Poop Playoff Odds according to footballoutsiders.com - Page 6

Buzzsaw · 11-04-2010, 01:32 PM

I just posted this on the Jets site (my team for those that don't know), thought you guys might like to see what the predictions/simulations are for KC

Compiled by Mike Harris

The playoff odds report plays out the season 10,000 times. A random draw assigns each team a win or loss for each game. The probability that a team will be given a win is based on an equation which considers the current Weighted DVOA ratings of the two teams as well as home-field advantage. (DVOA ratings are explained here.) Mean Wins is the average number of wins for this team over the 10,000 simulations. Odds are given for winning each playoff seed, as well as for winning the division (DIV), earning a bye week (BYE), or winning a wild card (WC). TOT represents odds of making the playoffs in any fashion.

Code:

AFC West
Team	Rec	WEI DVOA	Mean Wins	#1	#2	#3	#4	#5	#6	DIV	BYE	WC	TOT	CHANGE
KC	5-2	29.2%	11.4	19.1%	24.9%	23.2%	25.5%	0.5%	1.0%	92.7%	44.0%	1.5%	94.2%	0.4%
SD	3-5	17.0%	8.0	0.0%	0.6%	1.4%	2.9%	1.8%	7.3%	4.9%	0.7%	9.1%	14.0%	4.0%
OAK	4-4	-9.1%	7.0	0.0%	0.2%	0.4%	1.7%	0.3%	1.6%	2.3%	0.2%	1.9%	4.2%	2.3%
DEN	2-6	-24.2%	4.6	0.0%	0.0%	0.0%	0.1%	0.0%	0.0%	0.1%	0.0%	0.0%	0.2%	-0.3%

The table's a little messed up, here's a screen

Playoff Scenarios
This report lists the odds of each team

* Reaching the Conference Championship Game
* Winning the Conference Championship Game
* Winning the Super Bowl

Code:

 Conf App  Conf Win    SB Win
NYG	50.3%	30.6%	16.0%
KC	43.1%	23.6%	13.1%
NE	41.1%	22.1%	12.0%
PIT	36.6%	19.4%	10.4%
PHI	32.7%	18.8%	10.3%
ATL	46.0%	22.5%	9.6%
GB	30.7%	14.4%	6.5%
TEN	18.3%	8.8%	5.0%
IND	17.2%	7.8%	3.8%
NYJ	17.7%	8.0%	3.7%
NO	17.7%	7.2%	2.8%
BAL	14.2%	5.8%	2.6%
MIA	5.7%	2.4%	1.2%
TB	6.8%	2.0%	0.7%
SEA	6.6%	1.9%	0.6%
SD	2.7%	1.1%	0.5%
HOU	2.6%	1.0%	0.3%
SF	1.8%	0.5%	0.2%
WAS	1.7%	0.5%	0.2%
CHI	1.6%	0.5%	0.2%
STL	1.9%	0.5%	0.1%
MIN	1.4%	0.4%	0.1%
DET	0.5%	0.2%	0.0%
OAK	0.4%	0.1%	0.0%
ARI	0.4%	0.0%	0.0%
JAC	0.2%	0.1%	0.0%
CLE	0.1%	0.0%	0.0%
CIN	0.1%	0.0%	0.0%

http://www.footballoutsiders.com/stats/playoffodds

Mr. Laz · 11-04-2010, 02:39 PM

In the last 1/4 of the season, the teams in contention for the playoffs step-up and starting playing at a higher level. We still need to see if the Chiefs have that extra gear.

gonna be hard to do without a legit passing game

ToxSocks · 11-04-2010, 02:46 PM

Quote:

Originally Posted by Laz

In the last 1/4 of the season, the teams in contention for the playoffs step-up and starting playing at a higher level. We still need to see if the Chiefs have that extra gear.

gonna be hard to do without a legit passing game

I fully expect the Chargers to be right back in this thing by the end of the season. We're going to finish at 9-7 or 10-6 most likely. Hopefully the Chargers don't get on a roll like they usually do.

These next two weeks are going to go a long ways in securing the division. If we beat the Raiders this week and the Chargers drop to the Texans, I think it'd be a good bet that the Chiefs take the division.

If the opposite happens, I think we're in trouble.

**Gadzooks** · 11-04-2010, 02:46 PM

Quote:

Originally Posted by Detoxing

I was thinking the same shit. Where is Ugly Duck? That's a good Raiders fan....

Despite his avi, I agree.

**Gadzooks** · 11-04-2010, 02:50 PM

Quote:

Originally Posted by Detoxing

I fully expect the Chargers to be right back in this thing by the end of the season. We're going to finish at 9-7 or 10-6 most likely. Hopefully the Chargers don't get on a roll like they usually do.

These next two weeks are going to go a long ways in securing the division. If we beat the Raiders this week and the Chargers drop to the Texans, I think it'd be a good bet that the Chiefs take the division.

If the opposite happens, I think we're in trouble.

I agree.

**cdcox** · 11-04-2010, 02:51 PM

Geez, what does a guy have to do to get noticed around here? It's not like I haven't been doing the exact same thing since the days of the Star Board. Where was FO then?

http://files.nfl-forecast.com/

Some reasons why you should use my predictions:

1) My process is transparent. I use team efficiency ratings from http://advancednflstats.com. If you are willing to read long enough, the stats are completely defined on that web site. FO, is great and everything, but what they are based on is behind the curtain. Their playoff predictions will disappear next week, never to be seen again. Good luck tracking the accuracy of their predictions over time. They take a "trust me" approach. My predictions are archived on my blog.

2) I provide an application where you can analyze scenarios. For example, what happens to our odds if we win against the Raiders this week?

http://nfl-forecast.com/

3) My tiebreakers are rock solid. I know that the tiebreakers on FO were incomplete for a while. I still don't know if they accurately handle things like 3-way ties, which occur much more commonly that you would think.

4) Chiefs-specific support on this board. I routinely provide in depth analysis of the Chiefs playoff odds (or draft position odds if we suck) on this board. In 2006, I had the exact scenario described for the Chiefs to make the playoffs and updated it throughout the day. People that were following my threads knew what it would take for weeks leading up to the end of the season. At the end of the 2007 I announced the procedure that would be used to determine the Chiefs draft position as the final FG sailed through the uprights. It took the NFL another 24 hours to sort it out.

No one else is going to be more in depth, or accurate, or timely than me.

You go posting FO playoff forecasts on here like it is the greatest thing since sliced bread and you are sticking a knife in me, bro. WTF.

**Rain Man** · 11-04-2010, 03:22 PM

Wait a minute. You mean cdcox is doing this, too?

**Rain Man** · 11-04-2010, 03:29 PM

I see it here and I see that cdcox's predictions also show us with a very high probability, and something that our gut feelings keep collectively saying seems high. I suspect that the reason is that these models necessarily use past performance to predict future performance, but a key injury can change future performance quite quickly.

I wonder, cdcox, if there's a justification for putting a weighting in that says something like, "if the odds are greater than 50 percent, then adjust down by some weighted factor." That factor could be based off of something simple like the proportion of times in the past seasons that teams with >50% odds actually made the playoffs. So in Week 1 the adjustment might be 50% if half the Week 1 horses don't make it, and in Week 8 it might be 77% and in Week 16 it might be 95%.

This type of adjustment would reflect not just injuries but also a simple cooling down or loss of momentum. Whadd'ya think?

**007** · 11-04-2010, 03:34 PM

Quote:

Originally Posted by cdcox

Geez, what does a guy have to do to get noticed around here? It's not like I haven't been doing the exact same thing since the days of the Star Board. Where was FO then?

http://files.nfl-forecast.com/

Some reasons why you should use my predictions:

1) My process is transparent. I use team efficiency ratings from http://advancednflstats.com. If you are willing to read long enough, the stats are completely defined on that web site. FO, is great and everything, but what they are based on is behind the curtain. Their playoff predictions will disappear next week, never to be seen again. Good luck tracking the accuracy of their predictions over time. They take a "trust me" approach. My predictions are archived on my blog.

2) I provide an application where you can analyze scenarios. For example, what happens to our odds if we win against the Raiders this week?

http://nfl-forecast.com/

3) My tiebreakers are rock solid. I know that the tiebreakers on FO were incomplete for a while. I still don't know if they accurately handle things like 3-way ties, which occur much more commonly that you would think.

4) Chiefs-specific support on this board. I routinely provide in depth analysis of the Chiefs playoff odds (or draft position odds if we suck) on this board. In 2006, I had the exact scenario described for the Chiefs to make the playoffs and updated it throughout the day. People that were following my threads knew what it would take for weeks leading up to the end of the season. At the end of the 2007 I announced the procedure that would be used to determine the Chiefs draft position as the final FG sailed through the uprights. It took the NFL another 24 hours to sort it out.

No one else is going to be more in depth, or accurate, or timely than me.

You go posting FO playoff forecasts on here like it is the greatest thing since sliced bread and you are sticking a knife in me, bro. WTF.

It probably didn't come across that way but that was kind of my point in my post earlier in the thread. It seems to me that they pretty much copied your playoff percentages and let it go at that.

**Rausch** · 11-04-2010, 03:35 PM

I could be reading this wrong but it appears to question Milkman's sexual orientation...

Pasta Little Brioni · 11-04-2010, 03:38 PM

FO loves the Chiefs

**cdcox** · 11-04-2010, 03:55 PM

Quote:

Originally Posted by Rain Man

I see it here and I see that cdcox's predictions also show us with a very high probability, and something that our gut feelings keep collectively saying seems high. I suspect that the reason is that these models necessarily use past performance to predict future performance, but a key injury can change future performance quite quickly.

I wonder, cdcox, if there's a justification for putting a weighting in that says something like, "if the odds are greater than 50 percent, then adjust down by some weighted factor." That factor could be based off of something simple like the proportion of times in the past seasons that teams with >50% odds actually made the playoffs. So in Week 1 the adjustment might be 50% if half the Week 1 horses don't make it, and in Week 8 it might be 77% and in Week 16 it might be 95%.

This type of adjustment would reflect not just injuries but also a simple cooling down or loss of momentum. Whadd'ya think?

I've given the problem of too much certainty early in the season a lot of thought.

The challenge is that it is a zero sum game. If you drop the Chiefs by x amount you have to increase the other teams in the division or conference by the difference in order to get the totals to sum to 100%. It would be fairly simple if you were distributing the excess uncertainty to the other three teams in the division. But the WC makes it much more complicated because some of the uncertainty needs to be distributed within the division and some needs to be distributed to teams in other divisions that might win the WC.

A while back I tried to add uncertainty to the efficiency ratings. I used the end-of-season efficiency ratings as the "true ones" and calculated how much early season efficiency ratings had historically deviated from that as a function of week. Unfortunately (should have seen it coming) the distribution of wins is almost exactly the same whether your include uncertainty in team efficiency or not. For every season that a team got worse over the course of the year, there was a season where they got better. And sometimes when you get worse, your opponents get worse too.
So that approach didn't work at all.

Another approach would be to assign the efficiency ratings as a weighted average between the published rating of the team and 0.5, which is the mean rating. The weighting factor would increase toward the published one as the season went on. I would need to test that idea to see if the league wide predictions got better or worse by doing it. Probably won't get time to try it this year.

I'm open to hearing other ideas on how to make the early season predictions more reasonable.

**cdcox** · 11-04-2010, 04:03 PM

Quote:

Originally Posted by Guru

It probably didn't come across that way but that was kind of my point in my post earlier in the thread. It seems to me that they pretty much copied your playoff percentages and let it go at that.

No, I appreciated your earlier comment.

They use their power ratings for their predictions. Sometimes there are significant differences. For example, they have NE as a 90% playoff probability while I only estimate their odds as 72%. FO does legitimate work, it just struck in my craw. I guess I was in the mood to rant.

KCHawg · 11-04-2010, 04:51 PM

Quote:

Originally Posted by Detoxing

I was thinking the same shit. Where is Ugly Duck? That's a good Raiders fan....

And Overhead, he's alright for a 'nuck!

teedubya · 11-04-2010, 04:55 PM

I see the Chiefs winning one playoff game in January. But, that will be because of the run... when Cassel has to throw with the pressure of the playoffs, we will see a huge brown stain on his pants during crunchtime.

**Gadzooks** · 11-04-2010, 05:05 PM

Quote:

Originally Posted by KCHawg

And Overhead, he's alright for a 'nuck!

Over-head's probably busy in the shed.
(I also like RedNeckRaider as far criminals go)

11-04-2010, 01:32 PM
Buzzsaw Starter Join Date: May 2005 Casino cash: $10004900	Playoff Odds according to footballoutsiders.com I just posted this on the Jets site (my team for those that don't know), thought you guys might like to see what the predictions/simulations are for KC Compiled by Mike Harris The playoff odds report plays out the season 10,000 times. A random draw assigns each team a win or loss for each game. The probability that a team will be given a win is based on an equation which considers the current Weighted DVOA ratings of the two teams as well as home-field advantage. (DVOA ratings are explained here.) Mean Wins is the average number of wins for this team over the 10,000 simulations. Odds are given for winning each playoff seed, as well as for winning the division (DIV), earning a bye week (BYE), or winning a wild card (WC). TOT represents odds of making the playoffs in any fashion. Code: AFC West Team Rec WEI DVOA Mean Wins #1 #2 #3 #4 #5 #6 DIV BYE WC TOT CHANGE KC 5-2 29.2% 11.4 19.1% 24.9% 23.2% 25.5% 0.5% 1.0% 92.7% 44.0% 1.5% 94.2% 0.4% SD 3-5 17.0% 8.0 0.0% 0.6% 1.4% 2.9% 1.8% 7.3% 4.9% 0.7% 9.1% 14.0% 4.0% OAK 4-4 -9.1% 7.0 0.0% 0.2% 0.4% 1.7% 0.3% 1.6% 2.3% 0.2% 1.9% 4.2% 2.3% DEN 2-6 -24.2% 4.6 0.0% 0.0% 0.0% 0.1% 0.0% 0.0% 0.1% 0.0% 0.0% 0.2% -0.3% The table's a little messed up, here's a screen Playoff Scenarios This report lists the odds of each team * Reaching the Conference Championship Game * Winning the Conference Championship Game * Winning the Super Bowl Code: Conf App Conf Win SB Win NYG 50.3% 30.6% 16.0% KC 43.1% 23.6% 13.1% NE 41.1% 22.1% 12.0% PIT 36.6% 19.4% 10.4% PHI 32.7% 18.8% 10.3% ATL 46.0% 22.5% 9.6% GB 30.7% 14.4% 6.5% TEN 18.3% 8.8% 5.0% IND 17.2% 7.8% 3.8% NYJ 17.7% 8.0% 3.7% NO 17.7% 7.2% 2.8% BAL 14.2% 5.8% 2.6% MIA 5.7% 2.4% 1.2% TB 6.8% 2.0% 0.7% SEA 6.6% 1.9% 0.6% SD 2.7% 1.1% 0.5% HOU 2.6% 1.0% 0.3% SF 1.8% 0.5% 0.2% WAS 1.7% 0.5% 0.2% CHI 1.6% 0.5% 0.2% STL 1.9% 0.5% 0.1% MIN 1.4% 0.4% 0.1% DET 0.5% 0.2% 0.0% OAK 0.4% 0.1% 0.0% ARI 0.4% 0.0% 0.0% JAC 0.2% 0.1% 0.0% CLE 0.1% 0.0% 0.0% CIN 0.1% 0.0% 0.0% http://www.footballoutsiders.com/stats/playoffodds
Posts: 331

11-04-2010, 02:39 PM	#76
Mr. Laz Don't Tease Me Join Date: Dec 2000 Location: KS Casino cash: $11047037	In the last 1/4 of the season, the teams in contention for the playoffs step-up and starting playing at a higher level. We still need to see if the Chiefs have that extra gear. gonna be hard to do without a legit passing game __________________
Posts: 95,626

11-04-2010, 02:51 PM	#80
cdcox www.nfl-forecast.com Join Date: Sep 2000 Casino cash: $2221769	Geez, what does a guy have to do to get noticed around here? It's not like I haven't been doing the exact same thing since the days of the Star Board. Where was FO then? http://files.nfl-forecast.com/ Some reasons why you should use my predictions: 1) My process is transparent. I use team efficiency ratings from http://advancednflstats.com. If you are willing to read long enough, the stats are completely defined on that web site. FO, is great and everything, but what they are based on is behind the curtain. Their playoff predictions will disappear next week, never to be seen again. Good luck tracking the accuracy of their predictions over time. They take a "trust me" approach. My predictions are archived on my blog. 2) I provide an application where you can analyze scenarios. For example, what happens to our odds if we win against the Raiders this week? http://nfl-forecast.com/ 3) My tiebreakers are rock solid. I know that the tiebreakers on FO were incomplete for a while. I still don't know if they accurately handle things like 3-way ties, which occur much more commonly that you would think. 4) Chiefs-specific support on this board. I routinely provide in depth analysis of the Chiefs playoff odds (or draft position odds if we suck) on this board. In 2006, I had the exact scenario described for the Chiefs to make the playoffs and updated it throughout the day. People that were following my threads knew what it would take for weeks leading up to the end of the season. At the end of the 2007 I announced the procedure that would be used to determine the Chiefs draft position as the final FG sailed through the uprights. It took the NFL another 24 hours to sort it out. No one else is going to be more in depth, or accurate, or timely than me. You go posting FO playoff forecasts on here like it is the greatest thing since sliced bread and you are sticking a knife in me, bro. WTF.
Posts: 45,543

11-04-2010, 03:22 PM	#81
Rain Man NFL's #1 Ermines Fan Join Date: Jul 2001 Location: My house Casino cash: $3058491 VARSITY	Wait a minute. You mean cdcox is doing this, too? __________________ I'm putting random letters here as a celebration of free speech: xigrakgrah misorojeq rkemeseit.
Posts: 141,648

11-04-2010, 03:29 PM	#82
Rain Man NFL's #1 Ermines Fan Join Date: Jul 2001 Location: My house Casino cash: $3058491 VARSITY	I see it here and I see that cdcox's predictions also show us with a very high probability, and something that our gut feelings keep collectively saying seems high. I suspect that the reason is that these models necessarily use past performance to predict future performance, but a key injury can change future performance quite quickly. I wonder, cdcox, if there's a justification for putting a weighting in that says something like, "if the odds are greater than 50 percent, then adjust down by some weighted factor." That factor could be based off of something simple like the proportion of times in the past seasons that teams with >50% odds actually made the playoffs. So in Week 1 the adjustment might be 50% if half the Week 1 horses don't make it, and in Week 8 it might be 77% and in Week 16 it might be 95%. This type of adjustment would reflect not just injuries but also a simple cooling down or loss of momentum. Whadd'ya think? __________________ I'm putting random letters here as a celebration of free speech: xigrakgrah misorojeq rkemeseit.
Posts: 141,648

11-04-2010, 03:38 PM	#85
Pasta Little Brioni Consuming CP souls Join Date: Oct 2005 Location: U.S.A. Casino cash: $3728880	FO loves the Chiefs __________________ **Official TFWdemB Trivia Commisioner**
Posts: 68,848

11-04-2010, 04:55 PM	#89
teedubya Most Valuable Poster Join Date: Oct 2003 Casino cash: $9470002	I see the Chiefs winning one playoff game in January. But, that will be because of the run... when Cassel has to throw with the pressure of the playoffs, we will see a huge brown stain on his pants during crunchtime.
Posts: 36,659