Week 2 Results

NFLSim had another great week- even greater because it was only the second week of the season. Straight up picks went 9-7 (56.3%) overall. On the surface that's pretty disappointing, but check out how those wins are distributed within the confidence intervals. 50% confidence picks went 5-4 (55.6%) and 60% confidence picks went 3-2 (60%). Missed the 70% game (MIA over IND), that goes down as the most disappointing game of the week :(. The 80% game, DEN over CLE went pretty well.

The picks against the spread had another amazing week. Mia +4 was a tie so I left it out. Overall ATS went 10-5 (66.7%). 50% confidence games went 3-3 (50%), 60% confidence games went 3-1 (75%), 70% confidence games went 4-1 (80%). Come on. You can't argue with that.


That being said, O/U picks were a total debacle at 5-10 (33.3%). NFLSim was 50-50 on the PHI-NO O/U.


More analysis under the numbers...

STRAIGHT UP
CONFIDENCE	GAMES	CORRECT	% CORRECT
50-59	9	5	55.56%
60-69	5	3	60.00%
70-79	1	0	0.00%
80-89	1	1	100.00%
90-100	0	0	---
TOTAL	16	9	56.25%
ATS
CONFIDENCE	GAMES	CORRECT	% CORRECT
50-59	6	3	50.00%
60-69	4	3	75.00%
70-79	5	4	80.00%
80-89	0	0	---
90-100	0	0	---
TOTAL	15	10	66.67%
O/U
CONFIDENCE	GAMES	CORRECT	% CORRECT
50-59	9	4	44.44%
60-69	5	1	20.00%
70-79	1	0	0.00%
80-89	0	0	---
90-100	0	0	---
TOTAL	15	5	33.33%

Best game of the week: HOU +7 to win straight up against TEN at 54%... there's a moneymaker.
     Runner up: NYJ +4 to win straight up against NE at 67%... take THAT Brady.
Game I cried after: IND at MIA... I wanted it as much for Miami as I wanted it for me.
The game that no one saw coming: CIN beat GB... huh?
Worst pick of the week: JAX -3.5 ...

Last week's 'CLE at DEN' analysis predicted a fairly even match up with the expectation that DEN would take control on the ground and in the red zone. Buckhalter and Moreno took 76 and 75 yards a piece, with a TD for Buckhalter. Denver went 2-4 in the red zone w/ 2 TDs, Cleveland went 0-1. Denver dominated the air for most of the game.

Stay tuned for more predictions, more analyses, and more surprises next week!

Week 2 Picks

Got an early start this week. This week's spreadsheet can be found at...

Week 2 Picks & Stats
Week 1 Picks & Stats

I add games as the week goes by, so keep checking back!
During the course of the week, if the line changes by at least 1 point I'll redo the sim. Anything less than that only changes the % by 1 or 2 pts. I'll let everyone know of the change when it happens. When I info to the post, I'll separate them at the bottom with '---'.

 Watch out for the NE NYJ game... NYJ to win at 67%. I simulated twice to be sure. The numbers make sense, but Brady is a wild card.
---

The rest of the stats will be up a little later, Friday at the very latest.

---

Most of the stats are up.

CLE at DEN
Check out the Denver game! I don't think I've ever seen such a mismatch. DEN averaging 26 and CLE averaging 11 points. This is the first 70+% straight up game of the season... it's the first 80+%, too. Their passing games are pretty even. It looks like it really came down to the run game. Oddly enough, they perfectly matched each other in rush yds and carries, but on average, DEN scored 1 more rushing TD than CLE. 1.5 to 0.6. Sounds like red zone efficiency to me. Watch Buckhalter and Moreno destroy Cleveland's 32nd ranked run defense.

---

Alright, everything is up for the week. It looks like NFLSim went a little wild this weekend with its predictions. I'm going to chalk it up to the small amount of stats to work with. Then again, last week's predictions were a bit counter-intuitive and they worked out pretty nicely. Use these picks at your own discretion, don't rely on them for anything too important. Next week we'll be in business fo' sho'.

There's a 2nd 70+% straight up game: MIA over IND, who'da thunk it. In both passing and rushing, Miami was able to score more touchdowns.

Then again, there are some experts at ESPN and other sports news outlets that agree with some of the more eccentric picks...

2008 Season Results

CHECK OUT THE POST ABOVE, BLACK BOX NAME CHANGE CONTEST (win $20 $50, free to enter, just submit a new name!)

These results from last year, excluding weeks 1 and 2, which there wasn't enough data for. It's broken into 3 sections: Straight up Winner, Against the Spread, and Over/Under. I broke down the approximately 200 games into 5 intervals, each interval representing a confidence spread of 10%. When you see 96 games in the 50-59% interval, that means of the 200+- games NFLSim picked, 96 games fell in the confidence interval of 50-59%. Likewise, 70 games had winners predicted as 60-69% favorites (in the straight-up category). To the left of the 'Games' column is the 'Correct' column. That is simply the number of NFLSim's correct picks in that interval. Accuracy is the percentage of time that NFLSim correctly picked a game in that interval.

1. Straight up Winner
NFLSim was absolutely phenomenal when it came to picking the winner. At first glance, you see the season record at 139-70 (66.5%). That's nothing special until you take a deeper look at the statistics.

Straight-up	% Interval	Wins	Games	Win %
	50-59%	61	96	63.5%
	60-69	45	70	64.3
	70-79	27	36	75
	80-89	4	5	8
	90-100	2	2	1
	Total	139	209	66.5

First, lets take a look at the 'expected number of wins' for each interval. In the 50-59% interval, it can be expected that after a large number of games, 55% of those games should be picked correctly(the mid-point of the 50's). In the 60-69% interval, 65% should be picked correctly, and so on. The closer NFLSim's number of correct picks is to the expected number of correct picks, the better NFLSim approximates reality. This is very important, because if you know the exact probability of a team winning, you automatically know if Vegas' payout makes for a good bet or a bad bet. The closer that the predicted probability is to the actual probability, the smarter your bets are. The smarter your bets, the more $ you make! (This is the basis for the Bet Optimization Algorithm.... which is new and much improved this year.)

If we take a look at the first interval, 50-59%, 96 games took place. In theory, you should expect that 55% of those teams predicted to win should actually win. 55% of 96 games, 55% * 96, equals an expected number of wins of 53. NFLSim actually picked 61 winners, so it beat the expected number of wins by 8. This is great because it picked more winners than it was supposed to, but it was pretty far from the expected number of wins. From 60% on is where it gets very, very interesting. NFLSim reaches near theoretical perfection - the probability given by NFLSim is the actual probability that the team will win!!! If anyone has seen this feat achieved by anyone else please let me know.

Here is a table of expected numbers of wins:

	Expected Correct	Actual Correct
50-59%	53	61
60-69	46	45
70-79	27	27
80-89	4	4
90-100	2	2
Total	132	139

In the graph: Red is what should have happened, Blue is what did happen.



Now that I've gone through the whole explanation of the stats, I'll skip straight to the numbers.

2. ATS
ATS didn't do as well as straight-up. Looks like it was pretty close to a flip of a coin. The right team would win by the wrong score. I would hate to tamper with the program when the straight-up accuracy is so reliable after 209 games. I'll probably just stop posting ATS until I can fix it up.

ATS	% Interval	Wins	Games	Win %
	50-59	37	86	43
	60-69	46	78	59
	70-79	18	33	54.5
	80-89	0	3	0
	90-100	1	1	100
	Total	102	201	50.7

	Exp. Correct	Actual Correct
50-59%	47	37
60-69	51	46
70-79	25	18
80-89	3	0
90-100	1	1
Total	127	102

Interestingly, the slope of the trend line is almost what it should be. Had 1 of the 80% games won, it would be almost perfect (albeit far lower than it should be). That suggests that all I may need to do is to add/subtract from the calculated probability to arrive at the true probability... interesting. There is back testing to be done!

3. Over/Under
Finally, if you haven't gone to sleep yet, here are the O/U stats. O/U actually did much better than I expected. It performed much better than the ATS and its accuracy approached the expected accuracy.

Over/Under	% Interval	Wins	Games	O/U Win %
	50-59	77	145	53.1
	60-69	30	48	62.5
	70-79	3	6	50
	80-89	2	3	66.7
	90-100	2	2	100
	Total	114	204	55.9

	Exp. Correct	Actual Correct
50-59%	80	77
60-69	31	30
70-79	5	3
80-89	3	2
90-100	2	2
Total	121	114

O/U actually performed very well compared to what was expected. And hell, at 56% accuracy, you'll be making money anyway.












You've seen the performance, you've seen the stats, you know what to expect. What it comes down to:

Everyone's a winner with Black Box Sports. (especially if you win $20 $50 in the name change contest)

CHECK OUT THE POST BELOW FOR THE NAME CHANGE CONTEST (win $20 $50 for sending in your best name)

Jonathan
Black Box Sports

Week 10: Results

Week 10 was disappointing in terms of ATS and O/U, but NFLSim picked 11-3 straight up, so it wasn't all bad. The 50-59% games really got slammed, but if you read the Midseason Progressive Results post, you wouldn't have bet on them anyway. After this week, 60+% ATS and O/U picks are still 61.3 and 66.7% on the season, respectively.

I just found out about another NFL simulator over at www.whatifsports.com, a division of Fox Sports. Actually, they've got a pretty cool setup. They let you simulate games using teams from different seasons. I can't find their pick record, but someone posted that they are 69.5% straight up and 53.7% ATS on the season. No confidence numbers were available. On the season they beat AccuScore and NFLSim by a few % points, but when picks are made based on confidence, NFLSim has a substantial edge.

Straight	%	Wins	Games	Win %
	50-59	6	7	85.7%
	60-69	2	4	50.0%
	70-79	3	3	100.0%
	80-89	0	0	---
	90-100	0	0	---
	Total	11	14	78.6%
Spread	%	Wins	Games	Win %
	50-59	1	9	11.1%
	60-69	1	2	50.0%
	70-79	1	3	33.3%
	80-89	0	0	---
	90-100	0	0	---
	Total	3	14	21.4%
Total		Wins	Games	Win %
	50-59	4	12	33.3%
	60-69	1	2	50.0%
	70-79	0	0	---
	80-89	0	0	---
	90-100	0	0	---
	Total	5	14	35.7%

Week 10: Predictions

Ok, here's the new format. Same thing as before, but no I add the average scores as well. Underneath the matchup are the average scores. DEN's average score over the course of 500 simulations was 21.25 and CLE's average score was 21.40. Average scores aren't predictions of the score, they're just measures of relative performance. Every once in a while, one team will be favored to win even though on average they score less. This is the case with the DEN game. I wish there was an easier way to cram more stats into each game, but this blog format is really limiting.

I'm also planning on posting games as I finish them. One last change, instead of reposting everything an a "Results" section, i'm going to color everything in the "Prediction" section. A small "Results" section will sum up the weekend.

Only 2 upsets this week, DEN over CLE (52%) and HOU over BAL (57%). More 'UNDER's than usual. More favorites covering the spread, as well.

DEN at CLE -3.5, 45.5
DEN: 21.25
CLE: 21.40
Winner: DEN, 52%
Spread: DEN +3.5, 60%
O/U: UNDER, 61%

BAL -1 at HOU, 42
BAL: 21.57
HOU: 23.12
Winner: HOU, 57%
Spread: HOU +1, 59%
O/U: OVER, 55%

BUF at NE -3.5, 41.5
BUF: 22.10
NE: 23.41
Winner: NE, 55%
Spread: BUF +3.5, 54%
O/U: OVER, 59%

GB at MIN -2, 45.5
GB: 19.95
MIN: 23.45
Winner: MIN, 57%
Spread: MIN -2, 54%
O/U: UNDER, 59%

JAX -6.5 at DET, 44.5
JAX: 24.65
DET: 22.00
Winner: JAX, 56%
Spread: DET +6.5, 64%
O/U: OVER, 56%

NO at ATL -1, 50
NO: 22.22
ATL: 26.39
Winner: ATL, 59%
Spread: ATL -1, 58%
O/U: UNDER, 57%

SEA at MIA -8.5, 43
SEA: 17.19
MIA: 29.42
Winner: MIA, 79%
Spread: MIA -8.5, 54%
O/U: OVER, 59%

STL at NYJ -8.5, 44
STL: 19.05
NYJ: 27.06
Winner: NYJ, 70%
Spread: STL +8.5, 51%
O/U: OVER, 53%

TEN -3 at CHI, 38.5
TEN: 22.33
CHI: 18.47
Winner: TEN, 60%
Spread: CHI +3, 51%
O/U: OVER, 52%

CAR -9.5 at OAK, 37.5
CAR: 21.15
OAK: 19.93
Winner: CAR, 54%
Spread: OAK +9.5, 72%
O/U: OVER, 58%

IND at PIT -3.5, 39.5
IND: 17.97
PIT: 22.27
Winner: PIT, 62%
Spread: PIT -3.5, 54%
O/U: UNDER, 51%

KC at SD -15, 47
KC: 18.93
SD: 26.02
Winner: SD, 70%
Spread: KC +15, 74%
O/U: UNDER, 60%

NYG at PHI -3, 43.5
NYG: 18.88
PHI: 25.48
Winner: PHI, 65%
Spread: PHI -3, 58%
O/U: OVER, 52%

SF at ARI -9.5, 45.5
SF: 20.34
ARI: 24.25
Winner: ARI, 63%
Spread: SF +9.5, 70%
O/U: UNDER, 57%

Midseason Progressive Results

And now, what everyone's been waiting for. We're about halfway through the season, so I think it's time to post NFLSim's impressive progressive results. I chose to leave out week 2 because It was still too early for me to be making predictions based on 1 week of data. I'm going to post Accuscore's results as well for a comparison. They're the only other play-by-play NFL simulator I know of and they're a well established, well funded, syndicated, sponsored, and mathematically sophisticated operation. David and Goliath? Let's see...

I'm going to give you several different numbers. First, I'll give you the overall numbers, as in the collective 50-100% predictions for winner, spread, and o/u. Then you'll get their numbers broken down. I'll show you weekly trends, % trends, etc.

Overall numbers:
Winner: 65-34 (65.7%)
Spread: 47-45 (51.5%)
O/U: 50-43 (53.8%)
Spread and O/U combined: 97-88 (53.8%)

Accuscore numbers:
Winner: 87-43 (66.9%)
Spread: 51-50 (50.5%)
O/U: 67-50 (57.2%)
Spread and O/U combined: 118-100 (54.1%)

Accuscore has a slight edge when picking the winner, I have a slight edge picking ATS, and Accuscore has a sizeable advantage picking O/U. BUT! Let's look at how Black Box Sports picks compare when confidence is at least 60%. This confidence is really where NFLSim shines. Here is Black Box Sports's record when the confidence is over 60% compared to Accuscore's overall record (can't find any confidence values for picks). Fasten your seat belts.

Black Box Sports +60%:
Winner: 41-19 (68.3%) ... 22-8 (73.3%) when greater than 70%
Spread: 36-21 (63.2%)
O/U: 21-10 (67.7%)
Spread and O/U: 57-31 (64.7%)
Betting 100 units on the spread and over/under, you made +2290, ROI of 26%, halfway through the season.

Once again, Accuscore's numbers:
Accuscore:
Winner: 87-43 (66.9%)
Spread: 51-50 (50.5%)
O/U: 67-50 (57.2%)
Spread and O/U combined: 118-100 (54.1%)

Picking the winner, I'm ahead by 1.4% when at least 60% confidence, 6.4% when at least 70% confidence. Spread, I'm ahead by 12.7%. O/U, up by 10.5%. Combined, I'm up by 10.6%. That's what I'm talking about. Not to mention I offer all the picks for free... Spread the word everyone.

In tabular format:

Winner		Wins	Games	Win %
	50-59	24	39	61.5%
	60-69	19	30	63.3%
	70-79	17	24	70.8%
	80-89	4	4	100.0%
	90-100	1	2	50.0%
	Total	65	99	65.7%
Spread		Wins	Games	Win %
	50-59	11	35	31.4%
	60-69	24	37	64.9%
	70-79	11	16	68.8%
	80-89	0	3	0.0%
	90-100	1	1	100.0%
	Total	47	92	51.1%
Over		Wins	Games	Win %
	50-59	29	62	46.8%
	60-69	15	22	68.2%
	70-79	2	4	50.0%
	80-89	2	3	66.7%
	90-100	2	2	100.0%
	Total	50	93	53.8%

S&O/U
	50-59	40	97	41.2%
	60-69	39	59	66.1%
	70-79	13	20	65.0%
	80-89	2	6	33.3%
	90-100	3	3	100.0%
	Total	97	185	52.4%

Graphically:
Theoretically, as in, if NFLSim was a perfect analog of reality, those dashed lines would be perfectly in line with the thick black line. It would mean that the confidence values are always spot on and the games end exactly the way they should. If the trend (dashed) lines are below the thick line, the confidence values are not as accurate as reality. The more parallel the thick and trend lines are, the more accurate the changes in confidenc values are, i.e., as confidence increases, the accuracy increases at the correct rate. If that makes any sense. This graph shows all the picks.


You'll notice that above, the win % for 90-100 is at 50%. In week 3, the 90.03% favorite NE lost to MIA. 0.03% is just about a difference of 1 game in the entire set of hundreds and hundreds of simulated games. Had NE been an 89.97% favorite, the graph would look like this:


Check out the 'Wins' line. The Wins line overlaps the theoretical line. You can't even see it. That's absolutely absurd, especially after 100 games. The trendline has a slope of .09, compared to the theoretical line's .1. The "Wins" trendline has an R-squared value (a measure of how closely the data points fit the line) of 0.89. Absolutely insane. In general, I try to temper my enthusiasm, but this is unbelievable... This means that when NFLSim says a team will win 63% of the time, that team will win 63% of the time. For those of you math-minded people, the expected number of wins is approximately 64.15. The actual number of wins is 65. That blows my mind.

Here's a team-by-team accuracy breakdown:

	Win %	Cover Spread %	Over %
ARI	83.3%	40.0%	50.0%
ATL	66.7%	50.0%	33.3%
BAL	57.1%	42.9%	71.4%
BUF	66.7%	50.0%	66.7%
CAR	83.3%	40.0%	50.0%
CHI	50.0%	50.0%	66.7%
CIN	71.4%	71.4%	33.3%
CLE	42.9%	42.9%	57.1%
DAL	57.1%	57.1%	57.1%
DEN	50.0%	33.3%	20.0%
DET	100.0%	66.7%	40.0%
GB	100.0%	66.7%	66.7%
HOU	71.4%	57.1%	85.7%
IND	33.3%	50.0%	33.3%
JAX	33.3%	66.7%	60.0%
KC	83.3%	66.7%	83.3%
MIA	33.3%	66.7%	50.0%
MIN	83.3%	66.7%	50.0%
NE	33.3%	20.0%	20.0%
NO	66.7%	33.3%	50.0%
NYG	66.7%	16.7%	20.0%
NYJ	71.4%	57.1%	57.1%
OAK	57.1%	14.3%	57.1%
PHI	66.7%	0.0%	66.7%
PIT	50.0%	66.7%	66.7%
SD	83.3%	50.0%	33.3%
SEA	83.3%	50.0%	33.3%
SF	50.0%	33.3%	40.0%
STL	57.1%	50.0%	28.6%
TB	71.4%	42.9%	57.1%
TEN	100.0%	66.7%	83.3%
WAS	42.9%	28.6%	33.3%

I'll post some more stats if I have a chance.
Enjoy!

I'd love to hear everyone's reactions and questions, so don't be shy, send me some emails.