Creating Historical Players in NCAA 11

Printable View

Show 40 post(s) from this thread on one page

05-12-2010, 10:07 PM
JeffHCross

Creating Historical Players in NCAA 11

I want to use real statistics, and comparing them to current players, to create historical players in NCAA 11.

Now, obviously, this has a few limitations. You can only reasonably do this with recent years. Even Joe Montana, who is relatively recent, took a career to put up the numbers that Brady Quinn put up every season. However, you can possibly/reasonably work around this by using averages rather than totals. Still a limitation, but hopefully not a significant one.

What are you trying to do?
The primary purpose of my effort is to recreate the 1995 Ohio State Buckeyes. Expect to see a lot of Buckeyes used as examples to reinforce what I'm trying to do.

Now, a word about what I'm not trying to do: I am not trying to re-rate or recreate everyone. I'm more interested in integrating historical players into the current game. So, I'm trying to estimate how EA would be rating these historical players if they played today.

Using "Expected Performance"
To the best of my ability, I will be using a player's previous season performance or pre-season expectations for various ratings. For example, I won't be using Eddie George's 1995 Heisman performance as the barometer. Instead, I'll be using the fact that he was a 3rd-team preseason All-American. Postseason accolades can be considered for players that greatly exceeded their expectations. The reason for using previous season performance or pre-season accolades is to have a reasonable measurement for what the expectations were for a player heading into the season that's being recreated.

Phil Steele's PS# ratings will also be used for 'expected performance', if available or applicable. For example, a JR who was the PS #1 WR may use the OVR rating of Julio Jones or A.J. Greene when they're JRs in NCAA 11.

So what are you actually using for a rating?
For every rating I list in the future in this thread, I take a statistic or measure, whenever possible, and compare that to the performance of current players. Then, when NCAA 11 comes out, I'll take that player's rating, and plug it in for my created player.

Going back to the Eddie George example, this year's Phil Steele preseason 3rd team running backs were Noel Devine and C.J. Spiller. George's 'expected performance' would probably be used for OVR. So if I was creating Eddie George on NCAA 10, I'd attempt to match his OVR to either Noel Devine or C.J. Spiller's.

Sppppppeeeeeeeeeeeedddddd
Whenever possible, I will use legitimate 40 times. Often, these will be from the NFL combine. If the player in question is not a senior, then I'll be using an estimate of their dynasty offseason progression, so they have a reasonable chance of being equal to their actual 40 time by the proper year.

Where do you come in?
I hope that someone reading this will have suggestions for improvements that could be made to my technique. And maybe other people will want to use this to create their own teams.
05-13-2010, 06:06 AM
texacotea

So when creating these players all you are using is the OVR of the player that matches the one you are creating? How do you determine strength, speed, agility etc. for George? The 2 backs that you referenced are totally different backs than George was
05-13-2010, 11:10 PM
JeffHCross

I'll go into specifics as I get into it. I've got a decent base to start with (of attributes to compare) for now, but I don't want to post them up until they're slightly more final. I posted them on Utopia when I originally started this effort, but then found myself modifying them almost as quickly as I was posting them.

The quickest explanation is that I'll be taking measures that I believe correspond to the individual rating. For example, using Completion % for THA. I've also developed The Hayes Rating for other reasons, but I'll be using that for AWR. I'll find a comparable current player using the appropriate measure, then use that players rating for my historical player.

So, if I determined (through whatever means) that Eddie George had similar strength (and therefore STR) to Mark Ingram (just to pick a name), I would then put in Mark Ingram's NCAA 11 STR rating for Eddie George's. Does that make sense?

For SPD, it's all coming from my research on 40 yard dash times and their correspondence with SPD. I've developed an Excel sheet that will turn a 40 time into both SPD and ACC, then made a mathematical formula to convert SPD, ACC, height and weight to AGI. It actually works surprisingly well.

Again though, I'll detail this all when I get a chance. It's still very fluid right now.
06-23-2010, 11:09 PM
JeffHCross

I've learned to tease thanks to JB, so ...

Actually, just wanted to say "be on the lookout" for a big update to this coming soon. Made some major progress on this last weekend, and finished even more players tonight. With Teambuilder for '11 being released, progress should come really quickly now.
07-10-2010, 10:43 PM
JeffHCross

Alright .... let's do this.

For each rating listed below, I determined a metric to find a similar current player. I was not creating numbers (except where otherwise noted) using the statistics, but instead using the numbers to find a similar current player. I would then use that similar player's NCAA 11 rating for my player's rating.

A few constants to discuss first:
SPD: Uses the 40 times -> SPD conversions discussed in this thread.
ACC: Uses horribly (overly) complex calculations to compare a player's 40 time and the "average" 40 time for a player with that SPD rating. Then uses the "average" ACC rating for that SPD and modifies appropriately. In simple terms ... let's say an average 92 SPD, 92 ACC receiver (92 ACC is the average for 92 SPD) ran a 4.435 40-yd dash. If I had a player that ran a 4.40, and he had 92 SPD, he'd have slightly faster than 92 ACC. If my player ran a 4.45, with 92 SPD, he'd have slightly slower than 92 ACC. Excel made this easy, thankfully.
AGI: (0.57*SPD)+(0.36*ACC)-(0.08*Height in inches)-(0.01*Weight)+12.46. Came in pretty handy. Numbers are based on a regression of the entire NCAA 10 roster.
RET: Kickoff Return average. For anybody that wasn't a regular KR for Ohio State, I gave them 60 in RET.

Stamina: For skill positions, found players with similar amounts of plays per game. Number of touches, basically.
Injury: Looked for players with similar number of games played, games started, etc. Only really became a factor with players that had severe injuries.
07-10-2010, 11:01 PM
JeffHCross

Quarterback
OVR: Expected Performance
AWR: Hayes Rating, adjusted for year
Strength, THP: Yards Per Completion
Break Tackle, Trucking, Ball-Carrier Vision: Rushing Yards Per Carry
Elusiveness: Sack % (Sacks per Attempts)
Carrying: One of the harder stats to find ... fumble %. # of fumbles per rushing attempt.
THA: Completion Percentage

Example: Bobby Hoying
Hoying was the senior leader at quarterback for the Buckeyes in 1995. In the 1994 season, he threw for 2335 yards on 170 completions (56.5%) with 19 TDs and 14 INTs. He was also sacked 24 times.

OVR: First Team Big Ten, but not an All-American. This put him between Terrelle Pryor (1st time B10, but also an AA according to many) and 2nd Team B10, Ricky Stanzi. I have him at 92 OVR.
AWR: Near Taylor Potts and Andrew Luck. 84.
Strength, THP: Between Jevan Sneed and Kevin Riley. Ended up at 59 STR and 88 THP.
Elusiveness: Kevin Riley again. Hoying couldn't get away from much, so his 42 seems appropriate.
THA: Funny enough, he's near Terrelle Pryor. But I thought Ricky Stanzi, who threw 56.3%, was a more appropriate comparison. Ended up at 88 THA as well.
07-10-2010, 11:17 PM
JeffHCross

Running Backs
OVR: Expected Performance
AWR: Similar OVR player in the same year
Strength, Break Tackle, Trucking, Elusiveness, Stiff Arm, Spin, Juke: Player with similar Yards Per Carry and similar Running Style
Ballcarrier Vision: Yards per Carry, regardless of running style
Carrying: Fumble %
Catching, Spec. Catch, Catch in Traffic, etc: Player with similar receptions per game and receiving totals
Blocking Stats: Similar build player

Example: Eddie George
Eddie George, is, of course, the 1995 Heisman winner. However, for the purposes of this, I was using his 1994 campaign, which included 1,442 yards and 5.5 YPC.

OVR: 3rd team All-American and Heisman candidate. Similar players included Evan Royster and Ryan Williams. I aimed for a 93 OVR.
AWR: Used Royster, I believe. 90 AWR.
Break Tackle, Trucking, Elusiveness, Stiff Arm, Spin, Juke: I believe these came from John Clay. 90 BTK and 88 TRK, for example. TRK may be a little high.
Ballcarrier Vision: Not sure who. Dion Lewis had a similar YPC, but his 94 BCV is high. Ended up at 88 BCV.
Carrying: 90.
Catching, Spec. Catch, Catch in Traffic, etc: Evan Royster
07-10-2010, 11:28 PM
JeffHCross

Receivers & Tight Ends
OVR: Expected Performance
AWR: Similar OVR player in the same year
Strength, Break Tackle, Trucking, Elusiveness, Stiff Arm, Spin, Juke: Player with similar Yards Per Reception and similarly built/used (Speed, Possession, Slot, etc)
Ballcarrier Vision: Yards per Play. Total Offense per reception+carries
Carrying: Yards Per Reception
Catching, Catch in Traffic: Receptions Per Game
Spectacular Catch, Release, Jumping: Stylistic match. i.e. same type of receiver.
Blocking Stats: Similar build player

No example for receiver. A lot of this was done on the fly because receivers were very difficult to find similar players for. Plus, most of the Ohio State players in 1995 were almost complete unknowns (Terry Glenn had 15 career catches entering the season ... he had 9 against Pitt alone in 1995), so it was mostly a guessing game using Teambuilder.
07-10-2010, 11:51 PM
JeffHCross

Offensive Line
In a way, these were the most fun to do. But they're also going to be the hardest to explain, so bear with me.

For each player, using Expected Performance I had an OVR rating I was aiming for. For linemen this became even more important than at other positions.

In the late 1980s, a book called The Hidden Game of Football introduced this concept of Line Stats, also known as Line Yards. In general, the idea is that if you know a team's Rush Per Carry average, you generally know how well they run block. If you know their Sack Percentage, you generally know how well they pass block. If you take both of those rankings and average them together based on the number of rushing plays versus passing plays by that team for a season, you know how well they block ... period.

So that's where I went with this.

For Ohio State in 1994, I found that they were most similar, overall, to the Offensive Lines for UNC and Virginia Tech.
For rushing only, they ranked closely to Oklahoma State.
For passing only, they ranked closely to Wisconsin.

Not exactly strong comparisons in any of the three areas, in my opinion. But it wasn't a great offensive line ... outside of Orlando Pace.

So here's what I did. For Awareness, I summed up the AWR ratings of the OL on Virginia Tech and on North Carolina. I took the RBK ratings for Okie State and the PBK ratings for Wisconsin. I then summed the OVR ratings for all of my Ohio State linemen, found the individual percentage of that, and used that percentage to multiply by the sum for the three ratings I just mentioned.

This might make more sense:
Strength, Stamina, Injury: Player with close OVR in same year. Often the Expected Performance player.
Awareness: North Carolina / Virginia Tech
Pass Block: Wisconsin
Run Block: Oklahoma State

Each of the "next-gen" blocking ratings were calculated using a formula based on other ratings.
Impact Block: (.33 * Run Block) + (.27 * Pass Block) + (.17*Strength) + (.06*Acceleration) + (.07*SPD) + 9.03
Run Block Strength: (.4 * Run Block) + (.5 * Strength) + (.05*Acceleration) + (.16*SPD) - 6.38
Run Block Footwork: (.34 * Run Block) + (.27 * Strength) + (.1*Acceleration) + (.07*SPD) + 18.18
Pass Block Strength: (.53 * Pass Block) + (.41 * Strength) + (.05*Acceleration) + (.1*SPD) - 7.47
Pass Block Footwork: (.5 * Pass Block) + (.2 * Strength) + (.11*Acceleration) + (.02*SPD) + 12.26

Example: Orlando Pace
The star of the 1995 Ohio State line. Also, the reason the Pancake stat exists.
OVR: 3rd team All-American. #1 Offensive Linemen in the 1994 Recruiting Class. Compared him with Marcus Cannon of TCU, Clint Boling of Georgia, and Xavier Nixon of Florida. Aimed for 93 OVR.
Strength, Stamina, Injury: 91 STR, 77 STA.
Awareness: Pace's 93 OVR was 11.5% of Ohio State's OVR. 11.5% of 706 is 81.
Pass Block: 11.5% of Wisconsin's 826 PBK. 95.
Run Block: 11.5% of Okie State's 855 RBK. 98.

I would later adjust the PBK and RBK ratings because the OVRs became very inflated for the "good" linemen on my team. I usually took them down 2 or 3 points in each rating, until their OVR was about right.

After adjustment, his PBK was 88 and his RBK was 91.

Impact Block: 88
Run Block Strength: 91
Run Block Footwork: 86
Pass Block Strength: 88
Pass Block Footwork: 84
07-10-2010, 11:59 PM
JeffHCross

Defense
By far, the easiest. ("Same Position" means I'll find the closest player at the same position on the field. Not using Safeties ratings for LBs)

OVR: Expected Performance
Awareness, Press Coverage: Player with similar OVR and experience. Often same as Expected Performance player.
Strength: Player with Similar OVR, and similar build (ht, wt) at same position
Catching, Catch in Traffic: # of Interceptions
Jumping: Passes Defended / Broken Up
Tackling, Block Shedding: Solo Tackles, same position
Power Move, Finesse Move: Sacks, same position. Nearest player with similar Ht/Wt/Build.
Pursuit: Assisted Tackles, same position
Play Recognition: Tackles for Loss for Front 7, Total Tackles for DBs
Man Coverage, Zone Coverage: Passes Defended / Broken Up, with a twist. Say Shawn Springs compares with CB X from Florida. CB X has a 88 MCV and a 93 ZCV. Well, Shawn Springs was known as a Man Coverage expert. So that doesn't work. Instead, I'll take the 88 MCV and 93 ZCV and add them together, similarly to what I did for the linemen. Now I have 181 points. Then I'll look around for corners that have exceptional MCV ratings and total 181 MCV+ZCV. And I'll use that. Or at least that's the plan. Work-In-Progress right now.

Don't have any examples right now. Going to begin the defensive players ratings tomorrow.
07-11-2010, 12:04 AM
JeffHCross

Alright, that's it ... over 7 months of effort summed up in 5 short posts.

In a lot of ways that's horribly, horribly, sad. But at least it's done. I hope someone else finds it useful, or it spawns your own creative juices. Feel free to ask questions or offer suggestions.
07-11-2010, 06:31 AM
psusnoop

Really great work Jeff, your effort and hard work might have just been enough for me to make the 1994 Penn State Team.

1 other note, you really used Evan Royster alot for reference to Eddie George and I found that funny because I think Evan Royster is the most overratted RB in this years class. LOL
07-11-2010, 10:40 AM
JeffHCross

Quote:

Originally Posted by psusnoop

1 other note, you really used Evan Royster alot for reference to Eddie George and I found that funny because I think Evan Royster is the most overratted RB in this years class. LOL

Well, I didn't use him for anything that actually dealt with running ;)
07-12-2010, 12:15 AM
ebin

Great information, thanks for sharing this!
07-12-2010, 09:02 PM
thetwonewguru

Up until the past couple of years (2008) I was a firm believer that the game as far as players skill levels had not changed much. Even though the so called experts hammered this idea into my head every Saturday, I refused to believe this. Just as a note of reference I'm closing in on the ripe age of 30 in September. But in the 07,08 seasons I saw a drastic change of skill level occur on the top tirer programs. I did admit to myself in the mid 2000's that the playing field was leveling out. There were more talented players on each team mainly the Mid level programs. But the talent in say the SEC (my demographic) was very leveled out even Vandy. The lower teams in the SEC just didn't have the depth as the upper echelon.

I think the top teams from 1992 ro 2006 or so could have been competive in head to head action, sans the 95 Nebraska team. I think that team with the same scheme could have beaten the National Champ any year since. I don't care what they say about the 05 Texas team or the great teams USC fielded. To me, this team is the greatest team ever, hands down.

With all this said, I admire the system you have came up with compiling the classic teams. But Idon't think you can compare players of today with players back then. Beleive you me, thias is very hard for me to say, because I grew up in the 80's and 90's. This era has a lot of nostalgia for me. I have tons of games (specifically Alabama) that I rewatch on a yearly basis. I enjoy that era's style far more than this era's. But I think the top teams of today would wax the floor with the teams of that Era. Again, hard for me to admit.

I have created alot of classic teams. My mode of evaluation is through complete game watching. I size them up, and try to rate them as best as I can. The Dline and linebackers will not be as fast on an average like you use. My method is not very scientific. And in no way do I think I could be a professional scout by today's standard but I trust my eye from my era's standard. Sure, there are special cases where a lineman gets a high speed rating from that era but in no way does it compare to today's standard.
If you'd like to check out my rosters my gamertag is tcoaker on teambuilder. I too would love feedback from you. Since you do seem to share the same passion I do from this era. My teams are not heavy on the 90+ guys
07-12-2010, 10:18 PM
JeffHCross

thetwonewguru shared his thoughts with me over PM before posting them here, so I won't go in-depth in my response here. To put it simply, I agree with him completely. As much as I'd like to think otherwise, the mid-90s Buckeyes couldn't stay on the field with today's. That's probably even true, though lesser so, for the 2002 National Champions. The game is constantly evolving, especially in speed.

To quickly summarize something I noticed in the course of this research ... Bobby Hoying (4.67), Eddie George (4.47), and Terry Glenn (4.29) would have all been at, or among, the top players at their position in this year's NFL Combine. However, on the defensive side of the ball, Mike Vrabel and Matt Finkes, possibly two of the best defensive linemen to ever play for Ohio State (Finkes had 4 blocked kicks in 1994, to give you an idea of how quick he must have been). According to the 40 times I found (which I think may have been a little slow), they not only would have been the two slowest Defensive Ends at the Combine, they would have just barely beaten Ndamukong Suh. Pat Fitzgerald and Pat Tilman, two other mid-90s stars, would have been near the bottom at their positions in this year's combine.

So, it seems, that while the offenses have not gotten significantly faster, the defensive side of the ball has changed completely.

It's also worth noting that while I do have higher speeds than the "norm" for Offensive Linemen on NCAA 11, I'm also using equations based on 40 times for all linemen ... offensive linemen in NCAA 11 are generally much slower than defensive linemen ... and I chose not to go that route.

Thanks for the feedback, thetwonewguru and ebin.
07-21-2010, 08:28 PM
JeffHCross

One more example to share, to finish this off.

Example: Shawn Springs
A sophomore in 1995, Shawn Springs came off a 1994 campaign that earned him All-Big Ten Honorable Mention and Freshman All-American honors. Phil Steele named Springs as All-Big Ten Preseason 2nd team in 1995.

OVR: Shaun Prater of Iowa and Devin Smith of Wisconsin are 2nd team CBs this year. I'm aiming for a similar rating to Prater's 87.
Awareness, Press Coverage: Prater has 94 AWR, Smith has 81. I split the difference and went with 87. I used Prater's 82 for Press.
Strength: 50 from Prater.
Catching, Catch in Traffic: I used Springs' 1995 INT stats, which compared to Prince Amukamara of Nebraska for 2009. So I used the NCAA 10 ratings of 62 & 54, which are slightly lower than what Prince has this year.
Jumping: Comparable to Donovan Warren of Michigan, 92
Tackling, Block Shedding: 42 solo tackles compares well to Tyler Sash of Iowa and Prince, again. Sash is actually a safety in the game, so that won't work. Prince's 63 and 61, respectively, fit well here.
Power Move, Finesse Move: Springs compares favorably to Ras-i Dowling of Virginia. 25 and 25, ew.
Pursuit: Again compares favorably to Dowling, who has a 89 rating.
Play Recognition: Back to Prince. 94!!!
Man Coverage, Zone Coverage: Springs's numbers are similar to Michigan's CB Donovan Warren, as much as I hate to admit that possibility. Donovan's total of 176 worked well here, and I gave Springs a 92 MCV and 84 ZCV.

Almost done with my 1995 team ... ridiculously close. I'll probably be posting it tomorrow.
07-21-2010, 08:33 PM
JeffHCross

Just as a real quick aside, for an example of how overboard I'm doing this, for the 30 players on my defense, by the end, I will have looked up somewhere around 540 ratings, using about 150 different players.
07-22-2010, 08:34 PM
JeffHCross

I. Am. Done.

FINALLY!

Here's the team: http://www.easports.com/teambuilder#...665:1244367337

Here's a table comparing the overalls of the top two players at (almost) every position, for the 2010 Buckeyes and the 1995 Buckeyes. Just to give you an idea.

Anybody who gives them a try and provides feedback ... it would be tremendously appreciated.
07-22-2010, 10:39 PM
Oneback

Team Stats

Code:

1995 Ohio State Michigan 17 Score 21 17 First Downs 17 336 Total Offense 314 29-160-1 Rushes - Yards - TD 22-28-1 15-22-0 Comp - Att - TD 22-33-2 176 Passing Yards 286 5-11 (45%) 3rd Down Conv 5-11 (45%) 4-1-1 (50%) Red Zone - TD - FG 3-3-0 (100%) 2 Turnovers 2 1 Fumbles Lost 0 1 Intercepted 2 3 - 42.7 Punts - Average 5 - 43.0 4 - 30 Penalties - Yards 1 - 15 18:50 T.O.P. 19:28

Scoring

Code:

1st 2nd 3rd 4th OT OSU 7 0 10 0 0 17 Mich 0 14 0 7 0 21

First Quarter
3:54 TD OSU: R. Miller, returned interception 59 yards

Second Quarter
6:39 TD Mich: K. Koger, 21 Yd pass from T. Forcier
0:11 TD Mich: J. Hemingway, 15 yard pass from T. Forcier

Third Quarter
5:40 FG OSU: J. Jackson 25 yd FG
2:40 TD OSU: E. George 6 yard run.

Fourth Quarter
0:10 TD Mich: T Forcier, 5 yard run

Player Stats

Code:

Passing: Comp Att Yards TD INT Sack Tate Forcier Mich 21 32 264 2 2 2 Bobby Hoying OSU 15 22 173 0 1 0 Rushing: ATT Yards AVG TD LONG BTK Vincent Smith Mich 13 28 2.1 0 12 0 Eddie George OSU 19 105 5.5 1 22 9 Pepe Pearson OSU 6 46 7.6 0 15 2 Receiving: REC Yards AVG TD LONG DROPS J Hemingway Mich 6 45 7.5 1 16 0 M. Odoms Mich 5 50 10.0 0 15 0 Kevin Koger Mich 4 85 21.2 1 26 2 Eddie George OSU 5 59 11.8 0 19 0 Rickey Dudley OSU 4 66 16.5 0 20 0 D. Stanley OSU 2 27 13.5 0 23 0

In the end it was exactly what I expected. A strong running team with a good defense. Eddie George is a beat to bring down and it too everything I could to keep the running game in check (if you can call giving up 160 yards in check).

I like what you've created. Great Job.

Show 40 post(s) from this thread on one page