Farmers Insurance Open Picks, Plus a look back at Humana Challenge Projections

by thesanction1

I’m going to keep this short and sweet cause I’m tired as all get out. Manning is in the Super Bowl against Seattle. Can’t wait to see it go down. Machine didn’t speak its mind for Sunday’s games, but I did go 2-for-2 in my own personal picks against my dad. So it’s still possible, if Denver covers in the SB, that we could be splitting the cost of that Ruths Crist dinner… So there’s that.

The Humana Challenge went really well as far as the picks go, another successful week I’d say.  Although I missed the winner by a mile (I don’t think anyone anticipated the record setting performance by Patrick Reed) – most of my fantasy leagues paid out nicely and on DraftDay I split first place in the main GPP.  Not too shabby.  Now DraftKings, a major DFS site, has announced their season opening golf tournament with a $10k guaranteed payout – so I’ll surely roster a team over there.

Anyway – let’s get into it!

Look back at Humana Challenge Fantasy Projection Performance

The same caveats as last week apply.  The comparison picks are from Notorious at Rotogrinders.com, a very well respected name in the daily fantasy sports industry. If you don’t know what these graphs are showing (and you care) then just go to last weeks post and read it.  It’ll be good for you. 

Below is a grid showing Notorious’ draft street points projections (DSP), the converted mean field draft street points projections (%MFDSP), the machine’s %MFDSP and the actual %MFDSP’s:

 

Picks_Humana Challenge_2014_MFP_Grid

Again, like last week, my first impression was that the machine had been beaten out by Notorious’ simple (but very clever) methodology.  But again I was happily surprised to the contrary when I actually did the analysis. In fact, in some ways the Humana Challenge was a better performance than the Sony. The standard method for measuring a projection’s accuracy is called the root-mean-squared-error (RMSE) – and when I calculated the RMSE for both sets of projections here’s what I came up with (lower is better for an RMSE):

Picks_Humana Challenge_2014_RMSE_Grid

So the machine projections have quite a bit lower RMSE  than the projections of Notorious – which is good.  And once again – the most accurate possible projections were a blend of the two. I’ve got algorithms to find the exact optimal blend – this graph shows the outcome of them doing their work:

Picks_Humana Challenge_2014_Blending_Plot

The farther left you are on this line, the more of the machine’s projections you are using, the farther right here represents Notorious. So the optimal blending parameter for Humana turned out to be about 10% Notorious and 90% Machine.  I think this is roughly the split I’ll use personally moving forward.

Lastly – here’s the cut percentage for both of us:

Picks_Humana Challenge_2014_Made_Cut_Grid

And once again – pretty good for both of us!  This time the Machine was much closer to Notorious, beating his projections only by a narrow margin.

The Humana posed some interesting computational issues – it’s an unconventional format, a pro-am with 3 courses to be played, the cut comes in the 3rd round, etc.  This makes most historical data slightly less comparable than it would be for your average tournament.  The projections didn’t seem to suffer, per se, but I did have less confidence going in than I otherwise would have due to these factors.

The same cannot be said for this coming week’s tournament…

Picks for the Farmers Insurance Open

Here they are:

Picks_Farmers Insurance Open_2014A top 10 for K.J.? Pretty bold… Lee Westwood AND Bubba favored over Phil? Seems crazy… I cannot wait to watch. Hopefully these projections help inform some profitable decisions for you before tee-off this Thursday. I’m hoping for a repeat performance on DraftDay, a better showing on DraftStreet, and most important of all – an epic kick-off on DraftKings. Good luck everyone.