*Forty's Facts: An Analysis of Drafting in DSFL

[Forty Jordy]

Welcome to Forty’s Facts where the truth is murky and the media bonus is oh so real.

Now this may be coming in a bit late to impact my draft position in the DSFL draft but it doesn’t quite matter as it was confirmed by Bex, herself, and Gucci that I am 1OA.

In this post we are going to dig into the S21 NSFL draft class, specifically those players played by first time users. We are going to walk through a sampling of these players choosing from various traits such as high TPE players such as Leon McDavid, top DSFL picks such as Cal Lidious, and top NSFL picks such as Otis Allen. I sampled 30 or so players in these categories.

We will look at users’ forum activity looking at total posts and media topics submitted before the DSFL draft and at the time of this post, give or take a day. The DSFL draft occurred on Jan 13th; however, I set a post cutoff date of Jan 9th as the draft takes place behind closed doors prior to the presentation.

We will use this data to understand how pre-draft activity correlates with post draft activity, how forum activity affects draft position and how well GMs draft players into the DSFL by looking at their continued activity and NSFL draft position.

We will first look at the distribution of posts and media topics of prospects prior to the draft.


Looking at this data you will notice that posts have a nice distribution with most players posting 9-19 posts prior to the draft. There are a few standouts on each side. Those with very little posts which is somewhat do to accounts created prior to the cutoff date and those with a lot of posts. Those with more than 20 posts are more likely to have multiple media posts prior to the draft. Looking at media posts, most players have 0 or 1 topics prior to the draft. If they have a post is mostly either a back story or a scouting report.

Now let’s consider whether or not forum activity provides us with any meaningful data about players. We will first look at whether or not GMs used forum activity to aid in their draft picks.



We do see some correlation between number of posts and draft position. There is less correlation below 20 posts, but once a user posts more than 20 times they seem to be picked earlier in the draft.



Media topics seem to expand on the story. Users with 0 topics are all over the place while users with at least 1 topic are generally drafted earlier. Douglas Quaid had 5 posts prior to the draft which seem to reward him with the being the 2OA. The standout in this data seems to be Leon McDavid which had 19 posts and 4 topics, but wasn’t drafted until pick 40.

Now that we see that forum activity is loosely correlated with draft position let’s see how well pre-draft activity predicts continued activity.


There is very little correlation here.



The story is the same for pre-draft media and continued media production. It’s interesting to note that Leon McDavid only had 1 topic after being drafted. It seems difficult to predict a user’s forum activity based on pre-draft activity. This may just be due to timing of when accounts where created in relation to the draft.

Since it is difficult to predict forum activity with continued forum activity, let’s at least see if DSFL GMs did their job and compare DSFL draft position with NSFL draft position. This assumes the NSFL draft position is an accurate representation of a player’s value.


There is actually negative correlation! Did DSFL GMs do a bad job or it is just hard to predict? There are some notable data points in the plot above. There is Cal Lidous who seem to fall off the map after being drafted 1OA for DSFL and then being drafted 53OA in NSFL. Then there is Rotticus Scott who drafted 17OA in NSFL when he was drafted 82OA in DSFL. I believe this is mostly due to Scott creating an account just prior to the draft.

I don’t think we can make much of the previous plot, but let’s see if we can at least see how each DSFL team drafted with our sampling of players. We will plot the distribution of NSFL draft position of each DSFL teams players from the past season.


This is actually quite informative. The teams with the better records generally have players that drafted into the NSFL at better draft positions and the spread of those players is much smaller. This may not be all that surprising as if you can draft well you can have much better success in the league.

Considering the results of the DSFL draft, did current forum activity, predict the overall draft position of these players going into the NSFL?



This plot is quite refreshing after the plots from pre-draft activity. I assume this data is more telling not only because of the stronger correlation, but because now all players had the same amount of time to make posts. Summary: If you want 1OA post 120 times during your DSFL stint!



Roughly the same story for media topics, but just not as correlated. I assume this has to do with posts are more related to TPE gain whereas media posts only provide the green.

That’s all the data I have to share today. Please let me know if I missed anything glaring that would be useful to analyze. I probably won’t do a post like this again as it is a lot of work to track down users’ forum activity.

So here is Forty’s summary, you can’t predict future success of prospects’ future by forum activity, but you can at least have a solid pick if they are very active prior to the draft.

Secondly, if you want a competitive team in the DSFL, roll your dice and hope your players remain active.

Code:

Words 1009

[Troen]

Love the analysis here! Especially neat to see what intuition is actually backed up with outcomes like the activity to NSFL draft position part.

[Exilizer]

Great stats work. Love seeing those charts.