Of all of the formats of daily fantasy baseball, nothing rivals the popularity of large-field tournaments with guaranteed prize pools known as GPPs. A GPP affords the opportunity for a DFS gamer to land a large prize for a relatively small investment by defeating a field size that could run into the thousands. Typically, the real money is concentrated at the top of the leaderboard, so you will have to have a big game out of just about every roster spot in order to take one down.

Unlike cash games such as H2Hs, the goal is to put up a lineup with the highest probability of winning, which runs contrary to putting up a lineup with the highest projected score. I know this sounds illogical, but it is not. For example, if you stack the Colorado Rockies at home against a weak starter on a five-game slate, you are setting a roster of hitters that on paper should be expected to yield the highest score of the night.

If we were playing a game where the projected outcome were guaranteed, we would roster the Colorado stack every single time because we could not possibly put up a higher score. In an MLB DFS game, however, hitters are highly volatile on a night-to-night basis, so game theory dictates that if one set of unpredictable hitters will be highly owned, we should be looking for any reason to fade that set.

Even if the Rockies go off on that five-game slate, it puts us no better than a large portion of the field. Not only does it not assure us a top-10 cash, it very likely doesn’t even assure us a min-cash. If our other surrounding players fizzle, we do not cash. If half of our remaining players go off and half don’t, we are looking at best at a min-cash. If all but two guys go off, we have no shot at a top-tier finish. This is the huge downside of riding the popular wave in a GPP.

On the other hand, if you love a righty-stacked Marlins lineup in Philadelphia facing a bottom of the rotation lefty starter on that same five-game slate, you are more likely to cash for a high prize despite a lower probability of those players going off.

You should always look for that next layer play, as each step away from convention lowers the ownership percentages. For teams in a division with rotations such as New York and Washington’s, you can catch many GPP opponents asleep at the wheel when a team like Miami is coming off a week of getting shut down against monster pitchers and is now heading to Philly.

All the typical opponent of yours will be thinking about is how ESPN mentioned that the Marlins are hitting .177 the past week. There is a term for information that is counter to DFS usefulness: noise. Talking heads that take statistics out of context are some of the best friends of the sharp GPPer.

Another great example of noise is the oft-quoted statistic of what a batter is hitting against a particular starter. You will often hear how ‘Sano is 5-8 with 2 HR against Martinez.’ BvP statistics are mostly complete toss-outs due to sample size. We already know that hitters are volatile on a nightly basis and their true level comes out over a long sample of plate appearances.

There can be literally hundreds of reasons why a hitter got the measure of a given starter on one or two nights and most of those reasons are not applicable to the next event.

One strategy enjoyed by many good GPP gamers is multi-entry. In large-field tournaments, you can place a number of entries up to a predetermined limit. On nights where you like a few under-the-radar roster plays, you can put in entries using each one as your roster base. Odds are that as solid as your research and reasoning are, all of the lineups will not go off, so it is a comfort to have each one you liked in an entry.

In-contest adjustments are a unique feature to DraftKings and should be used strategically. If one late game is left and you’re in 30th place with Mike Trout left to play, take a look at how many of the lineups ahead of you have a player left as well. Quick calculations based on the open roster position can give you a good idea of who they have left.

If 20 people ahead of you also have Trout, your highest possible finish is 21st place; even if Trout hits five home runs, your ceiling is still capped. You can maximize your possible placing at that point by pivoting off Trout to another outfielder, and all of them will be cheaper so your cap is safe. None will be a better player, but in the high-variance nature of a single game for a hitter, it’s a move I’d make every time. Remember, all of the real money in a GPP is in the top spots.

In terms of GPP roster construction, you will have to stack hitters to win. You are looking to maximize upside, and the best way to achieve that is to roster a lineup of teammates who give each other the highest exposure to potential points.

When Bryce Harper hits one out, you want the two guys who were on base at the time to score for you as well. If the opposing starter shuts them down, you will not cash; this is fine and it’s the price to pay for the upside you need to hit the GPP leaderboard. It is a boom-or-bust format and you should be swinging for the fences so to speak.

There is a difference between contrarian and crazy and while you will want to dig for some upside potential with your hitters, your starting pitchers are another story. The dependability of high-quality starters to replicate their typical performance from game to game makes them playable in all formats, even GPPs.

When you can find good reason for a fade, such as a game in a tough hitter’s park or a quality starter who has put together a few consecutive starts of reduced K/9 and might be battling a tired arm, then by all means take the shot against them, but do not be contrarian at starting pitcher just for the sake of being contrarian.

DraftKings requires a second starting pitcher spot, so an under-the-radar starter who can get strikeouts and is in a good matchup; for example, a lefty facing a heavy lefty lineup makes for a nice SP2 plug-in.

Now, take what you’ve learned and sign up for DraftKings. DraftKings offers a 100% up to $600 sign-up bonus that’s released as you play. Use the code P5S when you create your account.