[bustabook]

for the next set of 1000 simulations, 12/7:

Simulation % (average of both %):

Denver -6.5 (55.1%) (56.4%) *1st set was w/spead @ -5.5
Denver/Philadelphia Under 217 (59.2%) (58.9%) *1st set was w/total @ 216.5
Portland -3 (71.8%) (70.9%) *1st set was w/spread @ -4
Portland/New York Under 201 (58.6%) (60.2%) *1st set was w/total @ 203
Oklahoma City -7.5 (71.6%) (70.5%) *1st set was w/spread @ -7
Oklahoma City Under 222 (70.2%) (70.7%) *1st set/2nd set total was the same (222)

i have an obligation tonight so i wont be able to get the 2nd set of the utah game posted by gametime.. sorry.

*note* will be editing as simulations finish

[suicidekings]

i can only backtest for this season, since the model runs on day-specific data and i dont know of any site which has this for past seasons.

if it is successful i will have to write another excel program to get day-specific data for past seasons.

--------------

as expected (with the exception of the Portland line), all lines are moving TOWARD my simulation lines.. i will be keeping a VERY close eye on the portland game, since there may be plays where the vegas lines move in the opposite direction of sim lines.

GL

The POR/NYK line is tainted by mising players and a JM (A) bet. This is one game that shouldn't factor into any changes / tweaks you make to the system. Same this is true of any line that has unexpected movement that is not a product of normal market activity.

I love what you're doing here. Good stuff.

[acarmelo1]

Hi BustaBook I am A software Devolper I am interested to know which data are you using for the simulation?

is it games won in the past? games won against opposing team in the past? Does each player represent a diffent variable?

By The way Nice Software you got there

[bustabook]

i use a few different stats, weighted differently. that being said, some of the stats i am using in beta 1.1 are offensive/defensive efficiency, shooting percentage (twos and threes), free throws/game, fg attempts/game.. i also add a few other things not mentioned here... some stats are randomized and some are not.

hope this helps,
GL

[acarmelo1]

i use a few different stats, weighted differently. that being said, some of the stats i am using in beta 1.1 are offensive/defensive efficiency, shooting percentage (twos and threes), free throws/game, fg attempts/game.. i also add a few other things not mentioned here... some stats are randomized and some are not.

hope this helps,
GL

By the way which language are you using? Visual Basic, C++, Java?

[suicidekings]

Golden State @ Okla City (-7) 222
Average Score: Golden State 99 - Okla City 115 (sim line: OKC - 16, 214)
Outright wins: Golden State 152 - Okla City 837 (OKC 84.4)
ATS wins: Golden State 298 - Okla City 678 (OKC 69.5)
O/U: Over 283 - Under 698 (Under 71.2)

[acarmelo1]

so how did it went??

I am gonna try and make one also looks fun

any insight you can provide me?

gonna do it using visual basic and SQL for the database

[bustabook]

i am using visual basic with an excel spreadsheet as my database.. i have done something similar to this 2 years ago in college football and was pretty successful with it.

my advice would be to go into this with achievable goals in mind.. and have more than one goal. for example, my first goal with this simulation program was to get the simulated scores as close as possible to the vegas lines. after i thought i conquered that, i wanted to find out what home-court advantage was worth and implement that... etc, etc.

good luck to you, i'm sure you will enjoy doing this as i have so far...

[bustabook]

Summary 12/7:

Side plays (all 1 unit):
Denver (-5.5)
Portland (-4)
Oklahoma City (-7)

Sides:
Denver/Philadelphia Under 216.5
Portland/New York Under 203
Golden State/Oklahoma City Under 222
San Antonio/Utah (Over 192.5)

Wow great night tonight.. the only one we missed was Portland, which i wondered aloud about earlier in the day.

-------------------
6-1, +4.9 units


Overall:
21-14, +3.9 units
--------------------

I know from the beginning i posted plays that were only over 60% in simulations but if i was playing EVERY play over 50% for 1 unit each (which i am now), i would be 39-22 (+14.8 units)... right around 65% w/volume..

Pretty awesome.

[MV]

it mustve been the Oden factor. i lost on the blazers too. but some of your picks made up for it

[bustabook]

i have no idea why that portland was so bad... but i assume you are right about oden being out. that is definately a weakness of this system. it doesn't account for injuries, current form, or other motivational factors.

i want to implement those things.. but its hard to put a numerical value on a player being out.. or having won 5 in a row.. or playing on back to back nights, etc.

[aznjeff07]

looking forward to tomorrows picks.

[bustabook]

here are predictions for 12/8, using data from 12/7... the results shouldn't differ all that much, but i will run another 1000 simulations tonight while i am sleeping with the most current data to be 100% accurate.

Early 12/8 (1000 simulations):


Denver @ Charlotte (3.5) 204.5
Average Score: Denver 96 - Charlotte 97 (sim line: Cha -1, 193)
Outright wins: Denver 439 - Charlotte 525 (Charlotte 54.5%)
ATS wins: Denver 350 - Charlotte 650 (Charlotte 65.0%)
O/U: Over 203 - Under 797 (Under 79.7%)

Minnesota @ Toronto (-8) 212
Average Score: Minnesota 102 - Toronto 112 (sim line: Tor -10, 214)
Outright wins: Minnesota 253 - Toronto 720 (Toronto 74.0%)
ATS wins: Minnesota 448 - Toronto 531 (Toronto 54.2%)
O/U: Over 534 - Under 438 (Over 54.9%)

Milwaukee @ Boston (-12.5) 191
Average Score: Milwaukee 87 - Boston 101 (sim line: Bos -14, 188)
Outright wins: Milwaukee 177 - Boston 810 (Boston 82.1%)
ATS wins: Milwaukee 453 - Boston 547 (Boston 54.7%)
O/U: Over 404 - Under 565 (Under 58.3%)

Cleveland @ Memphis (7.5) 192.5
Average Score: Cleveland 108 - Memphis 94 (sim line: Cle -14, 202)
Outright wins: Cleveland 807 - Memphis 173 (Cleveland 82.3%)
ATS wins: Cleveland 657 - Memphis 343 (Cleveland 65.7%)
O/U: Over 744 - Under 256 (Over 74.4%)

New Jersey @ Chicago (-7.5) 192.5
Average Score: New Jersey 90 - Chicago 92 (sim line: Chi -2, 182)
Outright wins: New Jersey 434 - Chicago 546 (Chicago 55.7%)
ATS wins: New Jersey 654 - Chicago 346 (New Jersey 65.4%)
O/U: Over 251 - Under 749 (Under 74.9%)

Sacramento @ New Orleans (-7) 211.5
Average Score: Sacramento 108 - New Orleans 106 (sim line: Sac -2, 214)
Outright wins: Sacramento 521 - New Orleans 453 (Sacramento 53.5%)
ATS wins: Sacramento 680 - New Orleans 294 (Sacramento 69.8%)
O/U: Over 571 - Under 429 (Over 57.1%)

Phoenix @ Dallas (-4) 214.5
Average Score: Phoenix 102 - Dallas 111 (sim line: Dal -9, 213)
Outright wins: Phoenix 254 - Dallas 717 (Dallas 73.8%)
ATS wins: Phoenix 350 - Dallas 629 (Dallas 64.2%)
O/U: Over 445 - Under 555 (Under 55.5%)

--------------------

Preliminary Sides:
Charlotte (+3.5)
Toronto (-8)
Boston (-12.5)
Cleveland (-7.5)
New Jersey (+7.5)
Sacramento (+7)
Dallas (-4)

Preliminary Totals:
Denver/Charlotte (Under 204.5)
Minnesota/Toronto (Over 212)
Milwaukee/Boston (Under 191)
Cleveland/Memphis (Over 192)
New Jersey/Chicago (Under 192.5)
Sacramento/New Orleans (Over 211.5)
Phoenix/Dallas (Under 214.5)

------------------

For those who can put in early action, the vegas lines will move toward these simulation lines almost inevitably!

GL

[1capper2008]

Hey Busta

I appreciate your analysis of these games, very helpful. I do have one question for you. For example, for tomorrow's Denver/Charlotte game, your system has the game total being under for 79.7% of the sims. Does this mean under the 204.5 (opening line) or under the 193 (sim line)? Keep up the good work!

Thanks!
1capper2008

[bustabook]

1capper, the 79.7% pertains to the vegas line... in other words, in my 1000 simulations, the game stayed under 204.5 79.7% of the time.

If i were to run 1000 simulations at my sim line (193), the over/under percentage would be right around 50%.

The "sim line" is merely the line that SHOULD BE on the game, according to my simulations.

Hope this helps, GL

[1capper2008]

Gotcha, thank you. That's what I thought, just wanted to confirm. Does your sim program use linear regression at all to come up with game totals, or is it not useful in your program?

[aznjeff07]

you siad your program dosen't account for multiple things, so i'm assuming it dosen't factor teams being tired, traveling, etc? Either way i'm goign to start playing your system it looks good and usually agrees with my pics-esp clev and the o/u on that game

[suicidekings]

Sacramento @ New Orleans (-7) 211.5
Average Score: Sacramento 108 - New Orleans 106 (sim line: Sac -2, 214)
Outright wins: Sacramento 521 - New Orleans 453 (Sacramento 53.5%)
ATS wins: Sacramento 680 - New Orleans 294 (Sacramento 69.8%)
O/U: Over 571 - Under 429 (Over 57.1%)

How do you go about balancing home & away stats in your simulations? The above line matches up with the teams' overall stats, but both of these teams are significantly better at home than on the road so far this season.

New Orleans: Home: 7-2 (96.3 pts for / 95.0 pts allowed) Road: 1-9 (98.2 pts for / 108.4 pts allowed)
Sacramento: Home: 8-3 (106.9 pts for / 102.5 pts allowed) Road: 1-7, (99.9 pts for / 107.4 pts allowed)

If I use overall stats, I get a line of SAC -2 (Total 205) as a fair line for a neutral court, but in New Orleans, I get a fair line or more like NOH -6.5 and a total of 199.5. Not to pick on your work, but I was just wondering how you came to that SAC -2?

[bustabook]

My understanding of linear regressions is that given x, you can determine y.. is this correct?

[bustabook]

aznjeff... no, beta 1.1 (this model) does not account for being tired, or playing on back to back days, or injuries...

i have started thinking about it and i think the main problem with it is that it is hard to put a numerical value on these things.. but i am sure with a little effort i could figure it out. i am just focusing on this thing until my ADD kicks in and i have to try something new, LOL.

as of right now i would suggest waiting a bit to play the system word for word.. i think it should be used in combination with other handicapping tools. there is some potential with the system, but i am sure one day i will go 1-9 and be looking for answers..

[1capper2008]

My understanding of linear regressions is that given x, you can determine y.. is this correct?

Yes, that is correct. But I'm not sure if the correlation is strong enough for NBA predictions.

[bustabook]

suicide-

i don't use home/away specific statistics... but i do account for home/away (if that makes any sense). i am of the opinion that a team will not beat the law of averages forever. i have my own way of calculating home-court advantage, and it may be completely wrong LOL.

as for your question on obtaining my sim line, i simply use average scores of the simulations (its almost arbitrary). i will tell you this, i dont know what Sacramento scores per game and i dont know what New Orleans scores per game on average. I know a bunch of other stats which tell me how much they score and how they score such as possessions/game, ft attempts/game, fg attempts/game, 3-attempts/game, along with respective percentages.

does this answer your question? i think this is what you wanted to know

[bustabook]

1capper,

i think you could get into some trouble using that to predict nba totals.. seems a little rigid, if that is the right word.

[bustabook]

also suicide-

i randomize most of the statistics, which makes the variance go way up... at first i only ran 100 simulations (which now seems completely ridiculous), and now i run atleast 1000 (which in a week may sound ridiculous also LOL)

[djiddish98]

Tonight was a good night - I'm guessing before actually seeing the data, but I think most teams played within their average ability with little to no outliers, which always make the predictions green. Keep up the good work.

[bustabook]

Tonight was a good night - I'm guessing before actually seeing the data, but I think most teams played within their average ability with little to no outliers, which always make the predictions green. Keep up the good work.

i havent compared last night's games to the averages, but i assume you are correct.

dji... i think you have hit the nail on the head. the system will work very well when teams play as they are supposed to. BUT isnt that the case with all handicapping? how many times have we said, "if (team A) played like they always do, i wouldv'e won."

Can you put yourself on a side better than "normal"? hopefully this system does exactly that.. put me on the "normal" side (and total).

thank you very much for the insight.

[Time2Score]

Bustabook good job with your system so far, its looking very strong. Have you been making tweaks to the OUs? Seems like they have been becoming more accurate. Cheers

[bustabook]

time2score-

i have only adjusted the system one time since i began posting... and that was because totals were coming in too low (because of an array not being populated correctly). that same day, i added a slight calculation in order to reward comparatively better offensive/defensive teams. this was all done after the first day of posting.

btw, it is my hope that everything gets more accurate with the more data i gather... thanks for the compliments.

[cortez11]

How many of these simulations do you actually place a wager on?

[bustabook]

i dont bet on simulations, they are too unpredictable.

[suicidekings]

suicide-

i don't use home/away specific statistics... but i do account for home/away (if that makes any sense). i am of the opinion that a team will not beat the law of averages forever. i have my own way of calculating home-court advantage, and it may be completely wrong LOL.

as for your question on obtaining my sim line, i simply use average scores of the simulations (its almost arbitrary). i will tell you this, i dont know what Sacramento scores per game and i dont know what New Orleans scores per game on average. I know a bunch of other stats which tell me how much they score and how they score such as possessions/game, ft attempts/game, fg attempts/game, 3-attempts/game, along with respective percentages.

does this answer your question? i think this is what you wanted to know

Yeah, that makes sense. I've never tried the route of simulations, but I can see that the stats you're describing are much more suitable for them. With respect to the home/away stats, there's only about 5-6 teams with a significant shift between their home and road performance, and as you say, over time it will even out.

Once again, I'm really enjoying following along with your progress. Makes me want to take a shot at it as well. Maybe a project for over the holidays.

[peterpan19]

interesting stuff thats going on in here... some things I am interested in are stuff like home court advantage...

as far as injuries to the starting 5 are concerned... every player has a numerical value (pr at least a bit)... like pts rebounds assist steals and so on per game... when a player is out (starting 5) the production should go down, but another player will replace some of the numbers... so maybe just lower the stats by a certain % according to how important a player was... e.g. a team scores 100 pts per game... player a is out and he has 25 pts on average... so the production should go down... as they normally dont have a 6th guy on the bench that can score 25pts per game too (or he would be starting too)... so the guy that comes in might score 15 pts... meaing the total average should drop to 90 pts... something like that... thats how you can add injured players...

GL

and really interesting stuff

[p0711142]

hey bustabook. great job there. is the simulation for 12/8 u have posted the lastest? or u have a final version of it later?

[Glitch]

i hope that bobcats shit is wrong.

[MV]

If I use overall stats, I get a line of SAC -2 (Total 205) as a fair line for a neutral court, but in New Orleans, I get a fair line or more like NOH -6.5 and a total of 199.5. Not to pick on your work, but I was just wondering how you came to that SAC -2?

i think it's just because the average scores of the simulations give SAC 108 NOH 106, that's where he gets the SAC -2 line.