Daily Fantasy Sports

NHL DFS – Short Slate Game Theory

on

This week marks the beginning of the bye weeks for the NHL, which gives us some less exciting slates than usual.

I haven’t gotten around to much game theory lately, and I want to talk about some things you can do on a short slate.

Much of this I am still testing, but there are others out there that I know swear by it, so I pass it on here. Let’s get to it.


Note that much of these strategies do not apply to FanDuel, as their roster constraints are different and you can’t have more than 4 from any team.

NHL DFS – Short Slate Game Theory

Normal Roster Construction

Embed from Getty Images
On most slates, I build my lineups typically the same way – two full lines, one with a correlated defensemen, a correlated goalie, and whatever defenseman fits in the last slot.

From this point, I will refer to this as 4-3-1 construction.

While this serves me well enough for big slates, it sometimes causes way too many duplicates on smaller slates, ESPECIALLY on 2 and 3 game slates.

We can still make lineups this way if we really like how it looks, but just know that you may be tied with 10+ other people.

We need to find a way to take a line we like (and everyone likes) and still make a contrarian construction with it.

Getting Contrarian

Completing the Power Play Stack

Embed from Getty Images
There are several ways we can alter our strategy here just to get a little different, so I will go over things I look at when I want to play the small slates.

Since we often look for power-play correlation when we choose our lines anyway, it would not be that hard to just extend our correlation to the 5th guy on the power play, right?

For instance, if we liked WPG1 tonight, its feasible on a normal slate that we would run it as the Scheifele-Wheeler-Connor and add-on Trouba for the PP correlation for our 4-3-1. We could simply just add the 5th guy on the power play to our stack just to get a little different (and in this case, its Patrik Laine!).

Some teams make this more difficult, however. A team like the Flyers do two things different on their powerplay that throws a wrench into this strategy. First of all, their top PP unit has no defensemen at all – its all forwards. Okay, we can play all forwards, that isn’t that detrimental.

Those 5 forwards span 3 lines at even strength though.

So we have 2 forwards from PHI1 and PHI2 and Wayne Simmonds from PHI3.

That is not really the best correlation as it is, as we aren’t getting properly exposed to all the even-strength lines we could be.

Further Variations

Embed from Getty Images
We can do a couple of different things to accommodate a situation like this. It isn’t necessary to have ALL 5 PP guys if it doesn’t make sense, so we can eliminate Wayne Simmonds in this case and just have the mini-correlation of Giroux-JVR and Couturier-Voracek. We can still play defensemen here, especially since the Flyers have two pretty solid defensemen that eat a lot of minutes and add a little something on offense in Gostisbehere and Provorov. So we could go with something like Giroux-JVR-Couturier-Voracek-Provorov and it will give us two mini-stacks from the top 2 even strength lines, 4 players from PP1, and a solid defenseman.

This strategy will still get you different on a 3 or 4 game slate, but if you want to get even MORE contrarian, there is a further step to take.

The Monster Stack

NHL DFS - Short Slate Game Theory

In the case of Philadelphia, for last night’s slate, I used a super stack that had as much correlation as I could fit in. I did exactly the same thing as I mentioned before, but I added another guy to complete the PHI1 line – Travis Konecny – which gives me more of that correlation as well as the minor one with Konecny-Provorov on PP2.

Also note that in these monster stacks, I am still stacking the goalie. We are betting on a team scoring multiple goals, so we are essentially betting on our goalie getting a win at least.

Get access to the stats used by the pros at StatRoute. Use promo code F6PNBA for 50% off!


This is something I have seen referred to by other hockey sharps as a “Josh Stack” – as a couple seasons back, Josh Harris stacked the top two lines of the Chicago Blackhawks, as well as Duncan Keith and the goalie to form the 7 man stack, and he took down many contests with.

Implementation of the Monster Stack

Experimentation

Embed from Getty Images
So that was just an explanation of how those stacks look. The next question is – when do we implement these strategies?

Last night, I tried a little experiment. I put 4 entries into the $8 Two-Man Advantage on DK, with each lineup being one of these mega-stacks of each team playing last night. I was curious as to see what could happen and what can go wrong on a two-game slate (everything, obviously).

Though I profited, it wasn’t by much. New Jersey put up 6 goals, with 4 of them coming from the top two lines (mostly the 2nd line), and some depth scoring from the third line and shorthanded. The Penguins scored 3 goals, but none from the top 2 lines. Winnipeg scored 1 goal from their PP2. Philly had 1 player with more than 1 point from the top 2 lines.

In short, you needed to have NJ2 to make any real money in large-field tournaments last night.

NJ1 did pick up a couple points, they just didn’t go off.

So, what is the lesson here?

We need to be more picky about making these monster stacks on short slates. We can’t just jam 6 lineups in tonight with each team stacked to the brim. Let’s try to set some parameters.

Parameters

Embed from Getty Images
Much of our correlation in these stacks involves the power play, so we want to pick a team that has a big power play advantage. Let’s look at the special teams numbers from last night’s slate:

TEAMPP% - SeasonPP% - JanuaryPK% - SeasonPK% - January
New Jersey18.414.384.784.4
Pittsburgh24.818.883.382.1
Winnipeg26.118.881.581.6
Philadelphia13.114.776.181.3

There is… not much to glean from this information. If anything, you could see that Winnipeg has been good all year, and Philadelphia has not been good on the PK. Looking at the more recent numbers – its a lot closer, and there is a lot less of an advantage.

So how did we get on to New Jersey last night?

Something else we often want to check out is opposing goalies. How often do these goalies give up 5+ goals?

That is a little harder to pick up with statistics, you just have to know what goalies are prone to blow up, and what defenses can stumble at times.

Matt Murray has been improved this year from last, but he is still prone to letting in some goals. On top of that, the PIT defense has been leaky at times.

The real way we got on to NJ is how good of a matchup NJ2 has against Malkin’s line. Malkin has been ABYSMAL on defense this season, and there is no sign of stopping.

And what do we do when we like a line on a short slate that may end up being popular or chalky?

Review

Embed from Getty Images
Things we are looking for in our monster stacks:

  1. Good Matchup at Even Strength
  2. Power-Play Advantage
  3. Opposing Goalie Known to Give Up Goals In Bunches
  4. Looking For Leverage On A Short Slate (These stacks aren’t viable on normal-to-large sized slates)

I have only begun testing the monster stacks, but have been making the 5 man’s for some time with some success. It is very much a boom/bust strategy, but you need to get different if you are going to play these shorter slates. There will be several of them in the coming weeks with the bye weeks, so if you want to get a taste of the action – you may want to try this out.

Get access to the stats used by the pros at StatRoute. Use promo code F6PNBA for 50% off!


Kind of a slow week, so I will hit one of those big slates next week with the normal format.

Until then, stay frosty.

>

About Kevin Walsh

Kevin Walsh hails from Pittsburgh, PA and has been playing fantasy sports for 15 years and DFS for 5 years. He is a huge football fan and and even bigger hockey fan. Perfect blend of sports fan and math nerd. Try to look past the Steelers/Penguins bias. Follow him on twitter @FuryOTStorm

Recommended for you

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.