The Gut Check No. 159 - High-Risk Series Part III (1 Viewer)

[Matt Waldman]

The Gut Check No. 159 - Crank Scores Part I (Part III of the High-Risk Series)

I'll have some projection methods for this approach with Part IV of the series. This portion is an explanation of the equation and how it can work out for you.

 

[stickboy]

Matt,

Good intro, an up to date summary of the Crank Scores rather than chasing through the articles at your previous site.

1. Will you be publishing a tool that allows us to enter our scoring system, league size, number of years to analyze, etc. to calculate Crank Scores?

2. Have you done any analysis of where consistency and high point projections occur together, and where they diverge? For example, if we are drafting 6th round QB's that are maybe QB7-12 in terms of point projections, are these more or less consistent than other parts of the curve where draft runs occur? What about the volatile WR position, especially in a PPR? Is there part of the point projection curve where I can get better consistency relative to the hot and cold WR's?

3. Are there any insights from the slope of the Crank Scores and where the fall off occurs most?

sb

 

[Matt Waldman]

Stickboy,

Good questions. I'll see if I can address some of those next week and/or the coming weeks.

 

[puckalicious]

I'm new to this idea (although I've heard of it) but the logic behind it all makes perfect sense. I'm a little bit skeptical of the predictive capability of this - not that you're advertising it as such - but does a player's past Crank score have much to do with future performance? Have you done a "rear-view" Crank analysis of predicted vs. actual player performance over the last several seasons?

I get the feeling that a particular player's Crank score could vary a lot from year to year based on their team situation, but maybe that's to be expected like everything else. Looking forward to Part IV.

 

[Dr. Octopus]

Great work as usual Matt. However, does the Crank Score and "risk" really go hand in hand? It actually seems like an approach that attempts to minimize risk.

I really enjoyed the first two parts of the series and look forward to incorporating the approach to some extent in my re-draft.

 

[hickbones]

I enjoyed the crank article. Nice to see someone quantifying what has historically been a gut check thing for me. I don't think I've ever drafted Ochocinco or Santana Moss because of inconsistency. It would be sweet if this could somehow be incorporated into the draft dominator. It could be a box to pull up and see crank scores when judging between two available and similar players. Or it could be something that is incorporated into the algorithm for draft advice.

I think Greg Allen did something like this a couple years ago with his Gold Algorithm that was used to market their draft program. I liked the idea a lot, but had my worst draft on record!

 

[Lord Fantasy]

I really like this piece, too, but I have to confess you kind of lost me when it came to actually calculating the crank score.

I'm going to try reading the piece again with fewer than six beers in me, but just in case that doesn't work is there a dumbed-down version for us who aren't so good with the math?

 

[stickboy]

I really like this piece, too, but I have to confess you kind of lost me when it came to actually calculating the crank score. I'm going to try reading the piece again with fewer than six beers in me, but just in case that doesn't work is there a dumbed-down version for us who aren't so good with the math?

Cliff Notes version:Week to week consistent scoring is better than one huge week followed by several bad weeks because we play head to head rather than total points.Standard deviation is the typical statistical measure for variations in scoring, but we want the upside, but not the downside, so another measure is needed.Crank measures how often a player exceeds a set of baseline values (points per week). The baselines represent performance tiers (elite = top 2 or 3 at a position, #1 tier = better than 12th player, #2 tier = better than 24th player, etc.) where the baselines are based on the size of your league -- will they likely outperform as a starter, a WR2, etc. There is also a subpar tier for the under performers.Since we all want a single number to use for comparison, it calculates (GOOD - BAD) relative to the number of starters. For a typical league that starts one QB and two RB:QB crank = (Elite games + #1 Tier games - subpar games), max score = 32RB crank = (Elite games + #1 Tier games + #2 Tier games - subpar games). max score = 48Note that this uses number of games rather than percentage, thus giving us a relative score within a position, and not a normalized number that can be used across positions. Use the tier % numbers if you want to compare across positions. This is a more recent change by Matt, and does not appear to be incorporated in a Crank Score calculator across positions that is available at another site.

 

[Matt Waldman]

Great work as usual Matt. However, does the Crank Score and "risk" really go hand in hand? It actually seems like an approach that attempts to minimize risk. I really enjoyed the first two parts of the series and look forward to incorporating the approach to some extent in my re-draft.

It does qualify as a risk in the sense that it's a relatively new strategy and you're trying to target players who are consistent when as a projection method there hasn't been a known approach...I'm hoping to work towards changing that over time.

 

[Dr. Octopus]

Great work as usual Matt. However, does the Crank Score and "risk" really go hand in hand? It actually seems like an approach that attempts to minimize risk. I really enjoyed the first two parts of the series and look forward to incorporating the approach to some extent in my re-draft.

It does qualify as a risk in the sense that it's a relatively new strategy and you're trying to target players who are consistent when as a projection method there hasn't been a known approach...I'm hoping to work towards changing that over time.

Makes sense in that light, thanks.

 

[Lord Fantasy]

@stickboy:

Thanks for trying to make it clearer, but I still don't get it. Maybe it will become more clear when the next article comes out.

CA

 

[stickboy]

@stickboy:Thanks for trying to make it clearer, but I still don't get it. Maybe it will become more clear when the next article comes out.CA

I was using the Crank Scores to help decide on the flex approach for one of my leagues so I had already dug into it based on a mention in another of Matt's articles. Since this required comparing across positions, I needed to fully understand how the magic score was generated vs. the individual percentages.If you can be more specific on where you don't get it, then perhaps Matt could clarify those points for you and others.

 

[greyone]

Im loving the crank. The article reinforces some of my seasoned thoughts and strategies. Difference being I didn't have a cool name for it.

But I don't consider crank a high risk strategy. Im a gambler with a more than an hint of conservatism. Taking players with higher floors is anything but risky.

Your article paints the crank team as a lower scoring team on the surface. While team A is loaded with big names to go along with BIG YEAR END stats. As you say when factoring the real time weekly h-2-h scoring Team B and its lesser names & higher floor with occasional high performance out burst beats out the big guns.

I don't want to waste time arguing semantics. High risk or not. This is great.

Reading both Crank & Jeff Terfetiller's PPR drafting series brought a ton of clarity to eachothers respective piece.

Part 2

 

[greggorymac]

Matt,

Good intro, an up to date summary of the Crank Scores rather than chasing through the articles at your previous site.

1. Will you be publishing a tool that allows us to enter our scoring system, league size, number of years to analyze, etc. to calculate Crank Scores?

2. Have you done any analysis of where consistency and high point projections occur together, and where they diverge? For example, if we are drafting 6th round QB's that are maybe QB7-12 in terms of point projections, are these more or less consistent than other parts of the curve where draft runs occur? What about the volatile WR position, especially in a PPR? Is there part of the point projection curve where I can get better consistency relative to the hot and cold WR's?

3. Are there any insights from the slope of the Crank Scores and where the fall off occurs most?

sb

I agree stickboy, a tool like this would be huge.

 

[Mike]

Matt,

Good intro, an up to date summary of the Crank Scores rather than chasing through the articles at your previous site.

1. Will you be publishing a tool that allows us to enter our scoring system, league size, number of years to analyze, etc. to calculate Crank Scores?

2. Have you done any analysis of where consistency and high point projections occur together, and where they diverge? For example, if we are drafting 6th round QB's that are maybe QB7-12 in terms of point projections, are these more or less consistent than other parts of the curve where draft runs occur? What about the volatile WR position, especially in a PPR? Is there part of the point projection curve where I can get better consistency relative to the hot and cold WR's?

3. Are there any insights from the slope of the Crank Scores and where the fall off occurs most?

sb

I agree stickboy, a tool like this would be huge.

Agreed. Whoever could come up with something like that would be genius.

 

[Lord Fantasy]

A site called FFToday has crank scores you can search. I think you can also personalize it based on your league.

http://www.fftoday.com/tools/crank.php

Now that I have this I don't need to understand how to calculate the actual score, which was what was making my brain hurt.

 

[GDogg]

Matt,

Good intro, an up to date summary of the Crank Scores rather than chasing through the articles at your previous site.

1. Will you be publishing a tool that allows us to enter our scoring system, league size, number of years to analyze, etc. to calculate Crank Scores?

2. Have you done any analysis of where consistency and high point projections occur together, and where they diverge? For example, if we are drafting 6th round QB's that are maybe QB7-12 in terms of point projections, are these more or less consistent than other parts of the curve where draft runs occur? What about the volatile WR position, especially in a PPR? Is there part of the point projection curve where I can get better consistency relative to the hot and cold WR's?

3. Are there any insights from the slope of the Crank Scores and where the fall off occurs most?

sb

I agree stickboy, a tool like this would be huge.

Agreed. Whoever could come up with something like that would be genius.

He had one at his previous site. I've been using the Crank scores for the last few years and used the crank projections for my league last season. Matt is awesome.

 

[gonzobill5]

Matt,

I truly enjoyed this one. Thank you for your devotion to exploring outside the box thinking in fantasy football. One question - in your analysis of FPPG vs. Crank, what assumptions did you make to come to that conclusion? Did you make a direct calculation, run a simulation, or just use last years results?

If you used last year's results, it would be interesting to see if it can be replicated. If it can, i think we might be able to assign a point value to your crank score so that we aren't left wondering who's better between a PPG of 25 with a crank of 13 vs. a PPG of 23 with a crank of 15.

 

[Matt Waldman]

Matt, I truly enjoyed this one. Thank you for your devotion to exploring outside the box thinking in fantasy football. One question - in your analysis of FPPG vs. Crank, what assumptions did you make to come to that conclusion? Did you make a direct calculation, run a simulation, or just use last years results? If you used last year's results, it would be interesting to see if it can be replicated. If it can, i think we might be able to assign a point value to your crank score so that we aren't left wondering who's better between a PPG of 25 with a crank of 13 vs. a PPG of 23 with a crank of 15.

I used last years results and I've replicated it every year I've written about it at the site I used to work and for Fantasy Pro Forecast Magazine for the past four years. However, as I mentioned in the piece the more consistent player with a lower average isn't always going to be outperform the less consistent with a higher average. What this concept shows is that it can happen and happen more frequently than you might imagine. I have an x-value projection system I'm using this year based on VBD and AVT principles. I am also going to look for common factors that the more consistent players possess. I'm hoping this can be used to project future performance in a more systematic way than simply past consistency. I'm glad you enjoyed this series. I think a lot of my colleagues do a great job of covering the nut when it comes to tried and true methods for success. It's fun to look for new ways and I appreciate FBGs offering me that opportunity to do so.

 

[gonzobill5]

Matt, I truly enjoyed this one. Thank you for your devotion to exploring outside the box thinking in fantasy football. One question - in your analysis of FPPG vs. Crank, what assumptions did you make to come to that conclusion? Did you make a direct calculation, run a simulation, or just use last years results? If you used last year's results, it would be interesting to see if it can be replicated. If it can, i think we might be able to assign a point value to your crank score so that we aren't left wondering who's better between a PPG of 25 with a crank of 13 vs. a PPG of 23 with a crank of 15.

I used last years results and I've replicated it every year I've written about it at the site I used to work and for Fantasy Pro Forecast Magazine for the past four years. However, as I mentioned in the piece the more consistent player with a lower average isn't always going to be outperform the less consistent with a higher average. What this concept shows is that it can happen and happen more frequently than you might imagine. I have an x-value projection system I'm using this year based on VBD and AVT principles. I am also going to look for common factors that the more consistent players possess. I'm hoping this can be used to project future performance in a more systematic way than simply past consistency. I'm glad you enjoyed this series. I think a lot of my colleagues do a great job of covering the nut when it comes to tried and true methods for success. It's fun to look for new ways and I appreciate FBGs offering me that opportunity to do so.

Very cool. I am in the process of doing the analysis using two teams with players with identical ppg and different standard deviations. However, as you explained, there are serious limitations to using standard deviation. I'd like to try it using Crank scores, I'll let you know if I make any progress.

 

[gheemony]

My hunch is that you'll find that players who have a higher % of their fantasy points generated by yards will have higher CRANK than those with a higher % created by TDs. This is because a player that averages 6 points/game from yards (e.g., 60 yards rushing) will likely score at least some points in each game (30 yards one game and 90 in the next still averages out to 6 pts/g). In contrast a player that averages 6 points/game from TDs (e.g., a TD a game) will have some games with 0 pts and some games with 12 points. That could push a player from the second tier into the elite tier. So TDs create greater game-to-game volatility.

If you have two players that average the same FPTS/game, take the one that generates more points from yards rather than TDs if you want more consistency (or less variance). This assumes the distribution of yards and TDs will continue to produce about the same means in the future.

 

[Dr. Octopus]

Matt,

Good intro, an up to date summary of the Crank Scores rather than chasing through the articles at your previous site.

1. Will you be publishing a tool that allows us to enter our scoring system, league size, number of years to analyze, etc. to calculate Crank Scores?

2. Have you done any analysis of where consistency and high point projections occur together, and where they diverge? For example, if we are drafting 6th round QB's that are maybe QB7-12 in terms of point projections, are these more or less consistent than other parts of the curve where draft runs occur? What about the volatile WR position, especially in a PPR? Is there part of the point projection curve where I can get better consistency relative to the hot and cold WR's?

3. Are there any insights from the slope of the Crank Scores and where the fall off occurs most?

sb

I agree stickboy, a tool like this would be huge.

Agreed. Whoever could come up with something like that would be genius.

He had one at his previous site. I've been using the Crank scores for the last few years and used the crank projections for my league last season. Matt is awesome.

 

[Matt Waldman]

Matt,

Good intro, an up to date summary of the Crank Scores rather than chasing through the articles at your previous site.

1. Will you be publishing a tool that allows us to enter our scoring system, league size, number of years to analyze, etc. to calculate Crank Scores?

2. Have you done any analysis of where consistency and high point projections occur together, and where they diverge? For example, if we are drafting 6th round QB's that are maybe QB7-12 in terms of point projections, are these more or less consistent than other parts of the curve where draft runs occur? What about the volatile WR position, especially in a PPR? Is there part of the point projection curve where I can get better consistency relative to the hot and cold WR's?

3. Are there any insights from the slope of the Crank Scores and where the fall off occurs most?

sb

I agree stickboy, a tool like this would be huge.

Agreed. Whoever could come up with something like that would be genius.

LOL!!!He is definitely demented. I've got a lot of respect for Mike MacGregor for developing that tool.

 

[GDogg]

Matt,

Good intro, an up to date summary of the Crank Scores rather than chasing through the articles at your previous site.

1. Will you be publishing a tool that allows us to enter our scoring system, league size, number of years to analyze, etc. to calculate Crank Scores?

2. Have you done any analysis of where consistency and high point projections occur together, and where they diverge? For example, if we are drafting 6th round QB's that are maybe QB7-12 in terms of point projections, are these more or less consistent than other parts of the curve where draft runs occur? What about the volatile WR position, especially in a PPR? Is there part of the point projection curve where I can get better consistency relative to the hot and cold WR's?

3. Are there any insights from the slope of the Crank Scores and where the fall off occurs most?

sb

I agree stickboy, a tool like this would be huge.

Agreed. Whoever could come up with something like that would be genius.

LOL!!!He is definitely demented. I've got a lot of respect for Mike MacGregor for developing that tool.

Is that "Mike" Mike MacGregor above?

 

[greggorymac]

Matt,

Good intro, an up to date summary of the Crank Scores rather than chasing through the articles at your previous site.

1. Will you be publishing a tool that allows us to enter our scoring system, league size, number of years to analyze, etc. to calculate Crank Scores?

2. Have you done any analysis of where consistency and high point projections occur together, and where they diverge? For example, if we are drafting 6th round QB's that are maybe QB7-12 in terms of point projections, are these more or less consistent than other parts of the curve where draft runs occur? What about the volatile WR position, especially in a PPR? Is there part of the point projection curve where I can get better consistency relative to the hot and cold WR's?

3. Are there any insights from the slope of the Crank Scores and where the fall off occurs most?

sb

I agree stickboy, a tool like this would be huge.

Agreed. Whoever could come up with something like that would be genius.

LOL!!!He is definitely demented. I've got a lot of respect for Mike MacGregor for developing that tool.

Is that "Mike" Mike MacGregor above?

Pretty funny... do you a link for the app Mike or is it a pay to play application?

 

[Mike]

Whoops, the jig is up. The link above, and here:

http://www.fftoday.com/tools/crank.php

is the program I developed to calculate Crank Scores based on Matt's work. It is free to use to your heart's content.

I actually didn't realize it was still on the FF Today site until Cooley's Angels linked it above, so that surprised me. Enjoy.

 

[stickboy]

Whoops, the jig is up. The link above, and here:

http://www.fftoday.com/tools/crank.php

is the program I developed to calculate Crank Scores based on Matt's work. It is free to use to your heart's content.

I actually didn't realize it was still on the FF Today site until Cooley's Angels linked it above, so that surprised me. Enjoy.

Does it incorporate all the changes to the final Crank Score, or is it an older algorithm?

 

[Mike]

Actually, it is older, before Matt went to the method of adding the number of games in each tier.

 

[Jeff Pasquino]

Good article Matt.

I'm interested in your thoughts on what I did last year in comparison to Crank.

I believe it was a more simplified (basic?) version of your formula.

Here's one example:

Qualty WRs - PPR

 

[Matt Waldman]

Jeff, I remember looking at it last year and liking what you did - it seemed we were on the same wavelength, looking at quality starts and our methods are similar. My first inclination years ago was to factor fantasy points per game into my equation for Crank, but I think it clouded the results a little. What you did is very straight forward and I like it. My approach to our common pursuit has been to make the info dynamic to the current season and flexible for lineup/league types.

I'm hoping to continue chipping away at it until there's a really strong method to project for it. I think I'm going to make good progress this season as I already have spotted some factors that should be interesting reading.

Just an FYI for everyone- here's what you should expect in the coming weeks.

Part IV - Projecting Crank (I'm laying a foundation for projecting consistency with the AVT style method I mentioned earlier).

Part V - What makes a player perform to a high Crank Score? I was going to combine this with Part IV, but I thought it would be better to break it into two columns.

Part VI - First crack at my Crank Projections for '09. I'll update them each week as necessary at the end of my columns throughout August.

 

[Jeff Pasquino]

Jeff, I remember looking at it last year and liking what you did - it seemed we were on the same wavelength, looking at quality starts and our methods are similar. My first inclination years ago was to factor fantasy points per game into my equation for Crank, but I think it clouded the results a little. What you did is very straight forward and I like it. My approach to our common pursuit has been to make the info dynamic to the current season and flexible for lineup/league types.I'm hoping to continue chipping away at it until there's a really strong method to project for it. I think I'm going to make good progress this season as I already have spotted some factors that should be interesting reading.Just an FYI for everyone- here's what you should expect in the coming weeks. Part IV - Projecting Crank (I'm laying a foundation for projecting consistency with the AVT style method I mentioned earlier). Part V - What makes a player perform to a high Crank Score? I was going to combine this with Part IV, but I thought it would be better to break it into two columns. Part VI - First crack at my Crank Projections for '09. I'll update them each week as necessary at the end of my columns throughout August.

Matt. I viewed it as complimentary to Crank and I think both have their place (and are valuable to know).I agree about projecting them as well - that's certainly tougher to do. It's more of a "why" than a math equation though, so it is understandably harder.I'll be doing these Quality Start articles soon as well (probably all before 8/1 if possible).

 

[Genester]