11 Responses

1. MikeB at |

Tanner awesome write up. My buddy and I are doing a \$400 draft and hold league and I was trying to use your original Create Your Own Rankings guide and I got stuck with the SGP. This helps out a ton!

1. Tanner at |

Thanks, Mike! Good luck in the league.

2. Jason at |

Maybe I’m wrong, but isn’t it a bad idea to use other league data. SGP caters to your history by analyzing the trend of category stats for how your league build their teams. When you pull vast averages from other leagues, you are using their trends instead of how your league trends.

That seems like a bad path to go down. Am I I wrong?

My league trends change from one season to another. For example in HR alone, the slope was 8.7, 10.5, 11.2, 10.7, 11.1 and in 2015: 11.7. Of course, I can trust my average slope over all those seasons of 10.6 or trust the last 3 seasons (11.1), or use a marcel weight on the last 3 seasons (11.25), or trust the NFBC of 8.42 (converted to a 10 team league = 10.10). The truth is that I can use any of these, they’re all pretty close to each other. But since my league is trending up I like the idea of using Marcel. For those who don’t know what that is, you weight the last season the most. =((11.7*5)+(11.1*4)+(10.7*3))/12 — 12 equals the weights added up (5+4+3 = 12).

The point is, aren’t you better off focusing on your history verse pulling averages from other sources? If I’m wrong and averages are a good starting point, then why aren’t they more readily available for use? For that matter, look at how much they have changed from your blog on SGP in 2014, HR denoms were 10.0, 10.4 from Razzball… and for 2015 NFBC Online for 2015 its 8.4. Almost 2 full points different, that’s a big swing in SGP totals. These numbers do matter, because they are being used to quantify the player’s value. So to use SGP as a primary valuation method, they need to be as accurate as possible… am I correct?

I do love this site… great stuff as always Tanner.

1. Tanner at |

Hi Jason, you ask a very good question.

When I wrote this article, I intended for it to be something those without sufficient (or any) league history to use. Some folks switch leagues, just start playing, or are just trying to learn SGP for the first time and can’t locate prior year data. So I wanted to provide them with something that seems more reliable than just my own league observations.

Coming into this, I also would have agreed with everything you said. That you’re better off using your own league history because then the values are tailored to your specific league’s tendencies.

But I’m very close to publishing my next piece that is not fully supporting those themes. Even though three different roto leagues might have HR SGP denominators of 10.0, 10.4, and 8.4, the end dollar values will be closer than you’d think.

Thinking about this in another angle. Assume you’re in a 12-team league that you’d consider “semi-competitive”. I could see an argument for using 12-team data from the “highly-competitive” setup of the NFBC format over your own history. For two reasons. One being that a team designed to win a “highly-competitive” league should also be able to win in a “semi-competitive” environment. And second, using NFBC data will eliminate a lot of the noise and statistical variation that must come from only three years of history from a single league.

I’m more speculating on this last paragraph. I haven’t looked into it at all yet. But they’re just concepts I’m wrestling with in my head.

The post I referenced earlier should be coming very soon.

Thanks for the well-thought question. It’s fun to think about these topics.

3. Jason at |
4. […] translate that into league standings information for each category. Once I had that, I went through this process to determine averages for each category place (that link will take you to various mixed-league NFBC […]

5. […] spent a lot of time acquiring and analyzing standings data of late (AL-only, NL-only, mixed).  And looking at that information got me to thinking that I could easily determine if finishing […]

6. keith at |

Thank so much for doing this! This is incredibly helpful. Small edit on the 2015 draft champions spreadsheet. On the wins tab, the entry DOUGHBOYS \$150 Draft Champions Nov 20 1:00 pm Lg.2977 is in there twice.

1. Tanner at |

You’re welcome! Thanks, Keith. I’ll take a look at that and try to get it fixed.

7. RJ at |

Tanner,
I have yet to find a post on your site that is not useful, so thank you so much for these articles. As a fantasy baseball newby your articles make a lot of terms and calculations easy to understand, and as such make FB more fun!

Dealing with the same problem as mentioned in the comments above (not having any historical SGP data) I am currently working on a Tableau report that shows player SGPs based on multiple data sources (2.14 season, 2015 season and multiple 2016 season projections) and user set SGP values. I will let you know when the first version is up and running, as I would love for you to have a look and provide feedback.

Cheers,
RJ