Was it many years before? I'm not sure.

Just started reading the thread you posted. Looks promising. Thanks!

lets go. may have given my computer cancer with 900 pop ups before they loaded but we sweating

damn 2-1, good thing we got + odds

0-2 on Faceoffs today. Our justification for betting these is that they aren't coin-flips but then we win just 28% YTD so maybe we are right but just on the wrong side. Good sweat though.

its ok. i believed in Marcus Smart today again. We all make mistakes.

it is very exciting and bookmaker paid instantly.

I'd heard it in college and then never heard it ever again until silicon valley made a running joke on it and then it became trendy.

Just checked and the Schrodinger joke there predated silicon valley bringing it to the cultural zeitgeist by 3 years. Possible it was big before the HBO show and I just didn't realize because I tend to ignore most pop culture.

As for the DFS thing, the scariest thing about it is that as people have mentioned had they not given block a lineups to one account and block b to another but randomized it then they'd have never been caught. Same with had they just used a random account name. They are only in trouble because very unsurprisingly two Instagram influencers are not the smartest two in the world.

They did dupe a lineup but it was the guy who entered 2x of the same so just a typo or something.

In other news, birdwings retired today and went atweet storm saying it's because of all multiaccouting and cheating and curtis painter has been tweeting his 6 figure losses so looks like a few more pros are dropping out of the game.

Holy crap. That is juicy. Can you link to curtis's twitter account?

I'm not happy with how we are weighting the probabilities of the opening skaters. I was using season-to-date historical opening faceoff data. Using that, we had an 83% chance Phillip Danault shows up for Montreals opening faceoffs (he's been the most heavily used opening faceoff guy in the whole league) and then here comes 5% chance Nate Thompson showing up. So I re-tooled things to use current season neutral zone 5v5 faceoffs. That probably undervalues Danault somewhat but it won't be some crazy event when Nate Thompson does show up too. Any other ideas?

Up 600k then a million dollar downswing to finish the year down 400k

Wow. Reminds me of the sharkscope graphs back in the day. That is gross.

life of 1% edge

FOs Daily: 0-2-0, -2.3u
FOs YTD: 2-6-0, -4.9u, 9.2u risked, -53.3% ROI

GAPs Daily: 5-4-0, +0.18u
GAPs YTD: 91-87-1, +18.81u, 195.26u risked, +9.6% ROI

SOGs Daily: 3-0-0, +3.17u
SOGs YTD: 219-209-2, -1.77u, 523.75u risked, -0.3% ROI

How hard do you press these low theoretical edges? I used to look for a 1% differential/2% edge, but later doubled it to 2%/4% to lower variance. I slept much better.

I wish I were confident enough in my stuff to bet 2% edges.

After some thought, this might be a candidate for the "Benter Boost" (credit @Gingfacekillah for the term, Billy Benter for the concept). I think of NZ 5v5 FO as the pool of available players to choose from. And the actual opening faceoff data as the YTD utilization of that pool. We shouldn't be disregarding the observed opening faceoff probabilities altogether. This is valuable information. So why not a Benter Boost as a way to ensemble the two? Or perhaps we can do some year-over-year correlations on the opening data and regress the observed to the percentage in the pool?

I could be way off base here since I haven't read the book you mentioned. But I looked up the concept and it seems to apply to regression and market odds. I also don't know much about hockey or the model you're using for face offs. But do you think you're accurately capturing regression in your model?

You mentioned betting on a guy who has won 83% of face offs ytd vs a guy who has won 5%. I would naturally want to bet on the 5% guy and fade the 83% myself, assuming face offs are close to a coin flip, mainly because of regression. I would also like to know if 83% is a sustainable FO% in hockey and what is the long term FO% of this guy and same for the 5%.

So I'm concerned that looking at YTD face off win percentage and treating it as predictive going forward is a mistake. Those are ad hoc numbers, not necessarily predictive of the future. In fact one of the ways they could be predictive is in fading outlier positive performances YTD and tailing outlier negative ones.

To put it simply you might be betting on guys who are running hot and betting against guys who are running below EV. This seems like a losing strategy to me, but I might not be understanding your model correctly.

What do we think about Knicks 1Q +4.5 -105


Sent from my iPhone using Tapatalk

Do you have data to test whether its predictive? A stream of face-offs in time sequence?

I don't have the data. I was just suggesting to apply some regression analysis to the current data set and to be cautious of assuming that YTD face off data on its own is predictive of future results.

So does the dfs syndicate need to worry about collusion?

Link to Yahoo winner of DraftKings Millionaire Maker contest colluding with her husband to create 298 unique lineups:

https://sports.yahoo.com/ex-bachelor...234640080.html

Great ideas and feedback. I was using the "Benter Boost" term loosely as a general method to combine two sets of related probability estimates. You are correct, Billy used the idea as a way to ensemble market probabilities with his model's fundamental valuation such that the market probability doesn't end up dominating his model's variables. After some additional thought, I think I have a better idea. Into the lab.

If anyone asks just tell them NO COLLUSION

So here's what I'm thinking.

1) Each day get updated lines/rosters for all teams and the day's games.

2) Predict Opening Faceoff Probability of all active skaters using 5v5 Neutral Zone Faceoff probability. This should be a small net improvement over just using raw opening faceoff data(?). We'll fit a spline or some other cubic model



3) Use a basic Marcel Projection system approach to project all skaters probability of winning a faceoff against league average competition.

4) Use our faceoff win% from (3) and probability of being in opening faceoff from (2) to calc each teams probability of winning opening faceoff. Use a log5 type adjustment to calc win% for that matchup. Bet situations where it's better than -115.

What do you think?

TomG SOGs almost all the way back nice work dude