Organizing Statistics and Building Databases

[CTA Gitz]

Hello all, first time post. I am in the preliminary stages of building a model for the nfl and I am having trouble getting the stats I need. I would like to build a program that goes through play by play data and separates it in excel into various categories. The play by play is on the website or each game is downloadable in pdf form. Do you think this is doable? I dont know any code how can I go about this? Any advice would be appreciated.

Thank you all very much

[Justin7]

It is certainly doable, but you may have to hire a programmer.

Getting play-by-play data opens up a lot of opportunities. If you can get it, it has lots of uses... from improved modeling to attacking props.

[GoIrish682]

Hey Justim,
i've been enjoyin gyour posts, and have discovered you on youtube...you are very smart, i have an important question..reading your posts, i get the strong feeling you are a big 'prop' guy and not so much game sides or totals...what can you reccommend to really find out what props to got after?
thanks

[Maverick22]

Get the relational database, mysql. Learn SQL.

You will thank me in the end

[Justin7]

Hey Justim,
i've been enjoyin gyour posts, and have discovered you on youtube...you are very smart, i have an important question..reading your posts, i get the strong feeling you are a big 'prop' guy and not so much game sides or totals...what can you reccommend to really find out what props to got after?
thanks

I actually bet more sides and totals than props.

Look at every single prop you can find anywhere. Price it. Go after them all at first. You'll find certain ones frequently offer value.

[CTA Gitz]

If I am unable to learn how to write this code in a timely fashion is there anywhere I can find this kind of data on the internet, even if I pay for it?

[Indecent]

Do it the hard way. You will learn a lot writing the scraper, and what you learn will help you in building the model.

[Maverick22]

Do it the hard way. You will learn a lot writing the scraper, and what you learn will help you in building the model.

I would have to disagree.

Unless the model is pretty basic. Or the scraper is pretty exotic. Not sure how what you learn from the scraper helps you with the model.

[Indecent]

I would have to disagree.

Unless the model is pretty basic. Or the scraper is pretty exotic. Not sure how what you learn from the scraper helps you with the model.

Well, I'm assuming he knows very little to start with. If you don't already know how to program, programming anything is better than not, and anything you learn will make the next task easier.

Maybe it's interpreting a compiler error message you've never seen before, solving a dependency problem, working with an external class/API for the first time, etc. There are lots of little things can come up, the more you program the more you encounter them and the easier the solution/fix can be in the future.

[Maverick22]

Well, I'm assuming he knows very little to start with. If you don't already know how to program, programming anything is better than not, and anything you learn will make the next task easier.

Maybe it's interpreting a compiler error message you've never seen before, solving a dependency problem, working with an external class/API for the first time, etc. There are lots of little things can come up, the more you program the more you encounter them and the easier the solution/fix can be in the future.

touche

[Flight]

Get the relational database, mysql. Learn SQL.

You will thank me in the end

+1

Don't use Excel for data storage. Use SQL. You can then build custom queries at any time to give yourself powerful views of your data.

[subs]

hi could u please explain that a little. i am using excel and find it easy and intuitive. what am i missing out on. my time is def finite so....

[boondoggle]

What the guys are suggesting is that you utilize a relational database for your data warehousing instead of a spreadsheet. This has many advantanges over excel.

1. create normalized database

2. now you can actually obtain your data via queries that can be ran on the fly to dig the results you want out of your model.

An example of SQL query for SQ SERVER 2005/8, "select * from table patterns where pattern = 'won_5_of_last_7'"

You can download SQL SERVER EXPRESS 2008 for free from microsoft. It scales and performs much better than excel and gives incredible flexibility for viewing data based on how you want to view it.

The only downside is that if you can do all of the above than you could probably make more money programming SQL than you will gambling

[subs]

thanks

i'm gonna look at a tutorial and if my head doesn't hurt too much after a few hours maybe get stuck in proppa.

BTW can't i do this via lookup tables? i mean ok it takes a bit longer to sort it all out but then so does learning something completely new?

[Borat38]

What the guys are suggesting is that you utilize a relational database for your data warehousing instead of a spreadsheet. This has many advantanges over excel.

1. create normalized database

2. now you can actually obtain your data via queries that can be ran on the fly to dig the results you want out of your model.

An example of SQL query for SQ SERVER 2005/8, "select * from table patterns where pattern = 'won_5_of_last_7'"

You can download SQL SERVER EXPRESS 2008 for free from microsoft. It scales and performs much better than excel and gives incredible flexibility for viewing data based on how you want to view it.

The only downside is that if you can do all of the above than you could probably make more money programming SQL than you will gambling

Excellent advice!

[CTA Gitz]

Alright thanks for all of the help everybody. I have since found ten years of data to use. 2000-2009 on armchairanalysis.com. I would like to learn all that is mentioned but in the meanwhile I am teaching myself vba so I can splice my excel data however Id like. I would like a macro that sorts each team's games into their own sheet. I did this, which did not work. The teams are in Column F for now.
Sub Team_Sort()
Dim This_Team, Next_Team As String
Dim i, j As Long
Dim Sheet_Num As Integer
i = 2
This_Team = Cells([i], 6).Value
Next_Team = Cells([i+1], 6).Value
Sheet_Num = 2
j = 2
For i = 2 To 2565 'This is the total # of rows'
If This_Team = Next_Team Then
i = 1 + i
End If
If This_Team <> Next_Team Then
Rows("j:i").Select
Selection.Copy
Sheets("Sheet_Num").Select
ActiveSheet.Paste
j = i
Sheets(Sheet1).Select
Sheet_Num=Sheet_Num+1
End If

Next

End Sub
Any suggestions on making this work?

[CFA]

If I am unable to learn how to write this code in a timely fashion is there anywhere I can find this kind of data on the internet, even if I pay for it?

Tell me how to send you an excel spreadsheet, and I will send it to you.