NBA Math

[apk2k6]

I don't know if this has been posted here or not, but I'm sharing an NBA Math system that I found. Apparently it produces around 100 plays in an NBA season, and has been hitting 70% pretty consistently. It doesn't start until 20 games have been played into the season, but it's 4-1 so far before the 20 game mark and there's another play tonight.


And here is the formula types exactly as it was given to me:

Example : Clev vs denver
Cle winning % is .600 and Denver is .057

Subtract the difference

Clev 600 - Den 057 = 543

Give 1 point for every 20% points Therefore 543/20 = 27, round to the nearest tenth

Home team is given 3 points, away team you subtract 3 points.

Subtract or add the difference
Clev is 27 and they are away, thus 27- 3 = 24

Then subtract or add point spread for the final number value

The spread for the game is Cle -9.5 points, thus 24 - 9.5 = 14.5

The final number must be 10 or greater for a PLAY. In this example, Clev is the PLAY.

If the spread is 10 or greater , do not play. If the number is less than or equal to -1, then you play the other team ATS.

FILTERS:

- Don't play if selected team played the night before (No back to back games)
- Dont play of one or more starters are out. Allow one week for return starters
- Don't play 1st 20 games of the season or 1st 3 games after all star break

[apk2k6]

Example for the game tonight: I use the team with the best winning % as the "comparison" team.

NOH is 7-5(.583), and Clippers are 2-11(.153) You have to get rid of the decimal since we will be dividing by 20. So 583-153 = 430. 430/20 = 21.5

Since Hornets are away, you subtract 3 points, giving you 18.5. The final step is to add/subtract the spread for the game. Since the spread for NOH is -5.5, you subtract.

18.5 - 5.5 = 13

This is of course greater than 10, so the play is the Hornets -5.5. They did not play last night, none of their starters are out, and they spread is less than -10.

Let me know if you have any questions, I will be tracking this all season.

Here is from the original posting: The record for 1999 was 60-25 and for 2000 it was 73-30

[apk2k6]

I use the team with the greatest winning percentage like I said earlier, but I'm still monitoring road favorites to see how they do. Hornets fall under this tonight.

[BeanSaunce]

I just started out, i mean totally fresh into this, but i am gonna try and use this math. hopefully i begin to understand it soon..

[roasthawg]

This one screams "regression to the mean" to me...it may have got lucky for awhile and hit at an amazing clip but I would not expect that to continue.

[Justin7]

Interesting. does your system look at injuries?

[UntilTheNDofTimE]

he said all starters have to be active so yes it counts injuries

[ipickwinners]

a lot of shit to do to get a winner, ill do what i do and look at matchupsand continue to make a profit in nba my way but thanks for the system

[apk2k6]

Ya it's a lot, that's why I made a Java program where I just enter the records and spread and it tells me the play

It doesn't factor in the filters of course, I have to do that.

[apk2k6]

NOH -5.5 (winner)

System is 5-1 now.

Warriors are a play tomorrow night, as long as they are no higher than -1. They are +1 right now at the Greek, great deal.

[pirate]

where do you find the winning percentages?

[apk2k6]

I just look at their records and calculate it. I'm sure the % is somewhere online though.

[The_Kid]

Interesting stuff, apk. Would you mind keep a thread like this throughout the rest of the season to track it? I'd like to see how this works out.

[ChuteBoxe]

It did decent last year. I even have a spreadsheet for it, that automatically does the equation for you.

[Poker_Beast]

Looks like an interesting perspective. Keep the post going so we can follow!

Taking a shot on GSW tomorrow!

[apk2k6]

It did decent last year. I even have a spreadsheet for it, that automatically does the equation for you.

Ya I have the spreadsheet as well. I'll be glad to keep a thread, hopefully it keeps doing well!

[Iwinyourmoney]

APK, looking at this system it looks good. Any way to get that java program up for the non math people?

[FreeFall]

APK, looking at this system it looks good. Any way to get that java program up for the non math people?

I can write one, but don't you also need to know if the team is away or home as user input? That changes the output number by +3 or -3.

[mundane]

so u only do this for the favorites?

[MAKAVELI 96]

sounds interesting. do you have the record of this over the last few years?

[pico]

nice. i got a spreadsheet for NBA calculations...so far nothing seems to be working. glad yours is making you money

[apk2k6]

I don't have the record over the last few years, because someone just dug it back up last year. From late December to the end of the season I believe it went 20-5 last year. I also posted the records from 1999 and 2000.

Here's the java file:

http://www.filefactory.com/file/f25d29/n/NBAMath2_jar

Input the spreads as + or - so it knows. Example, if you put -6 as the home spread, the away spread must be input as +6.

[mundane]

im curious....

hmmm, so even if gsw gets to -2, it's still a play as the number will be 10!?!

and... u basically start with ur calculations starting with the higher number (better winning %) to get subtracted by the lower number? in this case, gsw as it has a
higher number? it cant be the other way around?

and for the filters, they only apply on one selected team and not for both? like gsw... no injury, not back to back. it doesnt really matter for the opposing team even if they played back to back or has injuries

[durito]

you should also look at teams that lost by exactly 13 points on a tuesday if they are playing a team whose name starts with a C on friday after 8pm eastern

[dcbt]

you should also look at teams that lost by exactly 13 points on a tuesday if they are playing a team whose name starts with a C on friday after 8pm eastern

Hah, that sounds like Norm Hitzges...

[mundane]

err how do u open that *.jar? im on vista

[ICE-BLOOD]

this system is gonna have some 20+ point picks on teams that have a real good or bad record (EXTREME HIGH OR LOW WINNING %)
example- okcity around .100
lakers .900

[MexicanStallion]

Wouldn't Phoenix be a play tonight too?

[mundane]

If the spread is 10 or greater , DO NOT PLAY!

spread is 10 on suns/thunder game.

[MexicanStallion]

If the spread is 10 or greater , DO NOT PLAY!

spread is 10 on suns/thunder game.

Correct, good point. Thanks

[mundane]

from what i could understand... here are the calculations!

11/25

gsw/wiz
5 win (selected team)
8 loss (selected team)
1 win (other team)
10 loss (other team)
-3 -3 (away), +3 (home)
-1.5 spread
10 PLAY!

cavs/knicks
10 win (selected team)
3 loss (selected team)
7 win (other team)
6 loss (other team)
-3 -3 (away), +3 (home)
-7 spread
2 NO PLAY!

suns/thunder
9 win (selected team)
5 loss (selected team)
1 win (other team)
13 loss (other team)
-3 -3 (away), +3 (home)
-9.5 spread
16 PLAY!

pacers/mavs
6 win (selected team)
7 loss (selected team)
5 win (other team)
7 loss (other team)
3 -3 (away), +3 (home)
-8 spread
-3 PLAY Other Team ATS

nets/lakers
too high of a spread! 13.5

so accdg to this system:
gsw -1.5 up to -2 is a play.
and if suns goes down to -9.5, it's also a play!

correct me if im wrong.

[zackattack]

NOH -5.5 (winner)

System is 5-1 now.

Warriors are a play tomorrow night, as long as they are no higher than -1. They are +1 right now at the Greek, great deal.

i did the math for warriors and i don't see why they aren't a play even at -2......my figures are as follows
gs wp is .385 -.091 = 294/20 = 14.7 -3 for away team is 12 minus spread of 2 is still 10 so why isn't this a play?

[mundane]

only up to -2! accdg to my calculation. if it goes up to -2.5, number would be 9 by then.

[mundane]

ooops.. nvm! josh howard...

Latest News:
Mavericks forward Josh Howard (sprained ankle) ran and participated in non-contact drills on Sunday, but remains day-to-day. Howard has missed two games with the injury so far...if he can't play on Tuesday it would leave him with a maximum of two games next week. Hopefully we'll have an update later on Monday, so check back soon.

[JGILL50]

how about a spread sheet that will show how every game does in relationship to the numbers. like 10 or higher is a play one way 1 or 2 might be good other way. just a thought. pretty cool stuff though