Draft Strategy (1 Viewer)

[HULLOBUDMAN]

Let me begin by saying that this is my first stab at this system. I was reading a thread this past off season that discussed the virtues of individual player projections vs. rankings/buckets. Someone in that thread posted something that I felt was interesting and I thought I would give it a try this year. I have never been a huge fan of projecting individual players. I don’t have the time or patience to do it. In the thread that I was reading an individual brought up that the way they do their rankings/projections were to take an average of the scoring for each position over a certain time period. This would give you a historical average that QB1 out preforms QB 12 by x amount of points per year. So rather than projecting Aaron Rogers to throw for 7000 yards and 67 TDs this year, I am using QB1 (Rogers) to score about 408 fantasy points this year. I took a five year average and plan on using a 5 year look back for years in the future. Hopefully this somewhat takes into account trends of the game.

If I could get some help from the Shark Pool that would be great. Here are some of the things I find myself questioning myself about:

1) Over the past 5 years we have had some ridiculous record breaking performances. Brady and Moss in 2007. Tomlinson in 2006. I backed these record breaking out of the average. Not sure if I should do this or if it really matters as it only affects QB1, RB1, and WR1. All most likely first round picks anyway.

2) Our league has a starting lineup of 1QB, 1RB, 2 WR, 1 TE, 1 FLEX (RB,TE,WR). What is your opinion of the baseline that should be used for the RB, WR, TE position? Should TE even be factored into this?

3) When I look at FBGs projected stats using my scoring system vs. the historical averages, the numbers are quite difference. For instance, the age old theory of drafting QB later in the draft is supported by FBGs projections. The difference of projected stats shows a mere 20 points difference between the #7 and # 12 QB. In the actual scoring average in for the past 5 years the difference is about 44 points. The difference between #7 and #24 in the projection is only 68 points but average is closer to 100. Am I misinterpreting the data? Does this type of discrepancy hold any water?

I am not hoping to turn this into a projections are worthless/rankings are still projections battle. Just hoping for some constructive criticism and feedback.

To you statistics guys, try not to torch me too bad…

 

[Binky The Doormat]

Do you have the draft records of your league over the past few years?

 

[HULLOBUDMAN]

Do you have the draft records of your league over the past few years?

Yes. DD files.

 

[Binky The Doormat]

I like your approach - I use a combination of Projections Dominator (25% each), staff top rankings and my own judgement to create my own rankings.

Then I use ADP modified with the typical flow of our draft. I use this sheet to quickly see who is coming and for my own mocks.

 

[cheatsheetwarroom]

With the way the league changes from year to year, I think looking 5 years into the past is way, way too big of a sample size to consider. The only thing, to me, that is relevant is last year, period.

When I create my cheat sheets I look at 2 main factors: 1) each players 'fantasy point per game' output the previous season and 2) where other reputable sources have that player ranked. Between those 2 things, and any recent news regarding the player, I can make a good judgment to each player's value relative to other players of the same position.

I created a cheat sheet creation tool which integrates all of the aforementioned data into a web application where you can create your player rankings using drag and drop online, for free. I use the 'note' area of each player template to jot-down any recent news about each player, I use the supplemental rankings to make sure my own rankings aren't too far off, I use the 'bust', 'sleeper', and 'injured' tags to remind me of where that player stands, and that is really all you need to create an accurate spreadsheet.

I say that your time developing complex algorithms to rank players would better be used researching the most recent news possible about each player (using Google News Search) and integrating those factors directly into your cheat sheet. Knowledge is power and what happened 2 years ago is irrelevant, IMO

 

[HULLOBUDMAN]

With the way the league changes from year to year, I think looking 5 years into the past is way, way too big of a sample size to consider. The only thing, to me, that is relevant is last year, period.

When I create my cheat sheets I look at 2 main factors: 1) each players 'fantasy point per game' output the previous season and 2) where other reputable sources have that player ranked. Between those 2 things, and any recent news regarding the player, I can make a good judgment to each player's value relative to other players of the same position.

I created a cheat sheet creation tool which integrates all of the aforementioned data into a web application where you can create your player rankings using drag and drop online, for free. I use the 'note' area of each player template to jot-down any recent news about each player, I use the supplemental rankings to make sure my own rankings aren't too far off, I use the 'bust', 'sleeper', and 'injured' tags to remind me of where that player stands, and that is really all you need to create an accurate spreadsheet.

I say that your time developing complex algorithms to rank players would better be used researching the most recent news possible about each player (using Google News Search) and integrating those factors directly into your cheat sheet. Knowledge is power and what happened 2 years ago is irrelevant, IMO

I agree that 5 years may be a bit too far to look back but 1 year is not near far enough imo. Teams make adjustments like NE going to the spread or Miami going to the Wildcat. The league starts to copy cat those offenses or parts of them but first it seems that they need to attain the personel first. Minny getting Harvin, KC drafting McCluster, Cleveland getting Cribbs more involved. You don't see the true impact of these trends in just 1 year.Not sure what complex algorithims you think I am creating but adding 5 numbers and then dividing it by 5 is fairly simplistic. You are obviously not a proponant on VBD,that's cool. I haven't drafted using a printed cheatsheet since the DD came around. It is definitely the best organizational draft tool even if you don't buy into VBD.

Thanks for te comments.

 

[cheatsheetwarroom]

With the way the league changes from year to year, I think looking 5 years into the past is way, way too big of a sample size to consider. The only thing, to me, that is relevant is last year, period.

When I create my cheat sheets I look at 2 main factors: 1) each players 'fantasy point per game' output the previous season and 2) where other reputable sources have that player ranked. Between those 2 things, and any recent news regarding the player, I can make a good judgment to each player's value relative to other players of the same position.

I created a cheat sheet creation tool which integrates all of the aforementioned data into a web application where you can create your player rankings using drag and drop online, for free. I use the 'note' area of each player template to jot-down any recent news about each player, I use the supplemental rankings to make sure my own rankings aren't too far off, I use the 'bust', 'sleeper', and 'injured' tags to remind me of where that player stands, and that is really all you need to create an accurate spreadsheet.

I say that your time developing complex algorithms to rank players would better be used researching the most recent news possible about each player (using Google News Search) and integrating those factors directly into your cheat sheet. Knowledge is power and what happened 2 years ago is irrelevant, IMO

I agree that 5 years may be a bit too far to look back but 1 year is not near far enough imo. Teams make adjustments like NE going to the spread or Miami going to the Wildcat. The league starts to copy cat those offenses or parts of them but first it seems that they need to attain the personel first. Minny getting Harvin, KC drafting McCluster, Cleveland getting Cribbs more involved. You don't see the true impact of these trends in just 1 year.Not sure what complex algorithims you think I am creating but adding 5 numbers and then dividing it by 5 is fairly simplistic. You are obviously not a proponant on VBD,that's cool. I haven't drafted using a printed cheatsheet since the DD came around. It is definitely the best organizational draft tool even if you don't buy into VBD.

Thanks for te comments.

I'm not saying going back 5 years is completely worthless, just that when you go back farther than 1 year I think your expectations start to get skewed and that time would be better spent focusing on what happened in camp yesterday.

What is VBD?

 

[HULLOBUDMAN]

With the way the league changes from year to year, I think looking 5 years into the past is way, way too big of a sample size to consider. The only thing, to me, that is relevant is last year, period.

When I create my cheat sheets I look at 2 main factors: 1) each players 'fantasy point per game' output the previous season and 2) where other reputable sources have that player ranked. Between those 2 things, and any recent news regarding the player, I can make a good judgment to each player's value relative to other players of the same position.

I created a cheat sheet creation tool which integrates all of the aforementioned data into a web application where you can create your player rankings using drag and drop online, for free. I use the 'note' area of each player template to jot-down any recent news about each player, I use the supplemental rankings to make sure my own rankings aren't too far off, I use the 'bust', 'sleeper', and 'injured' tags to remind me of where that player stands, and that is really all you need to create an accurate spreadsheet.

I say that your time developing complex algorithms to rank players would better be used researching the most recent news possible about each player (using Google News Search) and integrating those factors directly into your cheat sheet. Knowledge is power and what happened 2 years ago is irrelevant, IMO

I agree that 5 years may be a bit too far to look back but 1 year is not near far enough imo. Teams make adjustments like NE going to the spread or Miami going to the Wildcat. The league starts to copy cat those offenses or parts of them but first it seems that they need to attain the personel first. Minny getting Harvin, KC drafting McCluster, Cleveland getting Cribbs more involved. You don't see the true impact of these trends in just 1 year.Not sure what complex algorithims you think I am creating but adding 5 numbers and then dividing it by 5 is fairly simplistic. You are obviously not a proponant on VBD,that's cool. I haven't drafted using a printed cheatsheet since the DD came around. It is definitely the best organizational draft tool even if you don't buy into VBD.

Thanks for te comments.

I'm not saying going back 5 years is completely worthless, just that when you go back farther than 1 year I think your expectations start to get skewed and that time would be better spent focusing on what happened in camp yesterday.

What is VBD?

Value Based Drafting. Are you a member of the website?