Jay Edgar NBA, Report on 5%-OFF method @ the Break

Jay Edgar · 02-18-07, 12:07 AM

Here are the results by %OFF
(comparing the Edgar ML and the median of the closing Pinny ML, what is the difference in what each line says about the play-on team’s predicted chance to win the game?)

ML RESULTS
%OFF/W-L/net$
5% 19-13 +1895
6% 18-28 +194
7% 9-18 -354
8% 24-22 +1530
9% 15-31 -85

10% 11-19 +874
11% 15-22 +744
12% 11-20 +234
13% 7-18 -213
14% 14-19 +735

15% 5-17 -851
16% 10-21 -127
17% 11-14 +862
18% 7-10 +738
19% 2-13 -139

20% 1-11 -1035
21% 4-9 -190
22% 8-10 +1677
23% 3-8 +500
24% 3-3 +993

25% 3-4 +107
26% 0-7 -700
27% 1-5 -230
28% 1-3 -140
29% 1-4 -300
30% 2-2 +255

31% 0-3 -316
32% 2-2 +333
33% 0-2 -200
34% 1-0 +315
35% 0-2 -200
36% 0-2 -200
37% 1-1 +65
38% 0-1 -100
43% 0-1 -100

Grouping the ML results
5-9% 85-112 +3180
10-14% 58-98 +2374
15-19% 35-75 +483
20-24% 19-41 +1945
25-30% 8-25 -1008
31%+ 4-14 -403

Same analysis for ATS results
%off/W-L/pct/net$
5% 22-10 .688 +1150
6% 25-21 .543 +295
7% 15-11 .577 +345
8% 30-16 .652 +1320
9% 26-20 .565 +500

10% 15-14 .517 +30
11% 24-13 .649 +1035
12% 17-14 .548 +230
13% 12-13 .480 -165
14% 21-12 .636 +840

15% 10-11 .476 -155
16% 15-14 .517 +30
17% 14-10 .583 +350
18% 8-9 .471 -145
19% 6-8 .429 -240

20% 2-10 .167 -850
21% 6-7 .462 -135
22% 9-9 .500 -45
23% 3-7 .300 -435
24% 4-2 .667 +190

25% 3-4 .429 -120
26% 2-4 .333 -220
28% 1-3 .250 -215
29% 2-3 .400 -115
30% 2-2 .500 -10

31% 2-1 .667 +95
32% 3-1 .750 +195
33% 1-1 .500 -5
34% 1-0 1.000 +100
35% 1-1 .500 -5
36% 1-1 .500 -5
37% 1-1 .500 -5
38% 1-0 1.000 +100
43% 1-0 1.000 +100

Grouping the ATS results
%off/W-L/pct/net$
5-9% 118-78 .602 +3610
10-14% 89-66 .574 +1970
15-19% 53-52 .505 -160
20-24% 24-35 .407 -1275
25-30% 12-20 .375 -900
31%+ 12-6 .667 +570

NET results for ML and ATS plays combined
5% +3045
6% +489
7% -9
8% +2850
9% +415

10% +904
11% +1779
12% +464
13% -378
14% +1575

15% -1006
16% -97
17% +1212
18% +593
19% -379

20% -1885
21% -325
22% +1632
23% +65
24% +1183

25% -13
26% -920
27% -450
28% -355
29% -415
30% +245

31% -221
32% +528
33% -205
34% +415
35% -205
36% -205
37% +60
38% 0
43% 0

Grouping the net results for ML and ATS combined
5-9% +6790
10-14% +4344
15-19% +323
20-24% +670
25-30% -1908
31%+ +167

COMMENT: Recently, there have been some good moneyline winners in that 20-24% off range. In fact, the moneyline plays over 20% are now unambiguously positive for the entire year.

Still, you can see that, for whatever reason, the best plays are the ones where the Edgar line is off between 5% and 15%. That has been true all year. The precise spot where a retrofit analysis would draw the “tread-carefully” line has moved around a bit during the year: generally, it remains somewhere between 15 and 25%. But more importantly, I think, the year-long results are establishing that it is a dotted tread-carefully line, not a hard-and-fast one. Obeying it isn’t really a necessity in the overall scheme of things. Blindly playing everything at 5% or more hasn’t hurt anybody – indeed, the net drag on performance from the highest %-OFF plays has been relatively minimal overall.

bigboydan · 02-18-07, 12:40 AM

Originally posted by Jay Edgar

COMMENT: So, are we slowing down? Well, last week was the worst week of all 16 weeks to date. (But it was also the shortest week.)

Jay, The same thing happens in baseball really. roughly 10 days before the all-star break. (the 20+ years trend) for both hoops and bases.

I look for the same thing to happen this year in hoops like years past. When you get to the final month of the season (stretch drive/teams jockeying playoff position) then thats when you really need to use caution. Then once playoff hopes start to die off for some teams, then you jump back in.

Just something you might wanna consider tracking for next years purposes.

Jay Edgar · 02-18-07, 01:05 AM

I'll certainly post the lines and track them from here until the end of the year, breaking them down every possible way. It'll be informative if nothing else.

Tomorrow I'll run through the YTD records by team. Both play-on teams and play-against teams. To me that's one of the more enlightening breakdowns.

Also we'll do breakdowns by closing pointspread, by gaps between the play-on and play-against teams in my rankings, by how much rest the teams have, and by the days of the week -- all the old favorites.

bigboydan · 02-18-07, 01:47 AM

Can't wait to see your findings Jay.

Dark Horse · 02-18-07, 11:41 AM

Great stuff.

Just a question. When do you pull the trigger on a bet? As soon as it comes within the 5% range, or is that based on the closing line?

And do you play both the ML and ATS when they qualify?

trustbutverify · 02-18-07, 12:13 PM

excellent work Jay. Does your spreadsheet allow you to run simulated bets through the entire season?

Jay Edgar · 02-18-07, 01:28 PM

DH-
Grading's based on the closing line, but for myself I look for the best number whenever it's available, including overnights on the ATS side sometimes. (Noodling around at Matchbook and working line movement has been great for getting ATS plays at far better than the -105. More like +110 much of the time.) And I am playing both ML and ATS pretty religiously.

TBV-
I'm not sure I understand your question 100%, but I think the answer's no. I have a pretty simple Excel spreadsheet working, and I'm at my limit on Excel proficiency. But I'm interested to hear what else you'd want a spreadsheet to do and why.

Jay Edgar · 02-18-07, 02:14 PM

PLAY-ON AND PLAY-AGAINST TEAMS
No surprise, the 5%-OFF method regularly taps the league’s ugly ducklings for play-on, and the league’s most popular teams for play-against.

PLAY-ON

Here’s how often each team has been a play-on team, and the results.

[Note that while there was finally a play-on PHX spot last week (and a pretty bad one), there has still never been a SAN play-on. That is, still never a game this year where the real line underestimated the Spurs’ current capability as I figured it. In fact, I don’t think it’s ever even been close. And by the way the only two Dallas play-ons have been against – that’s right – the San Antonio Spurs.]

PLAY-ON TEAMS, MONEYLINE
#PLAYS/TEAM/NET$
36 ATL 2380
37 NY 2093
11 UTH 1847
34 NOK 1763
40 POR 1618
39 CHA 1300
10 LAX 1011
14 WAS 744
6 CLE 452
2 DAL 418
21 MIN 405
6 CHI 247
11 ORL 242
14 HOU 159
5 DET 91
25 TOR 80
15 MIA 24
10 DEN 0
0 SAN 0
18 GS -8
1 PHX -100
20 NJ -142
15 SAC -312
35 PHI -602
13 IND -673
10 LAC -834
36 MEM -1046
28 SEA -1283
28 MIL -1584
34 BOS -1719

PLAY-ON TEAMS, POINTSPREAD
#PLAYS/TEAM/NET$
36 ATL 2380
37 NY 2093
11 UTH 1847
34 NOK 1763
40 POR 1618
39 CHA 1300
10 LAX 1011
14 WAS 744
6 CLE 452
2 DAL 418
21 MIN 405
6 CHI 247
11 ORL 242
14 HOU 159
5 DET 91
25 TOR 80
15 MIA 24
10 DEN 0
0 SAN 0
18 GS -8
1 PHX -100
20 NJ -142
15 SAC -312
35 PHI -602
13 IND -673
10 LAC -834
36 MEM -1046
28 SEA -1283
28 MIL -1584
34 BOS -1719

PLAY-ON TEAMS, TOTAL NET RESULTS
RANK/TEAM/NET$
1 NY 3333
2 UTH 2742
3 NOK 2703
4 ATL 2700
5 POR 1723
6 LAX 1501
7 MIN 1070
8 TOR 940
9 MIA 709
10 WAS 709
11 CLE 642
12 DAL 618
13 CHA 590
14 CHI 437
15 NJ 423
16 ORL 317
17 DET 181
18 DEN 180
19 GS 152
20 SAN 0
21 HOU -181
22 PHX -205
23 SAC -452
24 IND -703
25 PHI -1197
26 SEA -1453
27 MEM -1546
28 LAC -1679
29 BOS -1804
30 MIL -2064

PLAY AGAINST

Now the same three charts for play-against teams. The subtitle here could be “if only we never had to mess with the Dallas Mavericks . . . “

PLAY-AGAINST TEAMS, MONEYLINE
#PLAYS/TEAM/NET$
23 LAX 1905
24 DEN 1667
19 MIN 1271
23 ORL 1232
45 SAN 1211
27 UTH 1176
20 NJ 961
7 MEM 716
30 CLE 714
5 ATL 613
4 CHA 518
9 SEA 451
7 NOK 412
3 POR 357
18 SAC 134
32 CHI 107
8 PHI 35
24 LAC 22
4 MIL 0
4 NY -108
7 BOS -273
20 WAS -323
25 IND -382
32 DET -420
21 MIA -519
20 HOU -559
17 GS -575
10 TOR -610
43 PHX -674
43 DAL -2488

PLAY-AGAINST TEAMS, POINTSPREAD
#PLAYS/TEAM/NET$
14 MIN 875
16 LAX 865
15 ORL 660
16 UTH 550
19 CHI 535
5 ATL 500
14 LAC 455
4 CHA 400
12 NJ 360
17 CLE 335
3 POR 300
5 MEM 290
3 MIL 195
10 SAC 160
17 DET 125
4 NOK 85
5 SEA 80
10 HOU 55
10 WAS 55
2 NY -10
4 PHI -20
12 DEN -60
3 BOS -120
10 MIA -155
21 PHX -210
22 SAN -215
7 GS -350
3 TOR -435
16 DAL -710
9 IND -780

PLAY-AGAINST TEAMS, TOTAL NET RESULTS
RANK/TEAM/NET$
1 LAX 2770
2 MIN 2146
3 ORL 1892
4 UTH 1726
5 DEN 1607
6 NJ 1321
7 ATL 1113
8 CLE 1049
9 MEM 1006
10 SAN 996
11 CHA 918
12 POR 657
13 CHI 642
14 SEA 531
15 NOK 497
16 LAC 477
17 SAC 294
18 MIL 195
19 PHI 15
20 NY -118
21 WAS -268
22 DET -295
23 BOS -393
24 HOU -504
25 MIA -674
26 PHX -884
27 GS -925
28 TOR -1045
29 IND -1162
30 DAL -3198

OTHER CHARTS

Just a couple more charts on play-on/play-against.

First, the net results chart for all plays (for or against) involving any particular team. For whatever reason Utah, the Lakers, and Atlanta have really performed as the Edgar Line predicts game-by-game (although on the Hawks I must say I can remember at least four horrible efforts at home that cost some money. But they have more than made up for it on the road.).

ALL RESULTS
RANK/TEAM/NET$
1 UTH 4468
2 LAX 4271
3 ATL 3813
4 MIN 3216
5 NY 3215
6 NOK 3200
7 POR 2380
8 ORL 2209
9 DEN 1787
10 NJ 1744
11 CLE 1691
12 CHA 1508
13 CHI 1079
14 SAN 996
15 WAS 441
16 MIA 35
17 TOR -105
18 DET -114
19 SAC -158
20 MEM -540
21 HOU -685
22 GS -773
23 SEA -922
24 PHX -1089
25 PHI -1182
26 LAC -1202
27 IND -1865
28 MIL -1869
29 BOS -2197
30 DAL -2580

And finally, a new chart of PUBLIC TEAMS. If every play-against situation means a public vote for that team, and every play-on situation means a public-vote against that them, then the most popular teams are:

PUBLIC TEAMS
RANK/TEAM/PA-PO/NET
1 SAN 0-45 45
2 PHX 1-43 42
3 DAL 2-43 41
4 DET 5-32 27
5 CHI 6-32 26
6 CLE 6-30 24
7 UTH 11-27 16
8 DEN 10-24 14
9 LAC 10-24 14
10 LAX 10-23 13
11 IND 13-25 12
12 ORL 11-23 12
13 HOU 14-20 6
14 MIA 15-21 6
15 WAS 14-20 6
16 SAC 15-18 3
17 NJ 20-20 0
18 GS 18-17 -1
19 MIN 21-19 -2
20 TOR 25-10 -15
21 SEA 28-9 -19
22 MIL 28-4 -24
23 BOS 34-7 -27
24 NOK 34-7 -27
25 PHI 35-8 -27
26 MEM 36-7 -29
27 ATL 36-5 -31
28 NY 37-4 -33
29 CHA 39-4 -35
30 POR 40-3 -37

Looks right to me.

Jay Edgar · 02-18-07, 02:51 PM

DAYS OF THE WEEK AND DAYS-OF-REST CHARTS

Two sets of charts that have some relation to each other. First, the ever-popular day-of-the-week chart.

PLAYS BY DAY OF WEEK

MONEYLINE PLAYS
Wed 54-78 +4979
Fri 53-74 +3115
Thu 8-11 +444
Mon 24-47 +68
Sun 19-30 -41
Tue 19-52 -871
Sat 32-73 -1123

POINTSPREAD PLAYS
Thu 13-4 (.765) +880
Fri 72-53 (.576) +1635
Wed 73-56 (.566) +1420
Sat 57-48 (.543) +660
Sun 24-24 (.500) -120
Tue 35-35 (.500) -175
Mon 34-37 (.479) -485

ALL PLAYS
Wed +6399
Fri +4750
Thu +1324
Sun -161
Mon -417
Sat -463
Tue -1046

Wednesdays and Fridays are best, for some reason. (Day of practice beforehand usually? Days when team is most likely in that every-other-day rhythm?) Also interesting that Thursdays – the national TV night that most teams presumably get a little more ‘up’ for-- have been strong on a per-play basis.

DAYS OF REST

Saturday nights have been poor for these plays. And of course the most common back-to-back night is Saturday. (Though of course that is true for both play-on and play-against teams.)

Still, the charts below indicate that the Edgar Line is less reliable when the play-on team has played the night before.

(KEY: for example, “1-on-3” means the play-on team last played one day ago and the play-against team last played 3 days ago.)

MONEYLINE PLAYS
1 ON 1 284
1 ON 2 -930
1 ON 3 -1068
1 ON 4 +1047

2 ON 1 +540
2 ON 2 +5483
2 ON 3 +1120
2 ON 4+ +692

3 ON 1 -709
3 ON 2 +671
3 ON 3 -641
3 ON 4+ -200

4+ ON 1 +209
4+ ON 2 +78
4+ ON 3 -110
4+ ON 4+ +105

ALL “1 ON” -667
ALL “2 ON” +7835
ALL “3 ON” -879
ALL “4+ ON” +282

POINTSPREAD PLAYS
1 ON 1 +170
1 ON 2 -1170
1 ON 3 +40
1 ON 4+ -30

2 ON 1 +610
2 ON 2 +3105
2 ON 3 +15
2 ON 4+ +70

3 ON 1 -350
3 ON 2 +600
3 ON 3 -265
3 ON 4+ +200

4+ ON 1 +685
4+ ON 2 -130
4+ ON 3 -15
4+ ON 4+ +280

ALL “1 ON” -990
ALL “2 ON” +3800
ALL “3 ON” +185
ALL “4+ ON” +820

ALL PLAYS
1 ON 1 +454
1 ON 2 -2100
1 ON 3 -1028
1 ON 4+ +1017

2 ON 1 +1150
2 ON 2 +8588
2 ON 3 +1135
2 ON 4+ +762

3 ON 1 -1059
3 ON 2 +1271
3 ON 3 -906
3 ON 4+ 0

4+ ON 1 +894
4+ ON 2 -52
4+ ON 3 -125
4+ ON 4+ +385

ALL “1 ON” -1657
ALL “2 ON” +11635
ALL “3 ON” -694
ALL “4+ ON” +1102

The HG · 02-18-07, 05:39 PM

Originally posted by Jay Edgar

last week was the worst week of all 16 weeks to date.

Interesting, could there perhaps be a seasonal ebb and flow to the strength of your measurements? In other words, could there perhaps be certain times of the year where your measurements are particularly strong or not quite as strong as others? Maybe the uniqueness of the week before the All-Star Break throws a minor monkey-wrench into your numbers?

Jay Edgar · 02-18-07, 05:49 PM

Possible.

It's not at my fingertips to check, but I'd bet that for all games last week the total distance between all these predicted results and all actual results wasn't too much greater than the full-season average. We just fell a point or two short in some individual games in the first half of the week, and then there was no second half of the week. At least that's how it felt.

But we'll see what the next few weeks bring. As I say, it would not surprise me if the high water mark for the year is already in the rear view mirror at +124. Hope not, though.

bigboydan · 02-18-07, 11:29 PM

Hey Jay,

During that last month of the regular season. Make a note of my comments above and track that, because I'm quite curious to see how that effects your numbers.

bside · 02-18-07, 11:31 PM

I don't know Jay it sure felt like Wednesday all the home teams significantly out performed expectations. If I recall only two dogs covered. I used to have a hard and fast rule about no plays the week before or after all star breaks. Got a little lax this year and paid the piper. Down 12 units in NBA on a small slate of games last week. I think I will hold off next week unless I see something I really like.

Jay Edgar · 02-18-07, 11:45 PM

I agree about last Wednesday, bside. Home teams were 9-4 ATS and it felt more one-sided than that. That's the night of the year with the biggest mail-in danger, I think.

This coming week it's not an effort issue but rather that the break seems not only to interrupt momentum but reverse it. Thus it wouldn't surprise me if some combo of UTH, DET, CLE, MIA, TOR, ATL, SEA and even DAL (the teams that had it going) struggle, while some combo of the badly struggling teams (BOS, LAC, CHI, GS, WAS, POR, PHI, MIL, LAX) get hot immediately.

I like the method's primary tool (last ten games, weighted for recency) because it begins reflecting a team's change of direction almost (almost) right away.

Jay Edgar · 02-21-07, 10:39 AM

One more report/chart that might answer some questions and set expectations appropriately for the rest of the year:

If my method were a stock, here is the chart for this NBA season. Assume you started with 100 units and played one unit on each play. This would be your bankroll.

30-Oct 100.00
31-Oct 104.15
1-Nov 113.61
3-Nov 121.97
4-Nov 123.00
5-Nov 125.54
6-Nov 124.59
7-Nov 126.54
8-Nov 135.38
10-Nov 131.42
11-Nov 128.07
12-Nov 122.86
13-Nov 118.68
14-Nov 114.35
15-Nov 130.13
17-Nov 127.94
18-Nov 121.20
19-Nov 119.15
21-Nov 119.39
22-Nov 123.08
24-Nov 136.43
25-Nov 128.02
26-Nov 125.92
27-Nov 132.00
28-Nov 139.25
29-Nov 140.69
30-Nov 140.69
1-Dec 132.28
2-Dec 132.40
3-Dec 137.33
4-Dec 131.65
5-Dec 134.05
6-Dec 141.45
7-Dec 147.90
8-Dec 133.52
9-Dec 133.72
10-Dec 136.34
11-Dec 131.52
12-Dec 127.28
13-Dec 127.26
14-Dec 127.29
15-Dec 136.96
16-Dec 137.45
17-Dec 142.58
18-Dec 143.21
19-Dec 141.16
20-Dec 144.27
21-Dec 146.91
22-Dec 163.64
23-Dec 155.39
25-Dec 157.44
26-Dec 154.87
27-Dec 154.55
1-Jan 155.95
2-Jan 155.85
3-Jan 158.52
4-Jan 155.47
5-Jan 155.80
6-Jan 162.04
7-Jan 155.97
8-Jan 155.57
9-Jan 151.82
10-Jan 149.24
11-Jan 150.89
12-Jan 157.62
13-Jan 157.24
14-Jan 157.24
15-Jan 162.04
16-Jan 163.64
17-Jan 168.14
18-Jan 166.09
19-Jan 164.27
20-Jan 169.06
21-Jan 163.96
22-Jan 159.86
23-Jan 149.61
24-Jan 171.34
25-Jan 173.89
26-Jan 190.40
27-Jan 195.30
28-Jan 186.85
29-Jan 199.81
30-Jan 199.76
31-Jan 203.87
2-Feb 197.28
3-Feb 208.66
4-Feb 211.96
5-Feb 209.91
6-Feb 207.82
7-Feb 202.04
8-Feb 204.04
9-Feb 219.21
10-Feb 210.56
11-Feb 217.36
12-Feb 211.21
13-Feb 210.88
14-Feb 200.84
15-Feb 203.86
20-Feb 197.66

Plainly, there have been ups and downs, with a high water mark of +119 units on Feb 9th -- which capped a pretty stunning 17-day run from January 23, when the YTD figure had dipped below +50. So in those 17 days the thing was an amazing +70 units. The rest of the year it has been grinding out an additional proifit fairly steadily, in amounts that are adding up recently to between 25 and 50 units overall.

So, as you can see, jumping on board and playing everything for a few days in a row has been very profitable sometimes, neutral sometimes, and very not profitable sometimes.

It is just a feeling, but I am not especially optimistic about what the rest of February and March will bring. But we'll see.

Bottom line: I'm not posting these threads to tout anything. If anyone's looking for a tout telling them what to do, I'm not the guy.

I have been posting the threads for two reasons: (1) the discipline of having to post keeps me on top of it; and (2) I think my basic information -- the power ratings, the lines on each game, the comments that the thread generates -- can be extremely useful for somebody figuring out their own approach to the NBA. IMHO, they are as useful as anything else you will find on the internet. And the price is right.

bigboydan · 02-21-07, 11:40 AM

Originally posted by Jay Edgar

Bottom line: I'm not posting these threads to tout anything. If anyone's looking for a tout telling them what to do, I'm not the guy.

Seems like anyone that has a hot streak on these types of forums turns tout at some point. I'm glad to hear your not one of those types of guys Jay.

Jay Edgar · 02-21-07, 12:03 PM

I'm an amateur and always will be. For better or worse, no matter how profitable the hobby might get I have a real job that would make it phenomenally silly for me to either: (1) spend any more time at this than I already do; or (2) change my country of residence.