Made For You: Customizable Draft Pick Trade Chart
Posted: Sun Apr 09, 2017 12:58 pm
TL/DR at the bottom.
---> CHART HERE. TO USE IT, YOU NEED TO "MAKE A COPY" OR "DOWNLOAD AS" <---
I hate all draft pick charts. Too many people live by them without knowing the math behind creating them. Jimmy Johnson..... who isn't known for being the brightest bulb in the box..... created one that, for some reason, the entire NFL lives by. If you look up success rates vs draft picks, you will eventually find a link based on previous NFL Draft Data that shows JJs chart is actually flawed. High picks are valued too highly, and 2nd rounders aren't valued nearly enough, etc. Too many current draft pick charts are based on that first chart that Jerry Jones made.
This motivated me to make my own. I was a double major in college (statistics & mathematics), and went on to blah blah blah..... I listed my credentials here originally but who the bleep cares. Just know that I take this seriously. I had a ton of time to research the past few weeks, and I failed and gave up about 5 times before finally LOVING what I'm about to present to you.
================================================
MAKING THE CUSTOMIZABLE CHART
================================================
First, I needed to figure out what made a winning dynasty team. I analyzed all fantasy football playoff teams I could find (which amounted to dozens.... this was tiring) and I found out that the majority of these teams had on average, 2.07 "studs", 2.84 "starters", and then several "role players". It was actually lower than I realized, but one thing was for sure: You had to have a STUD, and most likely you had to have 2. No studs = no playoffs. For reference: Studs are players who you would give 2 or more 1st rounders for. Starters are worth 1 first. Role Players are worth a 2nd.
Next, I researched every draft pick from 2004-2014 using MFLs own ADP data from actual rookie drafts (I purposely didn't do 2015 & 2016 since those players are still changing in value). I then determined which of those players, on average, evolved to become studs (OBJ, Chris Johnson, AP, etc), starters (Frank Gore, Jon Stewart, Emmanuel Sanders, etc), or role players (Vernon Davis, Kendall Wright, etc). Anyone less than a "Roll Player" status was ignored, as they had little-to-no value in the long run. I also had to research the range of these players to go in the draft. For example, Mike Evans was both a 1.01 & 1.02 pick, so that was his range. Carlos Hyde was a 1.03 to 1.07 player, so that was his range, etc. Standard deviation was played with but ultimately dropped, since nearly every player in the middle of the 1st round had a SD of ~1 slot.
The result of where studs can be found in the draft was a little surprising. For example, there was barely any difference in success rate from the 1.01 to 1.03 selections. Success rate for these picks resulted in the same ~42% "Stud" rate. The 1.04-1.06 selections dropped to ~21% Stud Rate. The fail rate, which is where you do not draft one of the 3 kinds of players mentioned above, went from a 50% average in round 1, to an 80% in round 2, 84% in round 3, and a 94% fail rate in round 4. Talent dropped significantly from the top 3 picks to just the 2nd round (as it should, btw). It should look like a y=1/x line (where x>0) in order for the research to be considered valid, and when that happened, I figured I was onto something. All 4th round picks were shots in the dark. The success rate of the 4.01 pick (or any 4th rounder, for that matter) was not statistically significant over the 4.12 pick, so all 4th round picks were given the same value in my chart.
The next thing I had to do........ and this is where you come in....... is give the 3 types of players a numerical value. It all starts with role players and works up. I gave all role players a value of "1". Then, I had to determine how many role players it would take to equal a starter, assuming trade value. Given feedback on these forums, I assumed 3 role players = 1 starter. Last, I deemed Studs to have slightly greater value than 2 Starters, and since Starters were 3 points a piece, I made Studs = 7 points. The chart I made gives you the power to adjust your personal rankings, and therefore your own draft pick chart.
I then needed to find the expected value of each draft pick. For example, the 1.01 pick yielded a Stud rate of ~41%, a Starter rate of ~20%, and a Role Player rate ~5%. 41% of 7 plus 20% of 3 plus 5% of 1 = Your Expected Value for this draft pick. Since the numbers got small, I multiplied each one by 1,000 to give each draft pick what I consider to be a very accurate measure of value, using expected value based on previous draft data, which I think is the best predictor of future results.
The reason why I like this chart is simple: It follows the exponential decay associated with player value, and gives each user a personalized value based on their perception of player value. It's not a "one-size-fits-all" model. If you disagree with my assessment on player value, that's fine. You can make your own!
To use or play with it, just hit the link below. If you have a google account, go to "File - Make A Copy". If you don't have a Google Account, just go to "Download As..." and save it on your hard drive. I wrote steps on how to use the chart in the chart itself.
TL/DR: I researched and made what I consider (opinion) to be the most-accurate draft pick chart you will find, since it's based on outcomes of 10 years worth of drafts, merged with the success rates of fantasy football playoff teams:
---> CHART HERE. TO USE IT, YOU NEED TO "MAKE A COPY" OR "DOWNLOAD AS" <---
Open to comments, questions, or suggestions.
Skeptical? You should try it out on older drafts. 2015s and beyond. I just did on 2014 & 2013s. Look at the transaction history and compare that to how the draft shakes out. On my end, it was flawless. I didn't find one trade that didn't rule in favor of the winning side.
EDIT: Updated (4/9/17 - 4:36 PM)
---- I forgot to add in the variance formula. This gives slightly more value to the picks at the beginning of rounds 2, 3, and 4.
EDIT: Updated with several add-ons (4/10/17 - 6:52 PM)
--- Trade Calculator
--- Player Value Percentage Chart
--- Updated Rankings (Doesn't Change The Value Of Ratio Of Picks In The 1st Round, But It Does For The Rest)
EDIT: Updated (2/10/18)
--- 2015 Draft Stats Added / Draft Picks Adjusted
---> CHART HERE. TO USE IT, YOU NEED TO "MAKE A COPY" OR "DOWNLOAD AS" <---
I hate all draft pick charts. Too many people live by them without knowing the math behind creating them. Jimmy Johnson..... who isn't known for being the brightest bulb in the box..... created one that, for some reason, the entire NFL lives by. If you look up success rates vs draft picks, you will eventually find a link based on previous NFL Draft Data that shows JJs chart is actually flawed. High picks are valued too highly, and 2nd rounders aren't valued nearly enough, etc. Too many current draft pick charts are based on that first chart that Jerry Jones made.
This motivated me to make my own. I was a double major in college (statistics & mathematics), and went on to blah blah blah..... I listed my credentials here originally but who the bleep cares. Just know that I take this seriously. I had a ton of time to research the past few weeks, and I failed and gave up about 5 times before finally LOVING what I'm about to present to you.
================================================
MAKING THE CUSTOMIZABLE CHART
================================================
First, I needed to figure out what made a winning dynasty team. I analyzed all fantasy football playoff teams I could find (which amounted to dozens.... this was tiring) and I found out that the majority of these teams had on average, 2.07 "studs", 2.84 "starters", and then several "role players". It was actually lower than I realized, but one thing was for sure: You had to have a STUD, and most likely you had to have 2. No studs = no playoffs. For reference: Studs are players who you would give 2 or more 1st rounders for. Starters are worth 1 first. Role Players are worth a 2nd.
Next, I researched every draft pick from 2004-2014 using MFLs own ADP data from actual rookie drafts (I purposely didn't do 2015 & 2016 since those players are still changing in value). I then determined which of those players, on average, evolved to become studs (OBJ, Chris Johnson, AP, etc), starters (Frank Gore, Jon Stewart, Emmanuel Sanders, etc), or role players (Vernon Davis, Kendall Wright, etc). Anyone less than a "Roll Player" status was ignored, as they had little-to-no value in the long run. I also had to research the range of these players to go in the draft. For example, Mike Evans was both a 1.01 & 1.02 pick, so that was his range. Carlos Hyde was a 1.03 to 1.07 player, so that was his range, etc. Standard deviation was played with but ultimately dropped, since nearly every player in the middle of the 1st round had a SD of ~1 slot.
The result of where studs can be found in the draft was a little surprising. For example, there was barely any difference in success rate from the 1.01 to 1.03 selections. Success rate for these picks resulted in the same ~42% "Stud" rate. The 1.04-1.06 selections dropped to ~21% Stud Rate. The fail rate, which is where you do not draft one of the 3 kinds of players mentioned above, went from a 50% average in round 1, to an 80% in round 2, 84% in round 3, and a 94% fail rate in round 4. Talent dropped significantly from the top 3 picks to just the 2nd round (as it should, btw). It should look like a y=1/x line (where x>0) in order for the research to be considered valid, and when that happened, I figured I was onto something. All 4th round picks were shots in the dark. The success rate of the 4.01 pick (or any 4th rounder, for that matter) was not statistically significant over the 4.12 pick, so all 4th round picks were given the same value in my chart.
The next thing I had to do........ and this is where you come in....... is give the 3 types of players a numerical value. It all starts with role players and works up. I gave all role players a value of "1". Then, I had to determine how many role players it would take to equal a starter, assuming trade value. Given feedback on these forums, I assumed 3 role players = 1 starter. Last, I deemed Studs to have slightly greater value than 2 Starters, and since Starters were 3 points a piece, I made Studs = 7 points. The chart I made gives you the power to adjust your personal rankings, and therefore your own draft pick chart.
I then needed to find the expected value of each draft pick. For example, the 1.01 pick yielded a Stud rate of ~41%, a Starter rate of ~20%, and a Role Player rate ~5%. 41% of 7 plus 20% of 3 plus 5% of 1 = Your Expected Value for this draft pick. Since the numbers got small, I multiplied each one by 1,000 to give each draft pick what I consider to be a very accurate measure of value, using expected value based on previous draft data, which I think is the best predictor of future results.
The reason why I like this chart is simple: It follows the exponential decay associated with player value, and gives each user a personalized value based on their perception of player value. It's not a "one-size-fits-all" model. If you disagree with my assessment on player value, that's fine. You can make your own!
To use or play with it, just hit the link below. If you have a google account, go to "File - Make A Copy". If you don't have a Google Account, just go to "Download As..." and save it on your hard drive. I wrote steps on how to use the chart in the chart itself.
TL/DR: I researched and made what I consider (opinion) to be the most-accurate draft pick chart you will find, since it's based on outcomes of 10 years worth of drafts, merged with the success rates of fantasy football playoff teams:
---> CHART HERE. TO USE IT, YOU NEED TO "MAKE A COPY" OR "DOWNLOAD AS" <---
Open to comments, questions, or suggestions.
Skeptical? You should try it out on older drafts. 2015s and beyond. I just did on 2014 & 2013s. Look at the transaction history and compare that to how the draft shakes out. On my end, it was flawless. I didn't find one trade that didn't rule in favor of the winning side.
EDIT: Updated (4/9/17 - 4:36 PM)
---- I forgot to add in the variance formula. This gives slightly more value to the picks at the beginning of rounds 2, 3, and 4.
EDIT: Updated with several add-ons (4/10/17 - 6:52 PM)
--- Trade Calculator
--- Player Value Percentage Chart
--- Updated Rankings (Doesn't Change The Value Of Ratio Of Picks In The 1st Round, But It Does For The Rest)
EDIT: Updated (2/10/18)
--- 2015 Draft Stats Added / Draft Picks Adjusted