PatsFans.com Menu
PatsFans.com - The Hub For New England Patriots Fans

AccuScore simulation: AFC Championship game


Status
Not open for further replies.

weswelker#83

In the Starting Line-Up
Joined
Dec 13, 2007
Messages
4,528
Reaction score
4
AccuScore.com has run more than 10,000 simulations for Yahoo! Sports, calculating how each team's performance changes in response to game conditions, opponent's abilities, roster moves, weather and more. Each game is simulated one play at a time and the game is replayed a minimum of 10,000 times to generate forecasted winning percentages, player statistics and a variety of game-changing scenarios. Here's AccuScore.com's analysis of the AFC Championship game.
AFC CHAMPIONSHIP GAME

The Chargers pulled off the big road upset against the Colts, but things do not look good for the Chargers in New England. Even if LaDainian Tomlinson, Philip Rivers and Antonio Gates were not hurt, the Patriots would be heavy favorites at home. The Patriots are winning nearly 90 percent of simulations, converting over 53 percent of third downs. Tom Brady completes 70 percent of his passes for 333 yards and over 3 TD passes per simulation.

Laurence Maroney is actually out-rushing LT on both a per carry and total rushing yards basis (Maroney 4.3 ypc vs. LT 4.1). For San Diego to pull off the huge upset, it needs Rivers to pass for at least 2 TDs and no interceptions. In simulation losses, Rivers averages 1 TD pass and 1.1 INTs, but in simulation wins he has a 3-1 touchdown-to-interception ratio.

The Chargers' defense needs to intercept Brady twice. While San Diego managed to do that vs. Peyton Manning, the chances of Brady throwing 2 picks is under 10 percent. Look for Randy Moss to bounce back from his one-reception performance vs. the Jaguars with a huge 95 yards and at least 1 TD.

Simulation Full Results :

http://sports.yahoo.com/nfl/news;_y...-accuscore-nfl20080120017&prov=yhoo&type=lgns
 
Thanks for posting, good information
 
Accuscore has been leaning toward the favorites this season.
It has consistently overestimated the Pats in the 2nd half of the season and would lose the point spread. It has been giving too low a chance for Pats' opponents.
 
Accuscore has been leaning toward the favorites this season.
It has consistently overestimated the Pats in the 2nd half of the season and would lose the point spread. It has been giving too low a chance for Pats' opponents.
???? How did it overestimate predicting a winner by overestimating the Pats? Which opponent beat us? It seems to me that if Accurscore picked the Pats each time, then they are 17 for 17 with the Pats.

Don't worry if they gave the Pats an 85% chance of winning or a 65% chance. The key questions are three:

1. Who do they predict will win the AFCCG, and

2. What is their track record for predicting winners this year, and

3. Who did they predict would win the Colts-Chargers game?
 
Why does anyone care what a computer simulations thinks is going to happen? I guess you can say one thing for it though; it's not as bad as when they trumpet who won the Madden simulation before games.
 
???? How did it overestimate predicting a winner by overestimating the Pats? Which opponent beat us? It seems to me that if Accurscore picked the Pats each time, then they are 17 for 17 with the Pats.

Don't worry if they gave the Pats an 85% chance of winning or a 65% chance. The key questions are three:

1. Who do they predict will win the AFCCG, and

2. What is their track record for predicting winners this year, and

3. Who did they predict would win the Colts-Chargers game?

I meant against the spreads and actual final scores. I'm interested in accurate prediction rather than just who wins (of course Accuscore is 17 for 17 on the Pats). For this game, if I give significant weight to SD victory over Colts, then NE should have only 76% chance vs. 24% for SD. In many simulation scenarios, SD can both slow down Pats O, and put long drives on Pats D. Remember the first game, Pats O scored only 31 pts; 7 pts came from AD Thomas' pick 6; 7 other pts came from short field of Colvin's FR. Computer analysis shows that SD has become much better in the last 8 games on both sides of the ball.

If Billy Volek (whom I like back the days he backed up for Air McNair) has an inspired game a la AJ Feeley or Kyle Boller (or pull a Jeff-Hostetler or Frank-Reich), the game can be competitive. It's crazy but it's almost I rather have the devil we know (PR) than the one we don't know.

The 90%-10% split by Accuscore is as if all 17 games are counted equally, or they factor in LT, Gates, PR injuries.

But there are two major unknowns: 1) weather, will it be 22 MPH wind?; and 2) injuries. Many people I know will not touch this game because of these. Remember the NE-BAL and NE-Jets game in swirling wind and/or rain? Accuscore were way off. My pals (who never met an underdog they didn't like) took the weather dogs and both times. They just reminded me today.:)


Your questions 2 and 3:
2) I believe they are only 52-55% against spread. (which is the same as random; If they were better, everyone would have just followed them and Vegas would have adjusted their lines to them; Note that some dart boards and picks from a hat can do better)
3) Of course they predicted something like 80ish % Colts win over SD (I didn't have the exact record)
 
Last edited:
Why does anyone care what a computer simulations thinks is going to happen? I guess you can say one thing for it though; it's not as bad as when they trumpet who won the Madden simulation before games.

Apricissimus, some people like forecast for fun, like futures of oils or other commodities. They look at weather for the whole year and worry about corn/ethanol potential yields, impact at the pump, etc... It's a style of living...:)

BTW, Madden is a game, not serious simulation like Accuscore. A computer program that is best for devising strategy, but not for simulation is ZEUS.
 
Last edited:
I played 10,000 games on Madded 08 with Pats and Chargers and the Chargers won 85% of the time. Boy glad I did that now I don't have to worry about the Pats winning the game.
 
Isn't it funny how LT said that if we played you guys 10 times we would beat you 9 times and this says you have a 90% chance of winning? coincidence? maybe. But 10% is all we need because this game is only going to be played once, not 300 times. As the Schwam says: "and that's why they play the game."
 
Why does anyone care what a computer simulations thinks is going to happen? I guess you can say one thing for it though; it's not as bad as when they trumpet who won the Madden simulation before games.

Madden nailed the Eagles upset on the road of the Cowboys. It was eerie. No other "expert" or even accuscore saw it coming. You could have made a fortune if you could have bet that the Cowboys would fail to score a TD at home in that game, which is what Madden predicted.
 
Last edited:
Madden nailed the Eagles upset on the road of the Cowboys. It was eerie. No other "expert" or even accuscore saw it coming. You could have made a fortune if you could have bet that the Cowboys would fail to score a TD at home in that game, which is what Madden predicted.

coincidence. Where does superstition come from?
 
Isn't it funny how LT said that if we played you guys 10 times we would beat you 9 times
Lawrence Taylor said what?

Oh, you mean LdT, babyLT?

Who cares what he said.

LdT tweaked his knee and refused to come back into the game without an MRI.

What a guy. Gates must be so proud of him.

Better hope he doesn't get an ouchie against the mean ol' Patriots.
 
Lawrence Taylor said what?

Oh, you mean LdT, babyLT?

Who cares what he said.

LdT tweaked his knee and refused to come back into the game without an MRI.

What a guy. Gates must be so proud of him.

Better hope he doesn't get an ouchie against the mean ol' Patriots.

and you hating on everyone except yourself and the Patriots means what to me?
 



Look at it this way , the algorithm is right 99% (numbers and facts) , what happened in Chargers-Colts game is the computer has not processed the situation when the quaterback Chokes .
Don't blame the algorithm , but blame only Peyton Manning who screw up (fumbles/interceptions/red zone fiascos) what everybody predicted wich is a colts win .


I hope starting next year , they will add the "Choke" factor in their algorithm for a better result.


choke-a-duck.jpg
 
Last edited:
AccuScore? Do you really believe that if they had a trusted algorithm they would give it to the world?

Computer simulations are for fun. Sometimes they're right, and sometimes they're dead wrong - just like the humans who programmed them.
 
AccuScore? Do you really believe that if they had a trusted algorithm they would give it to the world?

Computer simulations are for fun. Sometimes they're right, and sometimes they're dead wrong - just like the humans who programmed them.


I bet you are a colt fan? tell me i am wrong now .


/Nope i was right , i checked your posts , and you are definitely a biaised colt fan.
 
Accuscore was 163--89 picking the winner in the regular season and is 5--3 so far in the postseason (picked dallas, colts and tampa bay). So, that probably makes them just a little better than a well-informed fan.
 
Status
Not open for further replies.


Tuesday Patriots Notebook 4/16: News and Notes
Monday Patriots Notebook 4/15: News and Notes
Patriots News 4-14, Mock Draft 3.0, Gilmore, Law Rally For Bill 
Potential Patriot: Boston Globe’s Price Talks to Georgia WR McConkey
Friday Patriots Notebook 4/12: News and Notes
Not a First Round Pick? Hoge Doubles Down on Maye
Thursday Patriots Notebook 4/11: News and Notes
MORSE: Patriots Mock Draft #5 and Thoughts About Dugger Signing
Matthew Slater Set For New Role With Patriots
Wednesday Patriots Notebook 4/10: News and Notes
Back
Top