Gentlemen, I have been reading several of the threads concerning scraping stats from various statistical websites. I have no programming experience and was thus wondering, are there any (reasonably priced) statistical services whereby you could obtain up-to-date stats (example: box scores for college football, NFL or baseball) that could be downloaded into your database and/or Excel spreadsheet?
-
-
My question exactly...Why do something yourself that would take an overwhelming number of hours when you can just pay someone to do it for you? lol
I've subscribed to Accuscore before ($99 a month) a service that gives you winning percentages for spreads, totals, moneyline..etc.
I've subscribed to Sportsinightsbefore ($200 a month) a service that gives you so-called smart money plays and steam moves in addition to live odds and public betting percentages from the major offshore sportsbooks
I've subscribed to Donbest before ($499 a month for real-time Vegas odds, or $99 for 1-2 minute delayed odds)
And despite all this, I'm still losing money lol
But I think I'd definitely be willing to pay someone or a service if they can scrape all the data from the websites for Korean men's, women's basketball and Korean baseball..organize and create a nifty program for me that figures out winning percentages against different spreads and totals and straight up
Trust me, I've looked for this type of service, and there isn't one ...at least not on google...
actually, you might wanna google "webscraping services"....there seems to be quite a few...might try one of these myself..i wonder if u can hire computer programmers to design these types of programs for u tooLast edited by kpoutlaw; 03-02-11, 11:35 AM.Comment -
Ok your query prompted me to do some more research myself and I found this website www.existonline.com..I just off the phone with them and they told me they can do everything you mentioned above in addition to organizing the data you want depending on the parameters you give them....so I said "how much??" he said i would have to send him an email of exactly what i want and depending on the amount of information you want...so i was persistent and said..well give me a ballpark figure and i said, for example, will it cost me several hundred, possible more than $500? and he said definitely not more than $500, and so i said, how does the payment work, ..do i pay upfront, how do I know your service is trustworthy? and he said, I pay 50% upfront and once theysend me what I want and I am 100% satistified, then I can pay the rest, or get a 100% refund..I might actually try this, just hope they don't screw me...and do exactly what i want, or maybe ill look around google a little moreComment -
kpoutlaw, thank you for your replies and suggestions. I looked at the www.existonline.com website, and while this certainly may be a possible solution, I was thinking more of an actual service that provides stats, box scores, etc. For example, where do all the sports sites (ESPN, Fox Sports, CNNSI, etc) get their stat information from (where does their baseball and football box scores come from)? Being unfamiliar with scraping, I was just curious as to how other's gather data for their databases and spreadsheets and thought there might be an alternative solution.Comment -
This question comes up every couple of months (probably more often for each specific sport). In general the thread goes a few replies and then dies with no result found, except for a few very specific, limited examples (if you want men's college football check out Soren Sorensen's site). I have myself requested this a few times over the years.
This leads me to believe the service does not exist in any economical way. The larger services of course do not provide it, as a cheap data feed would very easily allow anyone to set up a competing website. Similarly with the gaming sites. My guess is your best bet is to either code or hire a coder to scrape data from webpages. In general, I think its a terrible way to obtain information. It is prone to errors, keeps you in a constant battle of maintaining the code to conform to any new layout changes from the site, and is somewhat slow and bandwidth intensive. That seems to be the state of the art currently however.
I am interested in getting a dataset for this year's NCAAB touney. If you decide to go this route, PM me and maybe we can pool resources to get what we need. Without a quote, I cannot compare rates but I have also had some success hiring out jobs to http://rentacoder.com
Good luck,
SpektreComment -
Ok, I have a question....Let's say that I want to hire out a computer programmer or coder or web scraper or whatever you want to call it, to scrape data from a Korean baseball site to determine the win probability of each team based on 10,000 simluations similar to Accuscore. How far back do I go? 3 years? 2 years? And what parameters do I have to set for the coder..In other words, similar to Accuscore, I want to know
1. The win probability of a team against another team (moneyline)
2. The win probability of a runline (-1.5, -2.5)
3. The win probaility of a total (7, 8, 9)
Do I just tell him this stuff and he'll know what to do, or do I have to tell him specifically how to get this information? (Like do I have to tell him, you have to add up all the win totals for the team and blah blah)...I know this sounds like a stupid question, but I'm asking because I wouldn't know how to extract this information or create a program myself..I just know what I want, and I want something similar to Accuscore (even though I have no idea how they figure out their winning probabilities)Comment -
Don't take offense to my comments, or do
Your question is just too general.
At best I can say you are looking at what is generally considered three seperate activities.
1. Database building (which is what it seemed you were asking for in the original post and I commented on above). For this, in general you would specify exactly WHAT data you want in your database and WHERE the programmer should obtain it from, although the WHERE could possibly be left to an experienced programmer you trust. The WHAT is decided from an initial estimate of important data or simply decided to be all the data you can get your hands on
2. Modelling - Deciding what weighting and logic is given to each piece of information you deem important. This is determined by what output is desired from a simple "I want a power rating which shows a ranking of all the teams in order", to a more complex, "For any two teams on a given field what is the advantage of one team over another", to a very complex, "Using statistcal models, try to decide the outcome of each and very play in the game and then simulate the outcome of thousands of games". Generally a process called backtesting is used to validate the model and see how well it would have worked had you run it over a series of past data. There is no easy outsourcing of this. If you do not know (and do not CARE) what information goes into this decision, why not just use Accu...
3. Generating the desired output. A fairly simple step after the model has been developed.
SpektreComment -
i'm looking to outsource to india now. had this nice pakistani guy doing some work for me but when he learned it was for gambling he dumped me 2 weeks into the work!! grrrrrrr. then what's worse he tried to convert me!
i told him strictly speaking HE'S NOT gambling he's just writing code. he told me be ready to burn in hell. wouldn't even give me the code he'd already written.
top tip 1 don't go the muslim route elohel. i thought cool he won't be using my app cos he muslim but guess that backfired....
definetely nothing against muslims in general cos i am sort of half muslim, but just thought i'd share anyway.
Comment -
Subs, That sucks. That is why you never give a single programmer all of the program to write. Give a single person enough to do what they need, but not enough to figure out what exactly they are writing (if you can)
On the other hand... Maybe i really should go ahead and make this website after all... The one where people subscribe to me and get data feeds of the previous day's sporting events.Comment -
yea i really blew it there when he asked some tough questions.
i think the idea is a good 1 and pretty sure i'm not alone.
my initial thoughts r that i'm not sure there'd be massive profits for u because once u let the data out what's to stop some1 else selling it on for peanuts. not me but u know some gamblers always need money
the other thing that is probably going to fit is the SBR model where they sell info to the enemy. kinda wrong and may put some off, not me i'm a "pro" already. puke.
if u can do it fairly economically the profit should just tick over slowly.
wishing u teh best of luck with it. let us know what u decide.Comment -
you are likely right. i dont web program anyways. i'd have to make a website and super secure it. and figure out how to collect money...
too much effort, and the potential money has not made me care enough.
you pay me enough, i will write your program for you. long as it is not in python
Comment -
You really want to build the database and model yourself - you learn a ton while doing it.Comment -
No offense taken at all. If anything, you have been very informative and have validated my concerns that this isn't just a simple mathematical equation that you can ask your teacher to give you the answer to. I was really hoping this could be something you could just tell an experienced programmer with a knowledge in sports betting to do by just telling him "Look, I want a program that will tell me the win probability of a team ATS and straight-up. I don't know how or what to base these calculations on, but based on your extensive knowledge and since I'm giving you lots of money, just make sure it works." lolDon't take offense to my comments, or do Your question is just too general. At best I can say you are looking at what is generally considered three seperate activities. 1. Database building (which is what it seemed you were asking for in the original post and I commented on above). For this, in general you would specify exactly WHAT data you want in your database and WHERE the programmer should obtain it from, although the WHERE could possibly be left to an experienced programmer you trust. The WHAT is decided from an initial estimate of important data or simply decided to be all the data you can get your hands on 2. Modelling - Deciding what weighting and logic is given to each piece of information you deem important. This is determined by what output is desired from a simple "I want a power rating which shows a ranking of all the teams in order", to a more complex, "For any two teams on a given field what is the advantage of one team over another", to a very complex, "Using statistcal models, try to decide the outcome of each and very play in the game and then simulate the outcome of thousands of games". Generally a process called backtesting is used to validate the model and see how well it would have worked had you run it over a series of past data. There is no easy outsourcing of this. If you do not know (and do not CARE) what information goes into this decision, why not just use Accu... 3. Generating the desired output. A fairly simple step after the model has been developed. Spektre
Anyway, thanks for the reply. I'll have to do more extensive research myself, and alas, figure out how to learn/use these programs myself.
And in response to Subs, I sent you a PM, and sorry, but I couldn't help but crack up when you mentioned the Muslim story. LOLComment -
pushkin - if you are as smart as Pushkin the poet was - yo gotta understand - STATS DON"T MEAN SQUAD in wagering and capping. I have won quite a few SBR contests and I DON"T EVER USE STATS. You gotta analyze GAME FACTORS. Quit chasing that bait that bookies try to lure you on.Comment -
Arenas just over a month ago was on court all the time. But it was even on the news that his Knee hurt him in Humid weather. I wagered Against Orlando spread each time the weather was humid for Orlando during that period. That's an example.
A NBA team has just lost a top scorer and plays against a weak one with a 10 point spread. Even if the spread moved a bit, the team has to be disorganized, unless it's MIami with its Magnificent 3 Where 1 can be replaced by two. So the first game would be hard to beat a 10 or 8pnt spread.
Such games are rare maybe few-five a week, but those are surely your money-makers. Sports like NCAAB and College Football and MLB are the best in this regard.Comment
