# MLB FORUM

• zackthys

Let’s say, hypothetically, you decided to put 250 lineups into a GPP, and you were stacking. You’re going to stack 4 batters from one team, and mix up the other 4 batters. Would it be a higher EV strategy to stack the same team for all 250 lineups? Or would it be better to stack one team for 125 lineups and one team for the other 125? Or 50 lineups for 5 different teams etc.? My guess is that there comes a point where the law of diminishing returns kicks in, where adding an additional stacked lineup for one team does not increase the probability of winning the GPP enough to justify the entry fee for that additional lineup. I guess ultimately the question is “at what number of stacked lineups for one team is your EV maximized, and whether or not this number is greater than 250 which is usually the maximum amount of lineups allowed.” If it is higher than 250, then obviously the optimal play would be to stack one team for all 250 lineups.
I know a huge factor to play a part in this is game and slate dependent, but let’s ignore ownership and batter prices and just pretend stack A has a slightly higher chance of going off for 10 runs than stack B, which in turn has a slightly higher chance of going off than stack C, and so on.

• yanks237

Going to take a simplified statistics look at this with saying your keeping the same four player stack for each lineup.

That means you have 4 other roster positions to choose. Say for each position you have narrowed it down to a short list of 4 players you want in your lineups. You then have 4^4 or 256 combinations of those players. Meaning you would need 256 lineups to have the stack with your 16 player sized pool for the four positions.

Suppose you can reduce it to 2 players per position you are now at 2^4 combinations or 32 lineups. I think the issue isn’t going to be with the stack, rather your player pool. I’m not sure mathematically but I would rather narrow down the player pool and test out different stacks on a slate versus more player combinations.

Let me know if you have any other questions I will try my best.

• getlucky4

I think since your playing for first I assume that you should have as much variation as the slate allows. I don’t agree at all that your second lineup is worse than your first one. There is way too much luck involved for that to be true. If there wasn’t you would just see trains of the same team from every tournament player.

• TomCats83

Just my personal opinion here (and for the record I rarely enter more than 3-4 lineups into any tourny so i’m no expert on this matter) but if I was going to go balls out and put in 250 lineups I would stack the worst stacks I could with that 250 and combine those terrible stacks with the best chalk picks.

Take tomorrow for example… these are the best 5 guys in no particular order:

I would make 50 lineups with stacks for the 5 teams these players are going against – Nats, Pirates, Rangers, Dodgers, and White Sox and then i’d fill in the other spots with chalk picks.

In fact, I would very much consider taking the opposing pitchers against those 5 chalk guys as well because if the bad stacks hit then odds are their pitcher will win as well so it’s a double contrarian play – no one will have the stack and certainly no one will pick the pitcher going against any one of those 5.

My logic being that if you’re going to blow \$250-\$750 on a whole bunch of dart throws you’re going to need to take 1st place (or at least Top 10) in order for the return to make any sense. The odds on getting first in these large tournys with 250 chalk lineups, and a bunch of different combos of those chalk lineups, is less likely than throwing out 250 completely contrarian stacks instead.

Most normal people will play the chalk stacks that make sense – odds are out of thousands of lineups someone will have matched the right chalk combo and will kill all of your 250 lineups at once.

However, the vast majority will not stack the terrible stacks and if one of them hits and your combo of chalk picks within that lineup hits as well then your odds on winning are greatly increased because no one in their right mind would stack against Jose Fernandez.

I’ve thought about trying this but frankly I think you’d lose night after night after night until you finally hit one and then the question would become whether or not the consistent loss of \$250-\$750 was worth whatever prize payout you eventually got. Bottom line is I feel there’s more effective ways to spend \$250-\$750 every night (like with a nice combo of H2H and GPP plays).

Having said that the best finish I’ve ever had was last Tuesday – I took home 1st place in the FD Express slate (\$3 game – 3853 player pool). I won with Tomlin, Weiters, Napoli, Altuve, Valbuena, Lindor, Davis, Springer, Mazara.

The reason I won is because Tomlin went 8IP against Sale and Napoli, Lindor, and Davis all hit him – All 4 players were under 5% owned. Being a die hard White Sox fan I knew Sale had struggled his entire career against the Indians so I deliberately played the \$3 lineup on the off-chance i’d be right. Luckily I was and I took home \$1000.

I don’t think doing this every night is a sound strategy but if you’re going to fire off shotgun shells night after night you might as well do it in a way no one else will…

• getlucky4

Taking players in bad spots isn’t going to make you money in the long run if you keep playing the same volume. The line is pretty thin on what is good and bad in baseball though when you factor everything in.

• phillygamecox

@getlucky4 said...

Taking players in bad spots isn’t going to make you money in the long run if you keep playing the same volume. The line is pretty thin on what is good and bad in baseball though when you factor everything in.

Agreed 100% here. While it might not be a bad idea to throw a few lineups of rosters going up against the top pitchers of the day, I’d venture to say 8 times out of 10, the winner of a large GPP had a fairly chalky stack with low owned players around it – sort of the opposite of the always stack against the top pitchers approach. Certainly if one of those large pitchers gets blown up, you likely will head to the top of the leader board—but, I’ve found that Vegas is pretty darn good at this stuff and pitchers are highly favored for a reason. :)

I’ve tried all different stacking approaches, and even having taken down several over the last year or so – I still don’t have a preferred option. To be honest, one of the more generic options is to simply pick say the top 5 projected run scoring teams and stack them 2,3,5,10, etc..ways – whatever you feel most comfortable with. Maybe you have 5 pitchers you really like – intersperse them into the rosters as well. I just wouldn’t recommend doing this against every top pitcher every single night. :)

Another ideal thing to do is look at how the pros roster their teams. Most nights, you’ll see a guy like Max/Saahil pick his core set of players and simply build countless rosters around that. When it tanks, it tanks…but when it hits – you take down 1st, 7th, 12th, 18th, etc…

Another point is to not get too cute with pitching at the expense of putting in a stellar hitting stack. 99.9% of the time, all the hitting in the world won;t make up for a horrendous pitcher.

One last thing to make sure – if you are throwing a ton of rosters into play, start small. Maybe the .25 league if you can on DK or the \$1 Sac Fly on FD—try a few different things and figure out the best strategy for you. No matter what, you will likely lose more than you win – but by practicing smart bankroll strategy, you’ll eventually get a night when the stars align that pays for your entire summer of play. Good luck.

• TinkyTyler

• 79

• Ranked #57

• ### 2018 DraftKings FFWC Finalist

@getlucky4 said...

I think since your playing for first I assume that you should have as much variation as the slate allows. I don’t agree at all that your second lineup is worse than your first one. There is way too much luck involved for that to be true. If there wasn’t you would just see trains of the same team from every tournament player.

You do want variation yes, but I don’t completely agree with your 2nd comment.

If I’m playing 200 lineups, I know that there exists some ranking of those 200 lineups in terms of expected return, expected points scored, whatever. I might not know what that ranking is, but this isn’t terribly important. The key point is that, assuming you’re a winning player, even if each additional lineup you play has a slightly worse (but still positive) return than the one previous, that’s OK… the payout structure of GPPs makes the only important goal is to win the event. By playing more lineups you’re inherently accepting a lower EV per lineup, but drastically reducing variance which in turn lets you play for way more \$. Even though I know one of my lineups is objectively “better” than the rest, playing a train of it doesn’t really make sense: I’m not reducing variance at all, and in fact I’m altering the payout structure for myself in a detrimental way (the ROI of each additional tied lineup in the train is worse because you’re sharing with yourself essentially). I know a good single-entry player has a higher individual ROI per lineup than the multi-entry player, but the multi-entry player is going to realize their long term ROI quicker. There’s no best approach.

OP: The answer to your question comes down to what kind of risk tolerance you have. If you believe you have the “best” stack (obviously you have to take into account ownership), and you’re playing for an amount you’re comfortable losing, playing that stack 250 times is totally fine… there are WAY MORE than 250 reasonable combinations of a single stack as long as it’s a reasonably sized slate. If you have a few stacks that you think are good, dividing it up across those is a reasonable, more risk-averse approach, that lets you risk more \$ than the 1 stack does. I don’t think the right answer here is obvious, and it totally depends on what kind of risk you’re willing to accept.

• getlucky4

I understand your point and your technically correct but if you only play tournaments your going to enter more than 1 lineup.

• Razzle11

@TomCats83 said...

Just my personal opinion here (and for the record I rarely enter more than 3-4 lineups into any tourny so i’m no expert on this matter) but if I was going to go balls out and put in 250 lineups I would stack the worst stacks I could with that 250 and combine those terrible stacks with the best chalk picks.

Take tomorrow for example… these are the best 5 guys in no particular order:

I would make 50 lineups with stacks for the 5 teams these players are going against – Nats, Pirates, Rangers, Dodgers, and White Sox and then i’d fill in the other spots with chalk picks.

In fact, I would very much consider taking the opposing pitchers against those 5 chalk guys as well because if the bad stacks hit then odds are their pitcher will win as well so it’s a double contrarian play – no one will have the stack and certainly no one will pick the pitcher going against any one of those 5.

My logic being that if you’re going to blow \$250-\$750 on a whole bunch of dart throws you’re going to need to take 1st place (or at least Top 10) in order for the return to make any sense. The odds on getting first in these large tournys with 250 chalk lineups, and a bunch of different combos of those chalk lineups, is less likely than throwing out 250 completely contrarian stacks instead.

Most normal people will play the chalk stacks that make sense – odds are out of thousands of lineups someone will have matched the right chalk combo and will kill all of your 250 lineups at once.

However, the vast majority will not stack the terrible stacks and if one of them hits and your combo of chalk picks within that lineup hits as well then your odds on winning are greatly increased because no one in their right mind would stack against Jose Fernandez.

I’ve thought about trying this but frankly I think you’d lose night after night after night until you finally hit one and then the question would become whether or not the consistent loss of \$250-\$750 was worth whatever prize payout you eventually got. Bottom line is I feel there’s more effective ways to spend \$250-\$750 every night (like with a nice combo of H2H and GPP plays).

Having said that the best finish I’ve ever had was last Tuesday – I took home 1st place in the FD Express slate (\$3 game – 3853 player pool). I won with Tomlin, Weiters, Napoli, Altuve, Valbuena, Lindor, Davis, Springer, Mazara.

The reason I won is because Tomlin went 8IP against Sale and Napoli, Lindor, and Davis all hit him – All 4 players were under 5% owned. Being a die hard White Sox fan I knew Sale had struggled his entire career against the Indians so I deliberately played the \$3 lineup on the off-chance i’d be right. Luckily I was and I took home \$1000.

I don’t think doing this every night is a sound strategy but if you’re going to fire off shotgun shells night after night you might as well do it in a way no one else will…

This is FAR from the optimal strategy that was being asked about. In fact this post shows me that Mass multi-entering is a skill that not all DFS players have.

• Heterodox

@TomCats83 said...

My logic being that if you’re going to blow \$250-\$750 on a whole bunch of dart throws

Who’s talking about blowing money or throwing darts? Is that what you think Saahil, Yudacao, Chipotle, etc are doing when they max enter? I imagine they not only have fairly precise odds calculations of each player on their rosters reaching a certain level of production, but they probably have also simulated the entire contest to work out what the best combination of variance and reward is. They know exactly what their EV is (well, to the extent that it can be determined), and how it compares to the EV of doing something else, and they do these calculations every single slate.

So, if we accept that stacking is a way to solve a probability problem, because you have a way better chance of getting a stack right than picking all individual players, you should be able to get a value for each stack and compare. Just average the odds. It might be that on a given slate there are some that are close in value while on other slates one stands above. After that, just randomize the rest of the spots.

I’ve seen so much talk about “Anti-fragility” lately, and I think people have it completely wrong. They use that concept to mean simply using players/stacks that others aren’t on, when all it means is that you should set yourself up to benefit from randomness. Pick the optimal stack, and the optimal pitchers, because those are what you have the best odds of being right about, then randomize the remaining spots. I’m pretty sure this is what Saahil does, and further to the point about pros quantifying the value of their plays, when he stacked against his own pitchers last season (haven’t seen him do it this year, but last year he did it enough that someone started a thread about it), he wasn’t JUST doing it because no one else was and he figured it might have worked. He was doing it because his calculations for those slates told him that it was optimal, given all the information relevant to those slates.

I don’t mean to make it sound simple or easy. I don’t know how to actually handicap the players or the field, and there might be slates where your best “thing” doesn’t give you a huge edge over the field, and it’s actually better to play for as many cashes as possible, since a win will be pure luck that night. On a night like that, I would think that one line up entered 250 times would be best, but I don’t think I’ve ever seen a name I recognize do that. Anyway, the point is, the pros aren’t theorizing about this, they’re doing math and figuring it out. Their ability to do that (never mind the conceptual awareness TO do that) is what separates them from us.

TL;DR: this should be dictated by slate-specific quantitative analysis (notwithstanding the good points about personal risk tolerance elsewhere in this thread), but in general I would vote to use ONE optimal stack, with optimal pitchers, then randomize the remaining spots, and only vary the stacks if they are close enough in value to be basically identical. (And in this context, we’re using “optimal” to mean best combination of price, upside, variance, and ownership.)

• Tammy409

My clueless newbie’s 2-cents: First played dfs late last NFL season and when building 12 LU“s, thought they were all too similar (no clue about ownership; chalk; just too alike) so began putting a single “What if this Happens?” LU to balance out the other 11. In NFL Div Rd; thought was that CÀR Dst/Cam might blow game up early and thus limit their ROI while forcing RWilson & SEA WR“s to play catchup while CAR cruised. Obviously more time to cook up notions re game flow w/ NFL; but in the DK“s (small-slate/playoffs) 188,000 entry GPP, had 2 of 12 entries place uniquely (no ties) in Top 100. One of them was that “What If” and only 3 of 9 roster slots shared same player on both… Food for thought .

• stlcardinals84

• 242

• ### 2018 DraftKings FBBWC Finalist

@zackthys said...

I guess ultimately the question is “at what number of stacked lineups for one team is your EV maximized, and whether or not this number is greater than 250 which is usually the maximum amount of lineups allowed.”

This is an absolutely fantastic point, and admittedly one that is hard to give a concrete answer to. This is actually a good spot for a mathematics/statistics person to run some detailed analysis on – unfortunately, that’s a little beyond me.

If I had to give an educated guess, I would say that 250 stacks of one team seems a bit steep. I would likely do the 125/125 approach given your example here.

However, Saahil Sud is one of the best DFS players in the world and he frequently runs 200-300 lineups with just small variations on a stack of one team. It’s more boom/bust for sure, but he has had great success with it.

• bigloser11

I’m no expert. I play very small, 25 cents and 1.00 \$. A few to 5 lineups at most and I skip a day here and there.
I know there are people who do very well at this but I have to believe the majority don’t.jm

• FourPointPlay

There are various forms that a winning player’s skill edge manifests in.. one of them is lineup creation.. if one could quantify and enter all of the possibilities derived from their data then winning it all would be as simple as as @yanks237 worked out using “number of options raised to the power of number of decisions”

• mjordantmac

• ### 2016 FanDuel NBA Playboy Mansion Finalist

Very interesting question and something I struggle with often. For NBA things are a bit different due to the predictability of the sport but for MLB this can be challenging. I agree in the sense that if you’re running out this many lineups you’re going to need a top 1-2% score to make things worth it. So essentially you’re going to have to play for the win. Many times we see guys take down top spots with stacks or pitchers no one gave a chance. By picking just a few stacks and spreading them out across 250 lineups might not be an ideal way to achieve this.

Things to think about when calculating EV in MME strategies: How often will this stack hit/my exposure to them, pitcher ownership, how many lineups with pitcher(s), risk tolerance

• bigloser11

@mjordantmac said...

if you’re running out this many lineups you’re going to need a top 1-2% score to make things worth it.

I still feel like I’m playing lotto .If an ‘‘expert’‘ was so good they would need just a few lineups to win near the top.
The ‘‘expert’‘ could have losing nights but be ahead like the pro gamblers that win 56 %. Those pros still lose 44 %
of the time but the wins cover and put them ahead.To enter hundreds or thousands of lineups costs \$\$, time, and smells
like lotto.Just my 2 cents.jm