What's new
Fantasy Football - Footballguys Forums

This is a sample guest message. Register a free account today to become a member! Once signed in, you'll be able to participate on this site by adding your own topics and posts, as well as connect with other members through your own private inbox!

Data Mining Predictive Forums (1 Viewer)

Bolted

Footballguy
By trade I write data analysis software. When cruising different forums for information to help in my upcoming draft here is what I do…

1) Go to LAST YEARS posts.

2) Read them looking for posts that were well thought out AND ultimately happened similar to what the poster predicted.

3) Read a few more predictions from that poster.

4) If a high percentage of those predictions panned out, then go to that posters current posts and see what he thinks this year.

Before you get too excitted, it takes ALOT of work, and it doesn’t eliminate you from thinking things out for yourself. What it does do is help to eliminate a lot of the “noise”. It lets you focus on information provided by guys with a more proven track record. I honestly believe that this trick alone has helped me be very successful in my leagues over the last 3 years that I have been doing it.

Here is my plea to Footballguys.com!!!!

If you guys could set up a feature where people could go back and “rate” posts. You could give each post a thumbs up or down or neutral. People would then get rated on both the quality AND outcome of their posts. It is important to only allow this AFTER the season, because a rating of a prediction is useless until you can see what actually occurred. You then allow people to be ranked by those values.

Here is what this would help to drive on your site.

1) Loyalty – It simply gives additional incentive for a good poster to keep coming back instead of jumping to some new site.

2) Access to Better Information – Selfishly it would help me and others that do the same thing avoid the noise.

3) Drive Forum Participation During the Off-season - Let’s face it, it would give you something to do that would be fun and interesting after the season is over, or early in the season long before your draft.

4) Website Competitive Advantage - If you are the first to do it, then you will will have one more competitive advantage against your competition. Seriously, your historical forum data is some of the most valuable product you own. If you could help people “mine” that data for information, you win.

5) Better Posts – If people knew that they were a little more accountable for the content of their posts down the road, then you would probably drive more thoughtful posts with more meat then you get today. Some posters are awesome at this already, I just think they should be “rewarded” when that research pays off, and the predication turns into FFL value.

When you do implement this feature, all I ask is you float me a free subscription for life. Honestly, keep up the great work, you are the best forum out there.

:kicksrock:

 
It's a good idea in theory, but in practice I have my doubts. I've seen other forums that use rating systems and more often than not, the system is poorly used and simply becomes a childish war as people intentionally rate down good posts simply because of a dislike for the other users. This ultimately makes the ratings useless. If it works, great, but I wouldn't pin my hopes on it.

 
It's a good idea in theory, but in practice I have my doubts. I've seen other forums that use rating systems and more often than not, the system is poorly used and simply becomes a childish war as people intentionally rate down good posts simply because of a dislike for the other users. This ultimately makes the ratings useless. If it works, great, but I wouldn't pin my hopes on it.
See, that post would have gotten a "Thumbs Down" :confused: at the end of the year :yes: Just Kidding.That is why I emphasized "AFTER the season". It takes the emotional component out of it, and also at that point it is no longer theory. The only place I have seen it done is where the thumbs up and down happen right now. What good is that if it is all predicitions. If a guy posts that Marion Barber is going to be the #1 RB this year because of A,B,C, and D. People will kill him with "thumbs down". If he makes that post, and after the season is done he is sitting at #1 or #2, my guess is he gets a ton of Thumbs up.
 
Last edited by a moderator:
I like the idea, but since it is not going to get implemented this year, can you share what your system is telling you for this season?

 
It's a good idea in theory, but in practice I have my doubts. I've seen other forums that use rating systems and more often than not, the system is poorly used and simply becomes a childish war as people intentionally rate down good posts simply because of a dislike for the other users. This ultimately makes the ratings useless. If it works, great, but I wouldn't pin my hopes on it.
I think he is saying their rating would be exclusive to you, and the ratings would be kept under your username when you logon, so you can find and remember the posters you like. Good idea IMO...
 
Last edited by a moderator:
darn. reading the first few words of this post's title, i thought someone was going to suggest an algorithm for predicting how many fantasy points a given player was going to score in the upcoming week, given tons of data points about past performances. oh, well.

 
Yeah, because if people guessed right last year, they'll guess right this year. :)
A few years ago a blowhard FBG had a great season. All his success went into his signature for all to see and he began blowing harder than ever. For over a year many forum readers blew smoke up his dress encouraging him to continue his condescending annoying blowhard shtick. The problem is for the last few years his opinion blows. It's the primary flaw in the idea. Past performance is no guarantee of future results.
 
That is why I emphasized "AFTER the season". It takes the emotional component out of it, and also at that point it is no longer theory. The only place I have seen it done is where the thumbs up and down happen right now. What good is that if it is all predicitions. If a guy posts that Marion Barber is going to be the #1 RB this year because of A,B,C, and D. People will kill him with "thumbs down". If he makes that post, and after the season is done he is sitting at #1 or #2, my guess is he gets a ton of Thumbs up.
Fair enough. I still have my doubts about any system that relies on the masses to use it properly. The community here may be above the petty stuff and it could work well.
I think he is saying their rating would be exclusive to you, and the ratings would be kept under your username when you logon, so you can find and remember the posters you like. Good idea IMO...
I disagree. I think the system he is proposing was not for each individual, but rather a community rating. That being said, exclusive ratings would eliminate the abuse I expressed my concern with, but it would not address the other issues, such as:
1) Loyalty – It simply gives additional incentive for a good poster to keep coming back instead of jumping to some new site.2) Access to Better Information – Selfishly it would help me and others that do the same thing avoid the noise.3) Drive Forum Participation During the Off-season - Let’s face it, it would give you something to do that would be fun and interesting after the season is over, or early in the season long before your draft.4) Website Competitive Advantage - If you are the first to do it, then you will will have one more competitive advantage against your competition. Seriously, your historical forum data is some of the most valuable product you own. If you could help people “mine” that data for information, you win.5) Better Posts – If people knew that they were a little more accountable for the content of their posts down the road, then you would probably drive more thoughtful posts with more meat then you get today. Some posters are awesome at this already, I just think they should be “rewarded” when that research pays off, and the predication turns into FFL value.
 
Yeah, because if people guessed right last year, they'll guess right this year. :)
A few years ago a blowhard FBG had a great season. All his success went into his signature for all to see and he began blowing harder than ever. For over a year many forum readers blew smoke up his dress encouraging him to continue his condescending annoying blowhard shtick. The problem is for the last few years his opinion blows. It's the primary flaw in the idea. Past performance is no guarantee of future results.
Perhaps, but over time it would be self regulating. That blowhard would have had his posts in the second and third years downrated. Trends could emerge after a few seasons which tend to show those who tend to right more often....
 
Yeah, because if people guessed right last year, they'll guess right this year. :goodposting:
A few years ago a blowhard FBG had a great season. All his success went into his signature for all to see and he began blowing harder than ever. For over a year many forum readers blew smoke up his dress encouraging him to continue his condescending annoying blowhard shtick. The problem is for the last few years his opinion blows. It's the primary flaw in the idea. Past performance is no guarantee of future results.
Perhaps, but over time it would be self regulating. That blowhard would have had his posts in the second and third years downrated. Trends could emerge after a few seasons which tend to show those who tend to right more often....
Exactly. It is true that it won’t predict with any degree of certainty. It is still just a prediction. For the nay sayers, let me ask you a question…Do you treat your investment portfolio the same way? Sure there is no way to look at past performance and guarantee future returns, heck they even tell you that when you make an investment. But I would hate to buy any stock if they didn’t expose to me the past performance of a stock. Essentially that is what we are doing now, we consume all this info with no record of past performance. Sure, it isn’t perfect, but if someone in your league just always seems to win, and that same someone was posting a bunch of stuff on this forum, wouldn’t you want to know who that guy was? And if that guy that always seems to lose your league was also posting his thoughts each year, wouldn’t you want to know those posts so you could ignore them?
 
Yeah, because if people guessed right last year, they'll guess right this year. :thumbup:
A few years ago a blowhard FBG had a great season. All his success went into his signature for all to see and he began blowing harder than ever. For over a year many forum readers blew smoke up his dress encouraging him to continue his condescending annoying blowhard shtick. The problem is for the last few years his opinion blows. It's the primary flaw in the idea. Past performance is no guarantee of future results.
:goodposting:
 
how bout to negate the false negative rankings on even the great posts out of spite

how bout you just narrow it down to rating mostly the spotlight posts AND only a select few (staff) actually have ability to rank posts

 
Yeah, staff ratings on posts might be the only way to do this and bullet-proof it vs. the "idiot masses". I know we've had this discussion at least once each of the last 2 years, and similar systems have been proposed; I'd suggest you do a search for it, but the search functionality...umm, sucks :lmao:

Anyway, I'd love to see this work. Might even be able to "beta" this on a SharkPool 2.0 board??

 

Users who are viewing this thread

Back
Top