Re-Analyzing Trends in FanDuel's Perfect Daily Fantasy NBA Lineups

For the crowd that plays the daily version of fantasy basketball, the All-Star break can be bitter.

After 100-or-so days of grinding, creating lineups, and reacting to last-minute news, it all comes to a temporary end.

Sure you can take a crack at NHL, but you have to be careful not to let your bankroll dwindle too much because -- in just a few more hours -- it all begins again.

We've  already broken down some of the biggest trends that exist in the after-the-fact perfect lineups on FanDuel through the end of 2015, but with another 40-ish slates since then and time to brush up for the second-half push, why not see how things look now?

Based on the 106 NBA slates since the season tipped off, here are some interesting things to note from the optimal lineups on FanDuel.

FanDuel Perfect Lineup Analysis

Just so that we're on the same page -- in case you're new to DFS -- FanDuel requires you to roster a center and two of the other four positions (point guard, shooting guard, small forward, and power forward) while staying at or under the $60,000 salary cap.

Good? Good. Let's move on.

Leaving Money on the Table

In our first iteration, we saw that 10 of the 64 of the NBA slates through December 31st actually used all $60,000 of the cap -- 15.6% of them. Through the 106 slates leading into the All-Star break, a total of 16 used all of the cap, 15.1%.

In the 56 instances when there were seven or fewer games on the docket, only seven maxed out the salary cap (12.5%). On the 50 slates with eight or more games, nine totaled $60,000 (18%).

That suggests that zeroing out your balance doesn't need to be a top priority.

On the other hand, 70 of the 106 (66%) lineups totaled at least $59,500 in salary, so wasting too much salary might be limiting your upside.

Only 10 left more than $1,000 on the table (9.4%). Only two of those came when the slate was bigger than seven games, and only two came after Christmas Day (both were seven-game slates).

To Stack, or Not To Stack

That is the question: whether 'tis nobler in the mind to suffer the slings and arrows of outrageous fortu...wait.

Stacking. Right. Stacking in NBA contests seems to be counterintuitive because stats in basketball aren't generally linked. Aside from the assist-to-score, we're seeing Player A score, meaning Player B didn't score. Why, then, roster players from the same team if you can avoid it?

Well, it turns out that teammates are quite common in optimal lineups, even when the slate is fairly large. Of the 106 slates up until the All-Star break, 76 featured at least six games. Of those 76, only six were absent of at least one pair of teammates.

23 had exactly one pair of teammates. 33 had exactly two pairs. Six had three pairs exactly, so there have been as many optimals on six-plus game slates without any teammates as there were with three separate pairs.

Four of these lineups had a trio of teammates, and three featured a trio plus another duo.

One even had two duos and a trio of teammates from three different teams.

But what about the pricing?

Is it just that we see value teammates in a high-scoring game make the optimal? Is it two stars? What gives? Here's the breakdown of how the teammates are paired -- again, this comes from slates with at least six games. (To keep things easy, I omitted the instances of trios.)

Value players are key, yes, but you don't always need to avoid rostering mid-to-high-priced players from the same squad.

Stacking has its shortcomings. That's true. But over this sizable sample, we're seeing that it's a fairly common component to the highest-possible-scoring lineups.

It might feel like bad process to target teammates (and it would be if you target a game with a low over/under or a game that features two slow-paced teams), but when it hits, it can hit big.

Short Slates

Like I did in the prior installment, I'm drawing the line at seven games for a "short slate" so that we can try to keep bigger samples.

That gives us 56 slates. Of the 504 players on these 56 lineups, 95 cost $4,500 or less (18.85%). 74 (14.68%) cost at least $9,000.

Most of the value players came from three positions (though if you double the center spot, we're looking at just one outlier).

Value point guards haven't made their way to the optimal lineup very often, which seems odd to me, as these players "should" see an uptick in usage. Then again, perhaps the default to the facilitator role caps upside, so you might be best off looking elsewhere for value. Especially when you see where the star performers tend to play.

So, yeah, there are a good many high-priced point guards (and really only James Harden makes the cut at shooting guard most nights), but that doesn't mean that avoiding them has been the smartest play this season. Star point guards and small forwards have dominated the optimals when the slate has been fairly small.

Big Slates

Those small-slate trends are somewhat transferrable to the bigger slates.

94 of 450 players were priced at or below $4,500 (20.89%). 74 (16.44%) were $9,000 or more, so it skews a little more to the extremes on the bigger slates, likely because more value players are seeing court time. Based on hindsight, you should be using that to get some of those elite point guards in your lineups.

The frequency of elite point guards in the optimals, again, is contingent on the number of point guards priced that high on most nights, but even still, fitting them in is certainly not a bad idea.

Lineup Breakdowns

So, how do these lineups tend to look when all is said and done?

On small slates, the stars-and-scrubs approach (two or more players priced at or above $9,000 as well as at or below $4,500) on the top right has been the most common makeup. One star and multiple scrubs has been a distant second.

Everything else is a distant third and beyond. Finding value is key.

Big slates are much the same, though one star and one scrub is the second-most common setup. Only three lineups out of 50 (6%) did not feature a player below the bargain threshold.

Takeaways

You don't need an optimal lineup to be successful. Far from it. But studying optimal lineups can show us what happens when you piece together the best actual team possible.

Don't avoid teammates like the plague, and always seek value when you can. Just maybe not at the point guard position.