Football Forum

View the Forum Registry


Self-edit this thread


0 Subject: A new statistical look at TNSP production II

Posted by: dgrooves
- [454482113] Fri, Nov 08, 2002, 11:16

I got this idea from a thread (A new statistical look at TNSP production ) Mighty Geeks started last season. From that thread: " I have attempted to analyze the statistical performance of opposing defensive teams and apply that analysis to the average scores of several of the players I am looking at adding to my team. Like any mathematical approach to the game, the flaw is that nothing that ever happens is "average" on any given week. Nonetheless, I think it is useful to look at opponents when selecting players to add to your teams.

The actual method I used is fairly easy to understand, but difficult to explain, so please try to bear with me. First, I used a statistical source to capture the following stats for each defense, passing yards allowed, passing TDs, rushing yards allowed, and rushing TDs. I also computed NFL averages for these same stats. I could do the same for WR and TE yards and TDs, but I'm not yet sure that it is worth the effort. Then I took these average numbers and figured out how many TNSPs that would be worth per game. For instance, the average Defense gives up 232 passing yards and 1.4 passing TDs. This means an average net of 316 TNSPs. This formula does not take into account fumbles or interceptions. I think I like that, but I'm not sure yet. I then look at a particular Defense, such as Arizona and determine that they give up an average of 268 yards and 1.8 TDs. This results in an average TNSP of 376. I then convert this into a ratio of 1.19 and compute that any QB lucky enough to face the Arizona D will get 1.19 points for every point he usually gets. I then multiply the average TNSP/game for each QB by this factor and arrive at a predicted TNSP for each QB.

The next thing I do is perform the exact same analysis for the next three weeks to find players who have a nice cushy schedule coming up.

As a number cruncher from way back, I can assure you that these results will not come to pass, but they are interesting to look at sometimes. The more interesting stat for me is the three week outlook and I will try to paste my table for that as a reply. It might come out in HTML or it might not. I'll see what I can do."

I dont have access to the same data that he did, so I have instead used the TSNP allowed data that Goatlocker and RSF have been posting. Also, I have weighted all the players'/defenses' TSNPs scored and allowed. This adjusted average is 70% points per game over the player's/defense's last 4 games and 30% the player's/defense's overall average.

I have very little background in statistics so I am certainly open to suggestions and criticisms. Also, I want to reiterate that these numbers may seem like predictions, but it is highly unlikely that they will be correct in that sense. For example, last week's numbers had Travis Henry ahead of Mashall Faulk, but Faulk scored 542 and Henry only scored 136.






Week 10

Jeff Garcia

586

Brian Griese

448

Daunte Culpepper

371

Steve Mcnair

347

Kerry Collins

344

Marc Bulger

339

Aaron Brooks

337

Tom Brady

324

Trent Green

319

David Carr

310

Brett Favre

302

Donovan McNabb

266

Chad Pennington

261

Rich Gannon

235

Drew Brees

222

Kurt Warner

213

Jeff Blake

213

Jake Plummer

212

Jay Fiedler

209

Chris Redman

208

Marshall Faulk

534

Priest Holmes

515

Ahman Green

347

LaDainian Tomlinson

322

Clinton Portis

297

Charlie Garner

267

Corey Dillon

266

Eddie George

261

Curtis Martin

257

Tiki Barber

251

Duce Staley

251

Ricky Williams

246

Shaun Alexander

244

Antowain Smith

240

James Stewart

232

Marcel Shipp

222

Michael Bennett

220

Deuce McAllister

212

Kenny Watson

205

Stephen Davis

204

Terrell Owens

419

Travis Taylor

239

Derrick Mason

238

Jimmy Smith

220

Tai Streets

218

Santana Moss

212

Curtis Conway

201

Ike Hilliard

201

Koren Robinson

197

David Patten

194

Hines Ward

193

Michael Lewis

192

Troy Brown

190

Joe Horn

188

Dante Hall

188

Ed Mccaffrey

187

Rod Smith

183

Marvin Harrison

180

Bobby Engram

180

Plaxico Burress

178

Philadelphia D

321

GreenBay D

263

Arizona D

220

NewOrleans D

186

NewEngland D

182

NYJets D

181

Tennessee D

171

Jacksonville D

168

St.Louis D

150

Denver D

138

Washington D

131

Baltimore D

114

Cincinnati D

100

NYGiants D

95

Seattle D

92

Minnesota D

66

Pittsburgh D

56

Chicago D

54

Indianapolis D

50

Miami D

44







Week 11

Jeff Garcia

571

Drew Bledsoe

482

Brett Favre

399

Tom Brady

363

Donovan McNabb

355

Michael Vick

351

Rich Gannon

340

Trent Green

323

Brian Griese

314

Chad Pennington

309

Tommy Maddox

302

Aaron Brooks

289

Drew Brees

277

Marc Bulger

275

Shane Matthews

255

Brad Johnson

250

Mark Brunell

248

Kelly Holcomb

247

David Carr

245

Kerry Collins

218

Priest Holmes

588

Marshall Faulk

503

Corey Dillon

357

LaDainian Tomlinson

353

Clinton Portis

332

Deuce McAllister

306

Ahman Green

295

James Stewart

294

Charlie Garner

294

Travis Henry

277

Shaun Alexander

274

Fred Taylor

264

Tiki Barber

248

Antowain Smith

245

Duce Staley

244

Curtis Martin

238

Jamal Lewis

233

Garrison Hearst

219

Ricky Williams

216

Kevan Barlow

207

Peerless Price

327

Terrell Owens

322

Eric Moulds

297

Dennis Northcutt

270

Curtis Conway

261

Jimmy Smith

253

Marvin Harrison

231

Hines Ward

226

Santana Moss

222

Quincy Morgan

219

David Patten

219

Troy Brown

214

Donald Driver

214

Plaxico Burress

209

Derrick Mason

206

Michael Lewis

205

Joe Horn

200

Marty Booker

197

Bobby Shaw

196

Chad Johnson

194

TampaBay D

491

GreenBay D

340

Miami D

281

St.Louis D

268

Dallas D

265

Philadelphia D

254

Baltimore D

202

Jacksonville D

180

Denver D

178

Carolina D

160

Pittsburgh D

160

NYGiants D

150

Washington D

136

Indianapolis D

133

Cleveland D

124

Arizona D

118

NewEngland D

108

SanFrancisco D

99

NYJets D

85

Oakland D

75







Week 12

Donovan McNabb

403

Tom Brady

387

Rich Gannon

337

Drew Bledsoe

334

Chad Pennington

326

David Carr

319

Kelly Holcomb

304

Jake Plummer

303

Aaron Brooks

298

Jay Fiedler

294

Kerry Collins

293

Jeff Blake

273

Jeff Garcia

269

Michael Vick

267

Trent Green

267

Chris Redman

266

Tim Couch

265

Daunte Culpepper

264

Brian Griese

248

Tommy Maddox

237

Priest Holmes

601

Deuce McAllister

457

Ahman Green

415

Marshall Faulk

373

Tiki Barber

347

LaDainian Tomlinson

344

James Stewart

315

Curtis Martin

301

Michael Bennett

289

Ricky Williams

280

Duce Staley

278

Travis Henry

274

Shaun Alexander

268

Clinton Portis

257

Charlie Garner

237

Edgerrin James

230

Moe Williams

215

Eddie George

206

Kenny Watson

201

Stephen Davis

200

Hines Ward

293

Michael Lewis

286

Joe Horn

280

Plaxico Burress

270

Peerless Price

255

David Patten

248

Troy Brown

243

Jerry Rice

234

Eric Moulds

232

Koren Robinson

229

Terrell Owens

218

Santana Moss

214

Bobby Engram

209

Dennis Northcutt

202

Ike Hilliard

201

Antwaan Randle-El

196

Marvin Harrison

192

Tim Brown

191

Derrick Mason

190

Jerry Porter

187

Atlanta D

395

GreenBay D

301

Detroit D

268

Pittsburgh D

248

St.Louis D

226

Jacksonville D

216

Dallas D

209

Denver D

206

Tennessee D

177

Arizona D

171

Miami D

169

NYGiants D

161

NewEngland D

97

Washington D

95

NYJets D

87

Oakland D

85

Baltimore D

73

SanFrancisco D

73

KansasCity D

72

NewOrleans D

70







Next 3 Weeks

Jeff Garcia

1426475

Tom Brady

1074358

Donovan McNabb

1024341

Brian Griese

1010336

Aaron Brooks

924308

Rich Gannon

912304

Trent Green

909303

Chad Pennington

896298

David Carr

874291

Kerry Collins

855285

Marc Bulger

822274

Drew Bledsoe

816408

Michael Vick

808269

Daunte Culpepper

808269

Brett Favre

777259

Steve Mcnair

749249

Tommy Maddox

729243

Drew Brees

723241

Jake Plummer

672224

Jeff Blake

669223

Priest Holmes

1704568

Marshall Faulk

1410470

Ahman Green

1057352

LaDainian Tomlinson

1019339

Deuce McAllister

975325

Clinton Portis

886295

Tiki Barber

846282

James Stewart

841280

Corey Dillon

811270

Charlie Garner

798266

Curtis Martin

796265

Shaun Alexander

786262

Duce Staley

773257

Ricky Williams

742247

Michael Bennett

710236

Antowain Smith

665221

Fred Taylor

641213

Eddie George

612204

Jamal Lewis

601200

Kenny Watson

589196

Terrell Owens

959319

Hines Ward

712237

Michael Lewis

683227

Joe Horn

668222

David Patten

661220

Plaxico Burress

657219

Jimmy Smith

655218

Santana Moss

648216

Troy Brown

647215

Curtis Conway

643214

Derrick Mason

634211

Marvin Harrison

603201

Peerless Price

582291

Travis Taylor

580193

Ike Hilliard

576192

Koren Robinson

553184

Marty Booker

539179

Jerry Rice

537179

Chad Johnson

532177

Eric Moulds

529264

GreenBay D

904301

St.Louis D

644214

Jacksonville D

564188

Philadelphia D

522174

Denver D

522174

Arizona D

509169

Miami D

494164

TampaBay D

491245

Dallas D

474237

Pittsburgh D

464154

NYGiants D

406135

Baltimore D

389129

NewEngland D

387129

Washington D

362120

Tennessee D

355118

NYJets D

353117

Atlanta D

349116

Detroit D

307102

NewOrleans D

27792

Indianapolis D

21672
1Sludge
      Sustainer
      ID: 566332517
      Fri, Nov 08, 2002, 12:08
A good start, but it does not take into account the quality of the past defenses an offensive player has gone up against. For example, if a QB has played soft defenses all year, then his average TSNP per game is an inflated estimate of what his average performance would be versus the "average" defense.

An alternative approach: For each offensive player, build a regression model with TSNP as the dependent variable and the defensive statistics as the independent variables (i.e. Passing yards allowed, passing TDs, rushing yards allowed, and rushing TDs. This also allows you to throw in things like interceptions, fumble recoveries, sacks, home/away, etc. Have to be careful in your selection of the independent variables not to have too many.)

Drawback? Pain-in-the-ass because you have to do this for each offensive player.
2walk
      Leader
      ID: 338441813
      Fri, Nov 08, 2002, 12:52
Wow, made my selections prior to seeing this thread, and had a high degree of correspondence:

Garcia, Griese, Faulk, Holmes Owens, and Green Bay are all on my team. Now let's see if past performance is a predictor of future performance (it sure is when it comes to human behavior!).

Clearly, a team with some offensive stars playing KC makes it to the top...

- (dr.) walk
3dgrooves
      ID: 23105812
      Fri, Nov 08, 2002, 13:29
Sludge:
I am planning to adjust the players' averages based on past opponents, but I just havent had enough time to do so. Hopefully I will have it done by the time next week rolls around. I was thinking of calculationg the average of the multipliers of the opponents the player has played, then dividing the player's weighted average by it. For example, say a QB has played teams with multipliers of 1.06, 1.2, and .86 and is averaging 306 TSNPs per game. The average of these is 1.04, meaning this QB has, on average, played easy defenses, compared to the league average. The player's TSNPs per game should thus be decreased by .961, the inverse of 1.04. So instead of averaging 306 TSNPs, he would average 294 TSNPs. Is this the best way to account for schedule strength, so to speak?

I am unable to do the regression analysis because I am not using the independent variables you mention. I am using the data GoatLocker and RSF are providing.
4Sludge
      Sustainer
      ID: 566332517
      Fri, Nov 08, 2002, 14:34
dgrooves -

Is the method you mentioned the best way to account for the strength of the defenses a player has gone against? Probably not. It would be very difficult to impossible to know what the best way of doing it is. Only skimming over what you proposed, I would say that it's an improvement, though.

As to your statement that you can't use the variables I mentioned: Yes you can, and yes you are. In fact, I copied and pasted the four variables from your initial message. I simply state that you can also use the additional ones I mentioned.

In fact, a better way to do the regression would be to use TSNP as the dependent variable and both offensive statistics and defensive statistics for the independent variables. In this case, a different model would have to be built for each position, which is not unreasonable. At the very least it's much more reasonable than a different model for each individual player. The first type I mentioned suffers from a lack of data, which gives it less predictive power. The second type I just mentioned won't suffer from a lack of data, but it won't take into account an individual's ability to excel against a particular type of defense. Damned if you do, damned if you don't.
5dgrooves
      ID: 454482113
      Fri, Nov 08, 2002, 15:33
Sludge:

No, I cant use those variables. Please re-read my initial message. What you think I wrote was actually written by Mighty Geeks, as I say in that post. He was using the yardage, interception, etc numbers when he did this analysis last year. I dont have access to that data; I am instead using the TSNP allowed data from GL and RSF.
6Kyle
      ID: 381043820
      Fri, Nov 08, 2002, 22:13
I am just wondering why Farve's numbers are down these next three weeks? Ahman will make his numbers a little lower but Ahman catches the majority of Farve's completions
7root88
      ID: 359101014
      Sat, Nov 09, 2002, 02:50
I used Mighty Geeks info to help me out last year and it worked very well for me. Of course it's not a predection, but it can open your eyes to something you might miss schedulewise, or if you are stuck between two players this info could push you over the edge.

Thanks dgrooves.
8dgrooves
      ID: 454482113
      Sat, Nov 09, 2002, 10:56
Kyle, I dont see any problems with Farve's numbers. He is "predicted" to score more points than his average in Weeks 10 and 11 (Det and Min). In Week 12 he is playing the best defense in football, Tampa Bay. This means he will likely score less points than he is averaging.

Also keep in mind that I am using weighted average TSNPs per game. Farve is averaging 299 points per game during the entire season, but only 276 points over his last for games.
9Bandos
      ID: 39112921
      Sat, Nov 09, 2002, 21:53
Very fun to look at. Helpful too!
Thanks.

Just a few interesting observations...
Why am I holding Porter when M. Lewis is out there?

Carr and Fatpepper with better week 10 games than McNabb? wow!

Where is Driver?

Guess you gotta have Faulk and Holmes - the leaders won't (hall and Battery judging by their trade/bye situation)

Ward, Ward, Ward....is Tommy Maddox that good? Or is Ward.

How about Moulds andice next week. How about any WR vs. KAN every week. Holy crap.

Thanks again for the brain workout.
10Kyle
      ID: 239312618
      Sun, Nov 10, 2002, 21:13
Thanks for explanation D-grooves. I know who to start week 12 between Farve and Greise. I just don't know why farve wouldn't even crack the list that week. Then again u are the statistion (i can't spell).
Rate this thread:
5 (top notch)
4 (even better)
3 (good stuff)
2 (lightweight)
1 (no value)
If you wish, you may rate this thread on scale of 1-5. Ratings should indicate how valuable or interesting you believe this thread would be to other users of this forum. A '5' means that this thread is a 'must read'. A '1' means that this is a complete waste of time.

If you have previously rated this thread, rating it again will delete your previous rating.

If you do not want to rate this thread, but want to see how others have rated it, then click the button without entering a rating, or else click here.

Football Forum

View the Forum Registry


Self-edit this thread




Post a reply to this message: (But first, how about checking out this sponsor?)

Name:
Email:
Message:
Click here to create and insert a link
Ignore line feeds? no (typical)   yes (for HTML table input)


Viewing statistics for this thread
Period# Views# Users
Last hour11
Last 24 hours11
Last 7 days11
Last 30 days66
Since Mar 1, 2007485312