DFS Roundtable: Starting High Owned Players in GPPs

As part of our preseason roundtables, we asked our staff the following question. When do you start highly owned players in tournaments?

Jason Wood: If you look at who wins GPPs, it's not like they have a completely diverse lineup. Far from it. In fact, many times I'll see who wins the Millionaire Maker and the lineup seems so...logical...it confounds. My point being, not only should you be comfortable rostering heavily owned players, I don't think you can afford not to. 

The trick is to combine two or three consensus plays who have major upside with diverse alternatives.

David Dodds:

My tournament (GPP) formula is this:

Can this player hit at least 3X (FD) or 4X (DK) a higher percentage than his ownership level?  So if Julio Jones costs $8,300 and is 40% owned, can he hit 24.9 FP 40% of the time?  With 1/2 point PPR scoring, this would take a Herculean effort of something like 10 catches, 140 yards, and a TD.  Not impossible, but probably not going to happen 40% of the time either.

The answer is to nearly always fade the highest-priced stars in GPPs and create rosters with more people towards the middle of the pricing at the position. 

The same example with a min-priced running back who is starting because of an injury and costs just $5,000.  He just needs 5 catches, 65 combined yards, and a TD to reach 3X value on FanDuel.  That could easily lead to a situation where I think the player would hit that target 40+% of the time.

Phil Alexander: David's strategy is right on the money. GPP ownership decisions are all about weighing the implied odds of the player hitting value vs. his projected ownership percentage -- a concept many people miss. This line of thinking doesn't just apply to deciding when to fade a stud, it also helps you choose ideal GPP plays.

Let's suppose A.J. Green is facing a tough defensive matchup, his DraftKings salary is $7,500 (about $500 less than his current Week 1 price), there are a few other wide receiver studs with better matchups in his price range, and you project his ownership will be about 5% as a result. Should you play him at low ownership despite the bad matchup? Using Dodds' 4x multiplier Green would need 30 points to hit value. 30 points are no easy task for any wide receiver -- especially in a tough matchup. Green topped 30 DraftKings points in 33% of his games just last season. Ask yourself this: If the game were played 100 times, would Green hit 30 DraftKings points in more than five? If the answer is yes (it is), then you've found what you're looking for in GPPs -- massive upside, low ownership.

Now just because Green's probability of a blow-up game is much greater than his percent owned, it doesn't mean he's some sort of lock to score you 30 points. But it does mean rostering him is the correct decision over the long haul. Provided your ownership projections are consistently in the right ball park, you will profit over time.

Shout out to our John Lee, who was the first person I saw put these concepts into words I could understand. Without his advice, I'd probably still be the one seeking answers to questions about ownership rather than the one writing them.

Justin Howe: I tend to worry a little less about ownership than many others do. As Jason pointed out, winning GPP lineups are often put together logically and feature relatively chalky plays. The key is often in the construction. David is dead-on by pointing out that the high-priced studs are usually unlikely to hit their 4x or 5x multipliers. My GPP lineups are often weighted heavily around the middle of the salary pack, bloated around the $5.5K-$6.5K range. This tends to land me a lineup that's diversified plenty on its own, without forcing me to chase high-dollar guys (whose ownership is harder to project anyway).

Remember: "highly-owned" doesn't mean "more likely to excel." The NFL is wildly random, and high ownership is a product of perceived value, not of chances of success. If you're chasing a GPP prize pool, don't be afraid to fade the popular plays.

BJ Vanderwoude:

Aside from the reasons mentioned above, I also look to the entry limits of each tournament.  For instance, a single entry tournament provides a unique opportunity to fade a highly owned player and immediately separate yourself from the pack. On the flipside, in tournaments where I have 10+ entries, I am much more likely to include a highly owned player in my lineup variations. This is especially true for the players with massive ceilings, as it is very difficult to win a large GPP without having the one player who goes off for 40+ points in a given week.  

Jeff Pasquino:

I am with Justin on this one.  While I get David's math approach, if I am going to build a lineup that I think can put up a big number, I will care a lot less about uniqueness and just build that lineup.  Only in mega-GPPs (ones with 5,000+ entries) will I start to even consider uniqueness.

I would rather construct my lineup differently in another way (build with stud WRs when everyone wants stud RBs that week, or leave extra money on the table, or pick a contrarian game script and build around that).  Chasing uniqueness more often than not leads to a suboptimal play - but if you can find a player that few others are on and if he goes off, well, that can win you a tournament.  Just make sure you play that uncommon choice because you think he is a good play, not because less than 5% will be on him.  Those other 95% might be right about that decision.

I have had some monster weeks going nearly all in on a single player.  Jordan Howard's three touchdowns last year was a fantastic week.  If you are convinced that someone will be a good play - and the downside risk isn't too massive - then that is a great time to be "overexposed" to a given player. The rewards can be high.

Danny Tuccitto: My algorithm is the same as David's and Phil's: If projected ownership rate is less than projected probability of achieving tournament value, then include in my potential player selection. One thing I'll add that I think's worth noting is that these two considerations, ownership rate and value probability, are positively correlated, especially at the margins.

With that in mind, consider a quick scenario. Focusing only on the margins (aka the outliers), we only have four situations when doing this ownership vs. value calculus for a specific player:

Player X has high ownership and high value.
Player X has low ownership and high value.
Player X has high ownership and low value.
Player X has low ownership and low value.

Obviously, we'd always want to roster Player X when he's in Situation 2 and never roster him when he's in Situation 3 or 4. That leaves Situation 1, which is what this discussion is all about. But here's the thing: The correlation between ownership and value is precisely what this discussion is all about. Situations 2 and 3 represent negative correlation, which doesn't exist at the margins. (Players offering massive value will never have low ownership, and vice versa.) Situation 4 represents positive correlation, but it's exceedingly rare. Therefore, the only ownership vs. value decisions we ever make -- for all intents and purposes -- are when players offer massive value, but have massive projected ownership.

So what am I getting at here?

If the only real decisions we're making are when a player is in the high-ownership/high-value situation, then we're fading them some non-zero portion of the time. But if we're fading these players a non-zero portion of the time, then -- if our projections are good -- we're necessarily missing out on massive value. And if we're missing out on massive value, then we're not maximizing our tournament profits.

As I said at the top, I don't go this deep into the woods in practice: "If projected ownership rate is less than projected probability of achieving tournament value, then include in my player universe; else, exclude." It is a conundrum I do think about, though, as the outer edges of one's mind can be a place to find thin edges in the real world.

Chris Feery: If I’m sold on a chalky player for the week, I’ll go ahead and roster him without worrying too much about ownership percentage. I’m confident I can find diversity elsewhere, and sometimes the consensus plays are right on the money. That being said, I’m mindful of not being too heavy on the week’s chalk. Jason mentioned a solid benchmark in that regard: two to three consensus plays is just fine. If and when they go off, you’re not too far off the pace. If one or two only give you an average return, you’re not dead in the water as plenty of other lineups have the same problem.

For a general lineup building strategy, I’m looking for the aforementioned highly-owned players plus two to three value plays that I’m really sold on. I’ll fill the remaining spots with some contrarian plays. That approach provides you with more than enough diversity. Nailing a player or two with less than 10 percent ownership percentage can really help your lineup move up the leaderboard if all else goes according to plan.

Steve Buzzard: Danny is right on the mark with his comment and exactly the type of questions you should be asking yourself if you want to win some of the biggest tournaments.

It's easy to say I won't play guys at high ownership levels or I don't care about ownership levels but in reality we are debating between the tougher decisions when setting our lineup. Danny's quadrant 1 is where most decisions lie and what you need to wrestle with.

If Dez Bryant is going to be 40% owned and has a 45% chance of hitting value. What if he is 50% owned? What if he really only has 25% chance of hitting value?

I will be publishing projected ownership rates this year and they are arguably one of the most important tools you can have. How you use them can go along way towards your success this year.

Dan Hindery: My general GPP approach with players I expect to be highly owned is to spend extra time on them and really try to figure out whether I want to bet on them or against them. Then put them into one of three bins:

1. Green- These are my "all in" guys. In this example, I might go with 75% ownership if I'm expecting 40% to get ahead of the field. Sometimes the highly-owned guys are absolutely the right play and you can actually separate yourself from the field by giving yourself extra chances to put together the right combination of surrounding players.

2. Yellow- These are my "punts." Sometimes I just have a really hard time figuring out how I feel about a guy. I understand why other people like them and see the arguments both for and against them having a big week. I don't feel strongly enough to go all in but I don't feel so strongly that I want to bet against them either. In this example, I'd shoot for about 40% owership and try to find my edge elsewhere. Especially if it is a player who is capable of having a monster week (like Le'Veon Bell) where I'm terrified to go strongly against them. I try not to put too many guys into this category so I don't end up basically right alongside the field.

3. Red- These are my "fades." Sometimes I just think the crowd is wrong. Other times, I think there is a lower-owned player who has a similar floor and ceiling at the same position and I'm willing to pivot and go all-in on the lower owned player. It's not just about the highly-owned guy and his projections, but how I fell about the position in general in that week and my other options. In this example, I might fade the 40% guy by only having 5-10% ownership.

In this process, I try to look at projections, matchups, statistics, etc. and agree with the posts above. But at the end of the day, it also comes down to a gut feeling and trying to take a stand one way or the other on most of the popular plays.