04-06-2020, 02:02 AM
(This post was last modified: 04-06-2020, 03:14 AM by iStegosauruz.)
[div align=\\\"center\\\"]Background and Methodology[/div]
I simulated approximately 82,800 games so that you wouldn’t have to.
Unlike last time, that sentence isn’t misleading. The last time I studied cornerbacks I simulated approximately 100,800 games but had to toss out half the data because Windows accidently deleted half the files for it that I needed. The sample was only 50,400 games. This time, however, I simulated 82,800 games and used all of those games in the data set.
The last cornerback study I ran tested the effects of hands, intelligence, tackle, and strength at various levels on the average cornerback. This time around I tested the impact of all valuable attributes - agility, hands, intelligence, endurance, speed, strength, and tackle - on a low TPE cornerback. The control - or baseline - in this study is the average of the starting values for the various cornerback archetypes. Those averages look like this:
[div align=\\\"center\\\"][/div]
I tested all of the previously listed attributes in increments of 5. I tested them up to the values of the average CB in the league, which was calculated in my last cornerback study. For reference, those values are:
[div align=\\\"center\\\"][/div]
What this means is I would add 5 to each attribute until it was within 5 of the same attribute on the average NSFL cornerback. To keep values consistent, when I added the 5 to the average cornerback archetype I would round it up to the nearest value or 5 or 10. To make this clearer, that means I tested:
[div align=\\\"center\\\"]Agility at 75 and 80
Hand at 60, 65, and 70
Intelligence at 60, 65, and 70
Endurance at 65, 70, 75, and 80
Speed at 70, 75, 80, 85, and 90
Tackle at 50 and 55
[/div]
As always I used the Orange County Otters as the control team for this study. I edited their two starting cornerbacks - AJ Lattimer and Kacey Dream - each time to give them the appropriate value. I also modified their height and weight to be the average height and weight of the cornerback archetypes - 6’2” and 207 pounds.
The Otters played each other team in the league four times at each variable amount. Two of those games were at home and two of those games were on the road. That means there were four 900 simulation batches at each variable level. That means each variable level had 3,600 simulations in them. There were 23 different variable levels, meaning there were a total of 82,800 simulations run for this study.
I pulled two sets of data from each simulation - the game logs and the game stat logs for the players. I used the game log to calculate a variety of metrics including win percentage, average opponent total yards, average opponent pass yards, average opponent yard per attempt, average opponent rush yards, average opponent yards per rush, average opponent interceptions thrown per game, average opponent fumbles per game, and average opponent fumbles lost per game.
I used the game stat logs for the players to find a variety of individual metrics including the average interceptions each player had per game, the average number of interception return yards each player had per game, the average number of passes defended each player had per game, the average number of tackles each player had each game, the average number of fumbles each player forced per game, and the average number of defensive touchdowns each player had each game.
[div align=\\\"center\\\"]Agility
Team Statistics [/div]
For the three cornerback archetypes, agility starts at either 65 or 70. In general, increasing the cornerbacks’ agility had a marginal but not particularly substantive impact on a team’s winning percentage
[div align=\\\"center\\\"][/div]
The control group had a winning percentage of 37.08% while the highest agility value tested - 80 - had a winning percentage of 37.94%. This is a minor increase of 0.86%. The 75 agility group that was tested had the highest winning percentage of any of the groups at 39.22%. This is a 2.14% increase over the control group. There is such a large gap between the 75 agility group’s winning percentage and the 80 agility group’s winning percentage that I don’t think there is a strong correlation between agility and a team’s winning percentage. There is obviously an impact, both of the variable groups had a higher winning percentage than the control group, however it is not a consistent increase and there is a large gap between the two values produced by the variable groups.
[div align=\\\"center\\\"][/div]
Opponents consistently averaged more total yards per game as the cornerbacks’ agility increased. This was partially due to the rush yards allowed being consistently higher as the agility value increased. The pass yards surrendered increased consistently as the agility value increased as well. The increases were slight - 3.04 total yards per game more from the 80 agility group than the control group - so I don’t think there is a negative relationship between cornerback agility and opponent offensive production, however I don’t think there is a particularly strong positive correlation between an increase in cornerback agility value and a decrease in opponent offensive production.
[div align=\\\"center\\\"]
[/div]
There was not a substantive change in the amount of interceptions the opposing team surrendered per game. Cornerback agility does not seem to have an impact on the amount of interceptions a team forces.
[div align=\\\"center\\\"][/div]
[div align=\\\"center\\\"]Player Statistics [/div]
The increase in agility value did not cause a substantial or consistent change in any of the individual statistics I tracked. At best, agility might have a minor impact on the number of interception return yards each cornerback gets as well as the average amount of passes they defend per game and the average number of tackles they have per game. The correlation is not strong enough to make a definite conclusion about those theories, however.
[div align=\\\"center\\\"][/div]
[div align=\\\"center\\\"]Hands
Team Statistics [/div]
For the three cornerback archetypes, hands start at either 50 or 55. In general, increasing the hands value on the cornerbacks did not cause a consistent increase in the winning percentage for the Otters.
[div align=\\\"center\\\"][/div]
All of the variable groups allowed less total yards per game to their opponents, on average, however the decrease from the control group was not consistent. This is mainly due to the fluctuation in the average rush yards allowed to opponents per game. The average pass yards allowed to opponents per game consistently decreased - from 256.89 in the control group to 255.12 at the highest variable group - 70 hands. This is a decrease in 1.77 pass yards per game.
[div align=\\\"center\\\"]
[/div]
The increase in hands did not cause a consistent increase in the amount of interceptions the opposing team threw per game. The control group was the highest value. There does not seem to be a strong correlation between the increase of the hands value and an increase in interceptions.
[div align=\\\"center\\\"][/div]
[div align=\\\"center\\\"]Player Statistics [/div]
There is not a strong correlation between the increase in the hands value and the increase in any particular individual statistics. Average interception return yards per game, average passes defended, and average tackles per game all have variable groups that produce higher metrics than the control group, however they also have variable groups that produce lower metrics than the control group.
[div align=\\\"center\\\"][/div]
[div align=\\\"center\\\"]Intelligence
Team Statistics [/div]
For the three cornerback archetypes, intelligence starts at either 50, 55, or 60. In general, increasing the cornerbacks’ intelligence had a marginal impact on a team’s winning percentage. This impact is not consistent. The highest variable group tested - 70 intelligence - had the closest winning percentage to that of the control group.
[div align=\\\"center\\\"]
[/div]
Opponents averaged less total yards per game, on average, in the variable groups, however it was not a consistent decrease and there was one major outlier. The 60 intelligence variable group allowed 369.73 total yards per game to their opponents - 6.72 total yards per game more than the control group. This is such a large increase and all the other variable groups are lower than the control, so I think it's an outlier. Opponents' pass yards per game and rush yards per game did not consistently decrease either.
[div align=\\\"center\\\"]
[/div]
There was also not a consistent increase in the amount of interceptions the opponents averaged per game as the intelligence value increased either.
[div align=\\\"center\\\"][/div]
[div align=\\\"center\\\"]Player Statistics[/div]
The only individual statistics that was consistently impacted by the increase in the intelligence value was the amount of tackles each cornerback averaged per game. If you exclude the 60 intelligence variable group because it was such an outlier the average tackles per game gradually increase. The 70 intelligence variable group averaged 1.535 tackles per game - 0.056 more tackles per game than the control group.
[div align=\\\"center\\\"][/div]
[div align=\\\"center\\\"]Endurance
Team Statistics [/div]
For the three cornerback archetypes, endurance starts at either 55 or 60. In general, increasing the cornerbacks’ endurance had a marginal impact on the Otters’ winning percentage.
[div align=\\\"center\\\"][/div]
The 65 endurance variable group was lower than the control by 0.25%, however these values are close enough to each other I think it is just an outlier in the sim. The other three variable groups tested increased the Otters’ winning percentage consistently as the endurance value increased.
[div align=\\\"center\\\"][/div]
Opponents consistently averaged less total yards per game, pass yards per game, and rush yards per game as the endurance value increased. The only exception to this is for the total yards and total pass yards at the 80 endurance variable group, which is most likely an outlier in this situation. Total yards decreased from 363.01 per game in the control group to 360.21 per game in at the lowest - in the 75 endurance variable group. This is a 2.8 yard per game decrease. Pass yards per game decreased from 256.89 per game in the control group to 254.17 at the lowest - in the 75 endurance variable group. This is a 2.72 yard per game decrease.
[div align=\\\"center\\\"]
[/div]
Opponent average interceptions per game were not substantially impacted by the increase in endurance.
[div align=\\\"center\\\"][/div]
[div align=\\\"center\\\"]Player Statistics [/div]
In general, no individual statistics for the cornerbacks consistently increased as the endurance value increased. There were variable groups for all statistics tracked that were higher than the control, however it did not follow any form of a pattern.
[div align=\\\"center\\\"][/div]
[div align=\\\"center\\\"]Speed
Team Statistics [/div]
For the three cornerback archetypes, speed starts at either 60, 65, or 70. In general, increasing the cornerbacks’ speed had a significant impact on the Otters’ winning percentage.
[div align=\\\"center\\\"][/div]
The control group had a winning percentage of 37.08%. All other variable groups had a winning percentage higher than this. As the speed value increased, the winning percentage increased in all groups except the 75 speed variable group which saw a minor dip in winning percentage from the 70 speed variable group. This is most likely an outlier. The highest speed variable group tested - 90 speed - had a winning percentage of 44.97%. This is 7.89% higher than the control group.
[div align=\\\"center\\\"][/div]
The 70 speed and 75 speed variable groups allowed a small amount more total yards per game to their opponents than the control group, however the increase in both was so small it is probably just an outlier or a bump in the sim. From the 80 speed variable group onwards the Otters’ opponents averaged less total yards per game than the control. At the highest speed variable group - 90 speed - the Otters allowed 10.11 total yards per game less than the control group.
[div align=\\\"center\\\"][/div]
Opponents passing yards per game followed a similar trend to total yards per game. The 75 speed variable group saw a small increase in the average pass yards per game surrendered to opponents, however it is probably an outlier. Discounting that grouping, as speed increased the average pass yards per game allowed to opponents decreased. The highest variable group - 90 speed - allowed 10.72 less pass yards per game to opponents than the control group.
[div align=\\\"center\\\"]
[/div]
Team total interceptions were not substantially impacted by speed.
[div align=\\\"center\\\"][/div]
[div align=\\\"center\\\"]Player Statistics [/div]
As speed increased the cornerbacks saw a fairly consistent increase in the average number of individual interception return yards per game, the average number of individual passes defended per game, and the average number of individual tackles per game.
[div align=\\\"center\\\"][/div]
Average individual passes defended increased from 0.939 per game in the control group to 0.968 per game in the 90 speed variable group - a 0.029 passes defended per game increase.
[div align=\\\"center\\\"][/div]
Average individual tackles per game increased from 1.479 per game in the control group to 2.373 per game at the highest point - in the 85 speed variable. This is a 0.894 tackle per game difference.
[div align=\\\"center\\\"][/div]
[div align=\\\"center\\\"]Strength
Team Statistics [/div]
For the three cornerback archetypes, strength starts at either 30, 35, or 40. In general, increasing the cornerbacks’ strength had a marginal impact on a team’s winning percentage, although it was not a consistent increase.
[div align=\\\"center\\\"][/div]
The control group had a winning percentage of 37.08% while the highest strength value tested - 50 strength - had a 39.67% winning percentage. This is a 2.59% difference. The 50 strength variable group did not have the highest winning percentage of the variable groups, however. The variable group that produced the highest winning percentage was the 40 strength group. The 40 strength variable group had a 39.78% winning percentage - 2.7% higher than the control group. The increase in winning percentage for each of the variable groups implies to me that there is some correlation between strength and winning percentage but it is not a consistent positive correlation.
[div align=\\\"center\\\"][/div]
Opponent total yards per game decreased fairly consistently as the strength value increased. The 50 strength group had allowed more total yards per game than the 45 strength group, however it was a slight amount and is most likely an outlier. Opponent pass yards per game also decreased fairly consistently as strength increased. The control group allowed 256.89 pass yards per game, while the 50 strength variable group allowed 248.52 pass yards per game - a 8.37 pass yard per game difference.
[div align=\\\"center\\\"]
[/div]
Interceptions were not significantly impacted by strength.
[div align=\\\"center\\\"][/div]
[div align=\\\"center\\\"]Player Statistics [/div]
The average number of individual passes defended and the average number of individual tackles increased consistently as strength increased.
[div align=\\\"center\\\"][/div]
The average number of individual passes defended increased from 0.939 per game in the control group to 0.966 per game in the 50 strength variable group. This is an increase of 0.027 individual passes defended per game.
[div align=\\\"center\\\"][/div]
The average number of individual tackles per game increased from 1.479 in the control group to 1.762 in the 50 strength variable group. This is an increase of 0.283 individual tackles per game.
[div align=\\\"center\\\"][/div]
[div align=\\\"center\\\"]Tackling
Team Statistics [/div]
For the three cornerback archetypes, tackling starts at either 40 or 45. In general, increasing the cornerbacks’ tackling had a marginal but not particularly substantive impact on a team’s winning percentage. Its impact was very similar to that of agility’s.
[div align=\\\"center\\\"][/div]
The control group had a winning percentage of 37.08 while the highest value tested - 55 tackling - had a winning percentage of 37.61. This is a 0.53% increase. The 55 tackling group did not generate the biggest increase in winning percentage over the control group, however. The 50 tackling group produced a winning percentage of 37.94%. This is a 0.86% difference from the control group. These values are so close together across the board that I would hypothesize that tackling has a very minimal impact on a team’s winning percentage.
[div align=\\\"center\\\"][/div]
Opponent total yards per game consistently declined as tackling increased, however the reductions were minimal. The control group allowed 363.01 total yards per game to opponents, while the 55 tackling variable group allowed 362.65 total yards per game to opponents. This is a fairly insubstantial difference of 0.36 total yards per game. Opponent pass yards per game did not consistently decrease, nor did opponent rush yards per game.
[div align=\\\"center\\\"]
[/div]
Interceptions were not particularly impacted by tackling.
[div align=\\\"center\\\"][/div]
[div align=\\\"center\\\"]Player Statistics [/div]
Increasing the tackling attribute caused a minor increase in the average number of individual tackles the players got per game. That was the only metric to consistently change at all.
[div align=\\\"center\\\"][/div]
The control group averaged 1.479 individual tackles per game. The 55 tackling variable group averaged 1.515 individual tackles per game. That is a 0.036 individual tackle per game difference.
[div align=\\\"center\\\"][/div]
[div align=\\\"center\\\"]Conclusions [/div]
1. Speed is in a category of its own when it comes to cornerback attributes.
2. Strength is the second best cornerback attribute, in my opinion. It increases a team’s winning percentage, reduces opponent yards fairly consistently, and impacts two individual statistics.
3. Endurance has an impact, however it seems to be more of a complimentary attribute. I need to test its value in conjunction with other attributes. This makes sense in theory, however, because without other statistics to use there is nothing to be prolonged by endurance.
4. Intelligence and tackling seem to the third tier of cornerback attributes - if you exclude endurance. They produce minor changes in win% as they increase, minor reductions in opponent total yards, and both impact an individual statistic.
5. Hands and Agility seem to be two of the less impactful cornerback attributes.
6. There is the potential that agility is like endurance and is a complimentary attribute, however on its own or early in a cornerback build it seems to have fairly little impact. When tested it caused a minor winning percentage increase but did not reduce an opponent’s offensive output and did not impact any individual statistics.
7. Hands also did not cause a significant impact in winning percentage. It caused a small decrease in an opponent’s offensive output, however that impact was smaller than that of intelligence, tackling, and endurance. It did not cause a consistent increase in individual cornerback attributes. I also have the theory that hands could be a complimentary attribute as well. Essentially, having good hands does you no good if you’re never in the play. If you have enough speed and intelligence to be in the play hands might have a greater impact.
8. This is not a foolproof study by any means. I’m still figuring out my methodology in a lot of cases and there is so much to explore that I’ll inevitably do more work on cornerbacks.
[div align=\\\"center\\\"]Notes[/div]
1. I’m finishing this up at 2am for me, so pardon any grammar mistakes or potential errors. I try to proofread these articles fairly well but I’m very drowsy, so I might have missed something.
2. This is probably the last sim study I will publish before the draft. I need to preserve some element of competitive advantage for myself and my future team. I also have a few private theories to test before I have anything else public. Rest assured, however, at some point I'll publish everything I do. I've gotta make that sweet media money.
3. As always, all of my work is open source. You can find the data here.
I simulated approximately 82,800 games so that you wouldn’t have to.
Unlike last time, that sentence isn’t misleading. The last time I studied cornerbacks I simulated approximately 100,800 games but had to toss out half the data because Windows accidently deleted half the files for it that I needed. The sample was only 50,400 games. This time, however, I simulated 82,800 games and used all of those games in the data set.
The last cornerback study I ran tested the effects of hands, intelligence, tackle, and strength at various levels on the average cornerback. This time around I tested the impact of all valuable attributes - agility, hands, intelligence, endurance, speed, strength, and tackle - on a low TPE cornerback. The control - or baseline - in this study is the average of the starting values for the various cornerback archetypes. Those averages look like this:
[div align=\\\"center\\\"][/div]
I tested all of the previously listed attributes in increments of 5. I tested them up to the values of the average CB in the league, which was calculated in my last cornerback study. For reference, those values are:
[div align=\\\"center\\\"][/div]
What this means is I would add 5 to each attribute until it was within 5 of the same attribute on the average NSFL cornerback. To keep values consistent, when I added the 5 to the average cornerback archetype I would round it up to the nearest value or 5 or 10. To make this clearer, that means I tested:
[div align=\\\"center\\\"]Agility at 75 and 80
Hand at 60, 65, and 70
Intelligence at 60, 65, and 70
Endurance at 65, 70, 75, and 80
Speed at 70, 75, 80, 85, and 90
Tackle at 50 and 55
[/div]
As always I used the Orange County Otters as the control team for this study. I edited their two starting cornerbacks - AJ Lattimer and Kacey Dream - each time to give them the appropriate value. I also modified their height and weight to be the average height and weight of the cornerback archetypes - 6’2” and 207 pounds.
The Otters played each other team in the league four times at each variable amount. Two of those games were at home and two of those games were on the road. That means there were four 900 simulation batches at each variable level. That means each variable level had 3,600 simulations in them. There were 23 different variable levels, meaning there were a total of 82,800 simulations run for this study.
I pulled two sets of data from each simulation - the game logs and the game stat logs for the players. I used the game log to calculate a variety of metrics including win percentage, average opponent total yards, average opponent pass yards, average opponent yard per attempt, average opponent rush yards, average opponent yards per rush, average opponent interceptions thrown per game, average opponent fumbles per game, and average opponent fumbles lost per game.
I used the game stat logs for the players to find a variety of individual metrics including the average interceptions each player had per game, the average number of interception return yards each player had per game, the average number of passes defended each player had per game, the average number of tackles each player had each game, the average number of fumbles each player forced per game, and the average number of defensive touchdowns each player had each game.
[div align=\\\"center\\\"]Agility
Team Statistics [/div]
For the three cornerback archetypes, agility starts at either 65 or 70. In general, increasing the cornerbacks’ agility had a marginal but not particularly substantive impact on a team’s winning percentage
[div align=\\\"center\\\"][/div]
The control group had a winning percentage of 37.08% while the highest agility value tested - 80 - had a winning percentage of 37.94%. This is a minor increase of 0.86%. The 75 agility group that was tested had the highest winning percentage of any of the groups at 39.22%. This is a 2.14% increase over the control group. There is such a large gap between the 75 agility group’s winning percentage and the 80 agility group’s winning percentage that I don’t think there is a strong correlation between agility and a team’s winning percentage. There is obviously an impact, both of the variable groups had a higher winning percentage than the control group, however it is not a consistent increase and there is a large gap between the two values produced by the variable groups.
[div align=\\\"center\\\"][/div]
Opponents consistently averaged more total yards per game as the cornerbacks’ agility increased. This was partially due to the rush yards allowed being consistently higher as the agility value increased. The pass yards surrendered increased consistently as the agility value increased as well. The increases were slight - 3.04 total yards per game more from the 80 agility group than the control group - so I don’t think there is a negative relationship between cornerback agility and opponent offensive production, however I don’t think there is a particularly strong positive correlation between an increase in cornerback agility value and a decrease in opponent offensive production.
[div align=\\\"center\\\"]
[/div]
There was not a substantive change in the amount of interceptions the opposing team surrendered per game. Cornerback agility does not seem to have an impact on the amount of interceptions a team forces.
[div align=\\\"center\\\"][/div]
[div align=\\\"center\\\"]Player Statistics [/div]
The increase in agility value did not cause a substantial or consistent change in any of the individual statistics I tracked. At best, agility might have a minor impact on the number of interception return yards each cornerback gets as well as the average amount of passes they defend per game and the average number of tackles they have per game. The correlation is not strong enough to make a definite conclusion about those theories, however.
[div align=\\\"center\\\"][/div]
[div align=\\\"center\\\"]Hands
Team Statistics [/div]
For the three cornerback archetypes, hands start at either 50 or 55. In general, increasing the hands value on the cornerbacks did not cause a consistent increase in the winning percentage for the Otters.
[div align=\\\"center\\\"][/div]
All of the variable groups allowed less total yards per game to their opponents, on average, however the decrease from the control group was not consistent. This is mainly due to the fluctuation in the average rush yards allowed to opponents per game. The average pass yards allowed to opponents per game consistently decreased - from 256.89 in the control group to 255.12 at the highest variable group - 70 hands. This is a decrease in 1.77 pass yards per game.
[div align=\\\"center\\\"]
[/div]
The increase in hands did not cause a consistent increase in the amount of interceptions the opposing team threw per game. The control group was the highest value. There does not seem to be a strong correlation between the increase of the hands value and an increase in interceptions.
[div align=\\\"center\\\"][/div]
[div align=\\\"center\\\"]Player Statistics [/div]
There is not a strong correlation between the increase in the hands value and the increase in any particular individual statistics. Average interception return yards per game, average passes defended, and average tackles per game all have variable groups that produce higher metrics than the control group, however they also have variable groups that produce lower metrics than the control group.
[div align=\\\"center\\\"][/div]
[div align=\\\"center\\\"]Intelligence
Team Statistics [/div]
For the three cornerback archetypes, intelligence starts at either 50, 55, or 60. In general, increasing the cornerbacks’ intelligence had a marginal impact on a team’s winning percentage. This impact is not consistent. The highest variable group tested - 70 intelligence - had the closest winning percentage to that of the control group.
[div align=\\\"center\\\"]
[/div]
Opponents averaged less total yards per game, on average, in the variable groups, however it was not a consistent decrease and there was one major outlier. The 60 intelligence variable group allowed 369.73 total yards per game to their opponents - 6.72 total yards per game more than the control group. This is such a large increase and all the other variable groups are lower than the control, so I think it's an outlier. Opponents' pass yards per game and rush yards per game did not consistently decrease either.
[div align=\\\"center\\\"]
[/div]
There was also not a consistent increase in the amount of interceptions the opponents averaged per game as the intelligence value increased either.
[div align=\\\"center\\\"][/div]
[div align=\\\"center\\\"]Player Statistics[/div]
The only individual statistics that was consistently impacted by the increase in the intelligence value was the amount of tackles each cornerback averaged per game. If you exclude the 60 intelligence variable group because it was such an outlier the average tackles per game gradually increase. The 70 intelligence variable group averaged 1.535 tackles per game - 0.056 more tackles per game than the control group.
[div align=\\\"center\\\"][/div]
[div align=\\\"center\\\"]Endurance
Team Statistics [/div]
For the three cornerback archetypes, endurance starts at either 55 or 60. In general, increasing the cornerbacks’ endurance had a marginal impact on the Otters’ winning percentage.
[div align=\\\"center\\\"][/div]
The 65 endurance variable group was lower than the control by 0.25%, however these values are close enough to each other I think it is just an outlier in the sim. The other three variable groups tested increased the Otters’ winning percentage consistently as the endurance value increased.
[div align=\\\"center\\\"][/div]
Opponents consistently averaged less total yards per game, pass yards per game, and rush yards per game as the endurance value increased. The only exception to this is for the total yards and total pass yards at the 80 endurance variable group, which is most likely an outlier in this situation. Total yards decreased from 363.01 per game in the control group to 360.21 per game in at the lowest - in the 75 endurance variable group. This is a 2.8 yard per game decrease. Pass yards per game decreased from 256.89 per game in the control group to 254.17 at the lowest - in the 75 endurance variable group. This is a 2.72 yard per game decrease.
[div align=\\\"center\\\"]
[/div]
Opponent average interceptions per game were not substantially impacted by the increase in endurance.
[div align=\\\"center\\\"][/div]
[div align=\\\"center\\\"]Player Statistics [/div]
In general, no individual statistics for the cornerbacks consistently increased as the endurance value increased. There were variable groups for all statistics tracked that were higher than the control, however it did not follow any form of a pattern.
[div align=\\\"center\\\"][/div]
[div align=\\\"center\\\"]Speed
Team Statistics [/div]
For the three cornerback archetypes, speed starts at either 60, 65, or 70. In general, increasing the cornerbacks’ speed had a significant impact on the Otters’ winning percentage.
[div align=\\\"center\\\"][/div]
The control group had a winning percentage of 37.08%. All other variable groups had a winning percentage higher than this. As the speed value increased, the winning percentage increased in all groups except the 75 speed variable group which saw a minor dip in winning percentage from the 70 speed variable group. This is most likely an outlier. The highest speed variable group tested - 90 speed - had a winning percentage of 44.97%. This is 7.89% higher than the control group.
[div align=\\\"center\\\"][/div]
The 70 speed and 75 speed variable groups allowed a small amount more total yards per game to their opponents than the control group, however the increase in both was so small it is probably just an outlier or a bump in the sim. From the 80 speed variable group onwards the Otters’ opponents averaged less total yards per game than the control. At the highest speed variable group - 90 speed - the Otters allowed 10.11 total yards per game less than the control group.
[div align=\\\"center\\\"][/div]
Opponents passing yards per game followed a similar trend to total yards per game. The 75 speed variable group saw a small increase in the average pass yards per game surrendered to opponents, however it is probably an outlier. Discounting that grouping, as speed increased the average pass yards per game allowed to opponents decreased. The highest variable group - 90 speed - allowed 10.72 less pass yards per game to opponents than the control group.
[div align=\\\"center\\\"]
[/div]
Team total interceptions were not substantially impacted by speed.
[div align=\\\"center\\\"][/div]
[div align=\\\"center\\\"]Player Statistics [/div]
As speed increased the cornerbacks saw a fairly consistent increase in the average number of individual interception return yards per game, the average number of individual passes defended per game, and the average number of individual tackles per game.
[div align=\\\"center\\\"][/div]
Average individual passes defended increased from 0.939 per game in the control group to 0.968 per game in the 90 speed variable group - a 0.029 passes defended per game increase.
[div align=\\\"center\\\"][/div]
Average individual tackles per game increased from 1.479 per game in the control group to 2.373 per game at the highest point - in the 85 speed variable. This is a 0.894 tackle per game difference.
[div align=\\\"center\\\"][/div]
[div align=\\\"center\\\"]Strength
Team Statistics [/div]
For the three cornerback archetypes, strength starts at either 30, 35, or 40. In general, increasing the cornerbacks’ strength had a marginal impact on a team’s winning percentage, although it was not a consistent increase.
[div align=\\\"center\\\"][/div]
The control group had a winning percentage of 37.08% while the highest strength value tested - 50 strength - had a 39.67% winning percentage. This is a 2.59% difference. The 50 strength variable group did not have the highest winning percentage of the variable groups, however. The variable group that produced the highest winning percentage was the 40 strength group. The 40 strength variable group had a 39.78% winning percentage - 2.7% higher than the control group. The increase in winning percentage for each of the variable groups implies to me that there is some correlation between strength and winning percentage but it is not a consistent positive correlation.
[div align=\\\"center\\\"][/div]
Opponent total yards per game decreased fairly consistently as the strength value increased. The 50 strength group had allowed more total yards per game than the 45 strength group, however it was a slight amount and is most likely an outlier. Opponent pass yards per game also decreased fairly consistently as strength increased. The control group allowed 256.89 pass yards per game, while the 50 strength variable group allowed 248.52 pass yards per game - a 8.37 pass yard per game difference.
[div align=\\\"center\\\"]
[/div]
Interceptions were not significantly impacted by strength.
[div align=\\\"center\\\"][/div]
[div align=\\\"center\\\"]Player Statistics [/div]
The average number of individual passes defended and the average number of individual tackles increased consistently as strength increased.
[div align=\\\"center\\\"][/div]
The average number of individual passes defended increased from 0.939 per game in the control group to 0.966 per game in the 50 strength variable group. This is an increase of 0.027 individual passes defended per game.
[div align=\\\"center\\\"][/div]
The average number of individual tackles per game increased from 1.479 in the control group to 1.762 in the 50 strength variable group. This is an increase of 0.283 individual tackles per game.
[div align=\\\"center\\\"][/div]
[div align=\\\"center\\\"]Tackling
Team Statistics [/div]
For the three cornerback archetypes, tackling starts at either 40 or 45. In general, increasing the cornerbacks’ tackling had a marginal but not particularly substantive impact on a team’s winning percentage. Its impact was very similar to that of agility’s.
[div align=\\\"center\\\"][/div]
The control group had a winning percentage of 37.08 while the highest value tested - 55 tackling - had a winning percentage of 37.61. This is a 0.53% increase. The 55 tackling group did not generate the biggest increase in winning percentage over the control group, however. The 50 tackling group produced a winning percentage of 37.94%. This is a 0.86% difference from the control group. These values are so close together across the board that I would hypothesize that tackling has a very minimal impact on a team’s winning percentage.
[div align=\\\"center\\\"][/div]
Opponent total yards per game consistently declined as tackling increased, however the reductions were minimal. The control group allowed 363.01 total yards per game to opponents, while the 55 tackling variable group allowed 362.65 total yards per game to opponents. This is a fairly insubstantial difference of 0.36 total yards per game. Opponent pass yards per game did not consistently decrease, nor did opponent rush yards per game.
[div align=\\\"center\\\"]
[/div]
Interceptions were not particularly impacted by tackling.
[div align=\\\"center\\\"][/div]
[div align=\\\"center\\\"]Player Statistics [/div]
Increasing the tackling attribute caused a minor increase in the average number of individual tackles the players got per game. That was the only metric to consistently change at all.
[div align=\\\"center\\\"][/div]
The control group averaged 1.479 individual tackles per game. The 55 tackling variable group averaged 1.515 individual tackles per game. That is a 0.036 individual tackle per game difference.
[div align=\\\"center\\\"][/div]
[div align=\\\"center\\\"]Conclusions [/div]
1. Speed is in a category of its own when it comes to cornerback attributes.
2. Strength is the second best cornerback attribute, in my opinion. It increases a team’s winning percentage, reduces opponent yards fairly consistently, and impacts two individual statistics.
3. Endurance has an impact, however it seems to be more of a complimentary attribute. I need to test its value in conjunction with other attributes. This makes sense in theory, however, because without other statistics to use there is nothing to be prolonged by endurance.
4. Intelligence and tackling seem to the third tier of cornerback attributes - if you exclude endurance. They produce minor changes in win% as they increase, minor reductions in opponent total yards, and both impact an individual statistic.
5. Hands and Agility seem to be two of the less impactful cornerback attributes.
6. There is the potential that agility is like endurance and is a complimentary attribute, however on its own or early in a cornerback build it seems to have fairly little impact. When tested it caused a minor winning percentage increase but did not reduce an opponent’s offensive output and did not impact any individual statistics.
7. Hands also did not cause a significant impact in winning percentage. It caused a small decrease in an opponent’s offensive output, however that impact was smaller than that of intelligence, tackling, and endurance. It did not cause a consistent increase in individual cornerback attributes. I also have the theory that hands could be a complimentary attribute as well. Essentially, having good hands does you no good if you’re never in the play. If you have enough speed and intelligence to be in the play hands might have a greater impact.
8. This is not a foolproof study by any means. I’m still figuring out my methodology in a lot of cases and there is so much to explore that I’ll inevitably do more work on cornerbacks.
[div align=\\\"center\\\"]Notes[/div]
1. I’m finishing this up at 2am for me, so pardon any grammar mistakes or potential errors. I try to proofread these articles fairly well but I’m very drowsy, so I might have missed something.
2. This is probably the last sim study I will publish before the draft. I need to preserve some element of competitive advantage for myself and my future team. I also have a few private theories to test before I have anything else public. Rest assured, however, at some point I'll publish everything I do. I've gotta make that sweet media money.
3. As always, all of my work is open source. You can find the data here.