[Pensinger1]

I've been buying hooks/points in the NBA under the assumption that 2, 4, 6 and 8 are all key numbers.

As of late, I've heard some talk about 7 being a key number in the NBA, such that teams will stop fouling when they are down by 7 because the game has reached 3 possessions, leaving the opposition to run out the clock with ball in hand.

7 is a Key Number in the NBA.

[_] Agree
[_] Disagree

Thanks.

[Peep]

Agree.

[Ominous]

There are no key numbers in NBA....

the +-1 is less valueble but other than that they are all about the same

[Peep]

the +-1 is less valueble but other than that they are all about the same

Why do you think that? Gauch posted his list here a while ago of what numbers fall the most, I don't think they are "all about the same". BTW, the +/- 2 and 3 or no hell either. And I wouldn't recommend the +/-16 for frequency either......

[G's pks]

disagree...last week b. gordon(bulls) was fouled on a 3 he hit and got 4pts, so in a case like that a second three on a steal or after a rebound would mean 7 on two possessions. Or something like what happened in the Arz st/usc game where usc coach Floyd blew a gasket on a foul and turned two ft's into 4... thanks to 2 t's...

[Ganchrow]

7 is a Key Number in the NBA.

[_] Agree
[_] Disagree

In general, realized historical frequencies should not be answerable to referendum.

That said using Covers data from the 1990/91 season through February 12th of the 2008/09 season, push frequencies and associated standard errors given a spread radius of 2 follow:

Spread	Push %	Std. Err.
1	2.33%	0.18%
2	3.87%	0.21%
3	3.78%	0.19%
4	3.49%	0.18%
5	4.33%	0.20%
6	4.06%	0.20%
7	4.12%	0.21%
8	4.16%	0.23%
9	4.61%	0.26%
10	4.17%	0.27%
11	4.15%	0.31%
12	3.46%	0.32%
13	3.83%	0.39%
14	3.41%	0.42%
15	3.01%	0.47%
16	5.14%	0.75%
17	3.42%	0.79%

[Peep]

I like it better when we vote on what is real or not.

But I guess there is something to be said for facts.

Over 11,308 NBA games I have on file, by home team.

1) Home team wins by one 224 times.
2) Home team wins by two 368 times.
3) Home team wins by three 366 times.
4) Home team wins by four 344 times.
5) Home team wins by five 403 times.
6) Home team wins by six 388 times.
7) Home team wins by seven 430 times.
8) Home team wins by eight 401 times.
9) Home team wins by nine 384 times.
10) Home team wins by ten 348 times.
11) Home team wins by eleven 323 times.
12) Home team wins by tweleve 304 times.

Not enough that I would buy the half point, but enough that I pay attention, especially when looking at ML's for 2nd half.

[roasthawg]

5 and 7 are the two most likely numbers to be the final margin of victory in an NBA game. Ganch's list is probably more useful than that fact though because it takes into account the spread.

[donjuan]

Not enough that I would buy the half point, but enough that I pay attention, especially when looking at ML's for 2nd half.

How often all games end on certain numbers is not particularly relevant, especially in a sport that essentially has a normal distribution of outcomes. What is really important is how often push frequencies occur.

[Data]

How often all games end on certain numbers is not particularly relevant, only in a sport that essentially has a normal distribution of outcomes, where the closing line is the distribution mean. What is really important is how often push frequencies occur.

Nonsense corrected. If the distribution is non-normal, the MOVs are very important, e.g. baseball.

[donjuan]

Nonsense corrected. If the distribution is non-normal, the MOVs are very important, e.g. baseball.

Feel free to explain yourself.

[Data]

Feel free to explain yourself.

I was a bit disappopinted that instead of finally strarting contributing to this forum, which you seem to be long overdue, you switched from posting obvious to posting nonsense.

[Peep]

Since we are "explaining ourselves", I'll explain myself too.

When I am am betting a second half, and for example the team I want to bet is ahead (or down) by two, and I have a choice of +5 or ML +300, which one I will bet is largely determined by the chances of the bet hitting the final number. For example if the team is down two, I will usually opt for the moneyline, as the chances of the score "stopping at three" are not that high. They will usually either lose big or win the second half outright.

If they are up two, I probably take the points, as six and seven are live numbers.

[donjuan]

I was a bit disappopinted that instead of finally strarting contributing to this forum, which you seem to be long overdue, you switched from posting obvious to posting nonsense.

That hard to explain, huh?

[donjuan]

When I am am betting a second half, and for example the team I want to bet is ahead (or down) by two, and I have a choice of +5 or ML +300, which one I will bet is largely determined by the chances of the bet hitting the final number. For example if the team is down two, I will usually opt for the moneyline, as the chances of the score "stopping at three" are not that high. They will usually either lose big or win the second half outright.

There are factors you are ignoring by just looking at how often teams win by various numbers without regard to what the spread for those games were. I'd be much more inclined to use push percentages if you have them available. That said, I understand there are sample size issues for 2h push percentages.

[Peep]

Yes, even with 11K games, the sample sizes are much too small to look for push percentages.

[Peep]

Tonight is a good example. GS up by 6 at half of game, as 5 point dogs.

I can either get six points or +250 (I want to play GS either way). So I take the +250 since the inside numbers, i.e GS winning by 1-5 for game, are deader than outside numbers.

I am posting this before half goes off, so "we shall see". I think either lakers blow them out or they cover ML.

[SportsLockPicks]

totally disagree...there are a lot of reasons, mainly that there are only 3 increments of scoring and that one score can only put a team up a max of 3 points...

if there were any swing number it would be to buy down from 3.5

[Dark Horse]

I can hardly wait for proof that you should buy on/off the 7 and 3.



I don't think there's a white rabbit in this hole, but good luck chasing it anyway.

[smoke a bowl]

In general, realized historical frequencies should not be answerable to referendum.

That said using Covers data from the 1990/91 season through February 12th of the 2008/09 season, push frequencies and associated standard errors given a spread radius of 2 follow:

Spread    Push %    Std. Err.
1    2.33%    0.18%
2    3.87%    0.21%
3    3.78%    0.19%
4    3.49%    0.18%
5    4.33%    0.20%
6    4.06%    0.20%
7    4.12%    0.21%
8    4.16%    0.23%
9    4.61%    0.26%
10    4.17%    0.27%
11    4.15%    0.31%
12    3.46%    0.32%
13    3.83%    0.39%
14    3.41%    0.42%
15    3.01%    0.47%
16    5.14%    0.75%
17    3.42%    0.79%

Ganchrow, what do you think the logic is behind the 5 being worth almost a full % more than the 4? Obviously that is the case judging from the data (factoring in std. err. the difference could be as low as .5%) but what is the theory behind it? What makes the 5 so strong? I understand why the 7,8,9 are so strong(teams tend to stop fouling at that point) but can't figure the logic on the 5.Thx in advance.

Sidenote: What makes the 16 so much stronger than the 15 or 17? Even factoring in the Std. err. for the 15 and 16 according to the data above, the worst case for the 16 would be that it is still a full percent stronger than the 15. That makes very little sense to me but I'm sure I could be missing something. Any help would be greatly appreciated.

[smoke a bowl]

FWIW, my data only goes from 1997 - 2008 and it has the same bias for the 16 when running spreads of 12 or more as the chart indicates. I get 480 samples with the 16 pushing 25 times (5.2%) and the 15 pushing 14 times (2.9%). Weird but my only conclusion is that the standard error should be higher(I don't know how to calculate std err so just a pure guess) or this is just a rare case where #s just defy logic.

[smoke a bowl]

Though I addressed my post above to Ganchrow, I welcome anyone's help here as I respect most that post here in the tank immensely.

[Peep]

I have 796 games with favorites of 12 points or more.

Out of these games, 25 fell on "15" for final score differential and 43 fell on "16".

Does that answer your question?

[smoke a bowl]

I have 796 games with favorites of 12 points or more.

Out of these games, 25 fell on "15" for final score differential and 43 fell on "16".

Does that answer your question?

No Peep, I've gotten that far but thx for trying. I want to know why the 16 hits so much more than the 15 does.What would be the theory behind it because they seem like they should be worth the same amount.

[Peep]

All I can tell you is data falls like that Smoke a bowl, sometimes you can apply a "logical" reason for it, sometimes no.

In my college bk db, the 8 pt favorites our preform the nine pt favorites in many catagories. That is with a thousand samples or so. There is no reason for this possible IMO, I have gone over it, so I just try and "normalize" the data as best I can.

[smoke a bowl]

All I can tell you is data falls like that Smoke a bowl, sometimes you can apply a "logical" reason for it, sometimes no.

In my college bk db, the 8 pt favorites our preform the nine pt favorites in many catagories. That is with a thousand samples or so. There is no reason for this possible IMO, I have gone over it, so I just try and "normalize" the data as best I can.

Right, we are definitely on the same page here. I'm actually more curious about the difference in the 4 and the 5 in the NBA. Ganchrow's push rate chart indicates that the 5 is worth a good bit more (almost a full 1%) than the 4. I know that it is correct but was curious to what the logic is behind it. I'm pretty sure there is some reasonable logic behind this difference (not like the 15 and 16 which I think is just a product of sample size more than anything else) but can't figure it out myself.

[Peep]

Yes, I was thinking about the 4 and 5 too smoke a bowl.

Both are two score situations. If you are behind 4 and under 24, you foul, behind 5 you foul. The "four" then goes to "five" or "six", the foul at behind 5 goes to "six" or seven". So it might be it is not as likely to go back to a four, as the team behind will wildly throw up a three or two (probably).

What do you think?

[smoke a bowl]

Yes, I was thinking about the 4 and 5 too smoke a bowl.

Both are two score situations. If you are behind 4 and under 24, you foul, behind 5 you foul. The "four" then goes to "five" or "six", the foul at behind 5 goes to "six" or seven". So it might be it is not as likely to go back to a four, as the team behind will wildly throw up a three or two (probably).

What do you think?

I'm pretty sure that's the right idea. I'll do some research and talk to some people and report back when I come up with.

[Data]

Let's think what was the last score. If the last score was from the losing team then there should not be much of a difference in percentages of 4 and 5 (this is just a guess but it seems to be a reasonable). So, what if the last score was by the winning team? Here is another guess. Again, it seems reasonable to assume that the majority of the last scores when the games ended either on 4 or on 5 were made free throws.

If a team is down by 2 there is no reason for them to rush the shot. They will go for a buzzer beater. If they are down by 3 then they will likely rush and the if they miss they foul. Note, if they do not miss or if they miss a buzzer beater, the game will neither end on 4 nor on 5. So, they fouled being down by three. Since the average FT% is 0.76, the score difference will change from 3 to 5 1.5 times more often than it will change to 4.

Lastly, if we assume 50/50 probability of what team scored the last then 5 is a 2.5 to 2 favorite over 4, or 1.25 to 1. Note that 4.33%/3.49%=1.24.

[Madetowin]

I think only in the Bible

[Pensinger1]

Hmmm. Good stuff guys.

Maybe I should have titled differently... "Do Key Numbers exist in the NBA?"

Or maybe, "Are buying hooks/points in the NBA -E.V. in the long run?"

[Peep]

Interesting point about last score Data.

Can you parlay that bet with anything? I know some books offer "who scores last"?

[Data]

Can you parlay that bet with anything? I know some books offer "who scores last"?

No idea, the props are not my cup of tea.