everything on black.

 

Having said that, I agree with you on the trade down, however unlikely it might be. It seems this year in particular has a pretty deep value, and getting some mid-rounders in the 2nd would be similar to late 1st rounders. I'd say the late 1st rounders this year would be slightly lower (let's say 45-50%) and the early/mid 2nd rounders would be higher (let's say 35-40%). If this is something the F.O. believes, then the trade down makes a lot of sense.

Are you saying my probability is wrong? I thought you added the odds since they were mutually exclusive. The 2nd pick is like having a 2nd die. If I have a 55.3% chance to roll a 1 on the first roll, a 22.9% chance to roll a 1 with the 2nd die, and a 16.0% chance to roll a 1 on the 3rd die, wouldn't I have a 94.2% chance to roll a 1 overall? The more chances increases the odds?

 

Regardless, I just feel like unless someone drops unexpectedly, we should move down and improve our odds.

[bradjock]

Having said that, I agree with you on the trade down, however unlikely it might be. It seems this year in particular has a pretty deep value, and getting some mid-rounders in the 2nd would be similar to late 1st rounders. I'd say the late 1st rounders this year would be slightly lower (let's say 45-50%) and the early/mid 2nd rounders would be higher (let's say 35-40%). If this is something the F.O. believes, then the trade down makes a lot of sense.

 

I agree with you that based on the nature of talent, it makes sense to trade down. I'd add that it's more critical this year since we have so few players under contract for the 2014 season. Sure we will sign guys one way or another, but you want some degree of stability. It would be ideal if we had 10 draft picks this year in order to add competition and form a young core that would be together for the next few years. Angelo's poor drafting hasn't left us with much of a young base to work with. That being said, for being a "draft guy" Emery didn't hit any home runs in his first year. We need to see big strides from somebody this year. Be it McClellen, Jefferey, Hardin or Rodriguez.

 

 

[jason]

Are you saying my probability is wrong? I thought you added the odds since they were mutually exclusive. The 2nd pick is like having a 2nd die. If I have a 55.3% chance to roll a 1 on the first roll, a 22.9% chance to roll a 1 with the 2nd die, and a 16.0% chance to roll a 1 on the 3rd die, wouldn't I have a 94.2% chance to roll a 1 overall? The more chances increases the odds?

 

Regardless, I just feel like unless someone drops unexpectedly, we should move down and improve our odds.

 

It's actually the same die, rolled individually, multiple times. Same as flipping a coin. Each time you flip it, the odds are almost exactly 50/50. Doesn't matter how many flips you have, each time starts back up at 50/50.

 

I agree with the general idea this year because the draft seems deep.

[azbearsfan]

I agree with moving down. More cost effective.

[adam]

It's actually the same die, rolled individually, multiple times. Same as flipping a coin. Each time you flip it, the odds are almost exactly 50/50. Doesn't matter how many flips you have, each time starts back up at 50/50.

 

I agree with the general idea this year because the draft seems deep.

But it is not the same die, there is a pool of players that can't be picked twice, that reduces by one with every pick, so if the other team misses on a 5-yr starter, it actually increases our chances, that's why it is cumulative. We could just take our own picks for example, if we miss on the first pick, that actually increases our chances of picking a star the next pick because he is still available but there are less overall chances to pick him by the group.

 

Take this example, there are 3 HoFer quality players in this draft, yet we have no clue who they are, but they will be taken in the first round. The percentage would go up every time a pick is placed that is not one of those 3. That is how the draft works, yet there is a built in probability weighted towards the early picks.

 

 

[jason]

But it is not the same die, there is a pool of players that can't be picked twice, that reduces by one with every pick, so if the other team misses on a 5-yr starter, it actually increases our chances, that's why it is cumulative. We could just take our own picks for example, if we miss on the first pick, that actually increases our chances of picking a star the next pick because he is still available but there are less overall chances to pick him by the group.

 

Take this example, there are 3 HoFer quality players in this draft, yet we have no clue who they are, but they will be taken in the first round. The percentage would go up every time a pick is placed that is not one of those 3. That is how the draft works, yet there is a built in probability weighted towards the early picks.

 

But you're equating each player equally. That's flawed. The bolded sentence is reality. When you have 100 players, and take one away, if they are all equally weighted then it's 99 players. But in terms of value, the guys in the first round are worth more than one. So each time you take away a first rounder, the value from 100 goes down by more than one, and that is why the percentages of success decreases by round. Based on just 100, you'd have to give a value to the first round guys (let's say 1.5), second round guys (.7), third round (.4), fourth (.25), fifth (.15), sixth (.1), seventh (.05). I realize those are made up, but they come pretty close to 100. So when the first guy comes off the board, you don't have 99/100, you have 98.5/100. After the first round you have 52/100. After the second round you have 29.6/100. More than likely those 3 HOF players have already been selected. At least 2 of them. If it worked the way you're saying, everyone would trade out of the first two rounds, and get a ton of 6th and 7th rounders. You could draft every single pick in the 7th round if you tried hard enough as a GM to trade away your higher picks. But the simple fact is, even if you had 32 picks in the 7th round, you'd probably hit on less than if you had 10 picks in the first round.

 

BTW - The dice example is probable better explained like this:

1st round - Coin

2nd round - 4-sided die

3rd round - 6-sided die

4th-6th round - dungeons and dragons dice of increasing size

7th round - roulette wheel

[adam]

But you're equating each player equally. That's flawed. The bolded sentence is reality. When you have 100 players, and take one away, if they are all equally weighted then it's 99 players. But in terms of value, the guys in the first round are worth more than one. So each time you take away a first rounder, the value from 100 goes down by more than one, and that is why the percentages of success decreases by round. Based on just 100, you'd have to give a value to the first round guys (let's say 1.5), second round guys (.7), third round (.4), fourth (.25), fifth (.15), sixth (.1), seventh (.05). I realize those are made up, but they come pretty close to 100. So when the first guy comes off the board, you don't have 99/100, you have 98.5/100. After the first round you have 52/100. After the second round you have 29.6/100. More than likely those 3 HOF players have already been selected. At least 2 of them. If it worked the way you're saying, everyone would trade out of the first two rounds, and get a ton of 6th and 7th rounders. You could draft every single pick in the 7th round if you tried hard enough as a GM to trade away your higher picks. But the simple fact is, even if you had 32 picks in the 7th round, you'd probably hit on less than if you had 10 picks in the first round.

 

BTW - The dice example is probable better explained like this:

1st round - Coin

2nd round - 4-sided die

3rd round - 6-sided die

4th-6th round - dungeons and dragons dice of increasing size

7th round - roulette wheel

I got it, and based on the analysis, 11 picks from 150-189 would net you one 5-yr starter, almost guaranteed. The odds of getting 1 with each pick is 8.9%, so if you had 11 picks in that group you would have 11*8.9%= 97.9%. That is all I am saying. So collectively, to move back and get more picks (depending on where they fall) can give you better odds in finding a star, especially with certain picks. So for us, 20 could turn in 30-something and 70-something basically increasing our odds by 22.9% (again I know this is not the exact number but close).

[selection7]

The obvious thing that jumps out at me is that "5-year starter" was a bad choice, or at least not the best, for a useful metric. Everyone knows the highest picks get the most opportunities without having to prove anything (Cedric Benson anyone?). Tom Brady, for example, wouldn't count as a 5-year starter because he didn't start his first year. So a metric that misses a Tom Brady could stand to be improved. They should have gone with something like "Players that by the end of their fifth year...had started at least years 3-5". If you're starting your third-through-fifth years, you are a quality NFL player with likely a full-length career ahead of you, regardless of what happened your rookie year.

 

It follows, then, that the chances of finding a quality player in the highest positions is actually lower than advertised by the OP's referenced stats. Tom Brady is just an extreme example of it...Tim Jennings would be a more typical case.

[adam]

The obvious thing that jumps out at me is that "5-year starter" was a bad choice, or at least not the best, for a useful metric. Everyone knows the highest picks get the most opportunities without having to prove anything (Cedric Benson anyone?). Tom Brady, for example, wouldn't count as a 5-year starter because he didn't start his first year. So a metric that misses a Tom Brady could stand to be improved. They should have gone with something like "Players that by the end of their fifth year...had started at least years 3-5". If you're starting your third-through-fifth years, you are a quality NFL player with likely a full-length career ahead of you, regardless of what happened your rookie year.

 

It follows, then, that the chances of finding a quality player in the highest positions is actually lower than advertised by the OP's referenced stats. Tom Brady is just an extreme example of it...Tim Jennings would be a more typical case.

Yeah I guess I didn't clarify that it was a 5-yr starter out of 7 years, and they only had to start 8 games to qualify. This seems reasonable since it is well beyond the length of rookie contracts and gives them 7 years to be a starter in 5. It is not perfect, but seems like a good stab. You essentially could break this down any way you wanted, based on Pro Bowls, All-Pro teams, etc.

[AZ54]

They freely admit it's hard to pick what starter means. I think they've done a good job of giving a reasonable definition to it by making it for 5 years. My remark about Jacksonville having 22 starters is just because I've seen so many below average players hang around as starters.

 

Years ago I read some scout of front office personnel person said the league is really about blue chip players. I now lean more toward that way of thinking. Every team has it's "starters" or as Bill Parcells said JAGs (Just Another Guy). Add in a couple blue chip players to the mix like a young Urlacher and it dramatically changes things. Scheme and coaching are what you need to keep the JAGs above average productive. New England is a great example because you have blue chip Brady and a few other good players (Welker) plus some JAGs. How many players left their system and didn't put up the same numbers? Or didn't even look like the same player?

 

So 5 yr starter means something but the draft is where you need to find some elite talent. Picking later in the first round doesn't mean it can't happen and I'd say Ozzie Newsome is one of the best at this. Ed Reed 24th overall. Ray Lewis 26th overall. Terrell Suggs 10th overall (when many said he was too slow).

 

Right now Emery's first draft doesn't overwhelm with potential for the future beyond saying there might some 5 yr starters.