Keep Your Lineup Building Process Simple: Winning DFS Lineups With The SaberSystem

Welcome back to the third episode of the Saber System, a framework for beating DFS. My name's Jordan Chand. I'm the head coach here at SaberSim. In the first two episodes of this series, we walked through a five step process for beating DFS in the first episode. And in the second episode, I walked through how to apply those principles in the process of entering contests for a day of baseball on DraftKings.

In this episode, we're going to be focusing on building lineups for, at the very least, the main slate for the contest that we entered here, but probably also mix in the showdown that we entered and possibly one of the turbo slates here. The five principles we outlined in the initial video were first to understand the odds and variance of DFS, Second, to optimize your portfolio, which was a big part of what we were doing when we were entering contests.

Third and fourth, which will be the two big points that we focus on in this video, are keeping your lineup building process simple and paying attention to the news. And finally, the fifth principle is to study and review the lineups. learning from the best players from the sharks. And we'll focus on that more in one of the last episodes here.

So without much further ado, let's start actually by building some lineups, jumping into the slate here. And I'll talk through how you can keep your process lean and efficient while still being able to add your value, your takes, your research, but a way that lets you play high quality lineups into the variety of different lineups, entries, and slates.

You're probably playing. If you are listening to our advice about how to build a strong. contest and lineup portfolio in the way that you actually enter your contests, you're going to be playing a lot of lineups. You're going to be playing a lot of slates. You need an efficient process to make sure you're putting good lineups and diversified lineups into all of those different entries.

So let's start here with the contest. This is a little bit earlier than I would typically sit down to build. This slate starts at 5. 07 here. It is 4. 18 my time. That is a little bit earlier than I would normally do, but we have most of the starting lineups here. And I want to give myself a little bit of extra time to talk through this stuff here.

The first thing we do want to do though, in general, as a part of paying attention to the news, is make sure that we are building with confirmed starting lineups here. So, in most of these sports, where not every game starts at the same time, like baseball here, you're not always going to have the confirmed starting lineups for every team at lock, which is why late swap is important.

But you should have the confirmed starting lineups, at the very least, for the teams that are starting at lock. And this is a very common leak that I've seen from people, especially those new to DFS or new to particular sports. is they're struggling to build good lineups. They're struggling to get good results.

Sometimes it is as simple as you didn't build with the confirmed starting lineup for each team. You were lacking the most up to date information or you were lacking putting the players that were actually playing into your lineups that day. So right off the bat, first thing as it relates to paying attention to the news and staying on top of the news is making sure that you're seeing the confirmed starting lineups here.

And you can use this as a little bit of a guide of when you might need to come back to Lateswap. So if I build right now and enter, I'm gonna have to come back at least before the 7. 40 game and make sure that my lineups are still correct for these four teams playing after that point. But I also want to pay attention here for any potential late scratches at the locks of each of these other games on the slate.

Now we'll focus more on what Lateswap actually looks like in the next episode in this series here. But the other important news element for baseball is to pay attention to the weather. We want to be aware of any potential delay risks or postponement risks on the slate. For that, I do recommend Kevin Roth's uh, weather dashboard over on Rotogrinders.

I have a full separate video on our YouTube channel that breaks down my approach to dealing with weather, but we want to prepare ourselves here for any weather concerns that we have on this slate. It looks like we're For this slate. The only potential weather concern is in Colorado with a couple scattered showers and storms.

It doesn't really look like we have any real postponement risks, so I'm not gonna make any adjustments to my build in this case. Again, if you wanna see more specifically how I handle weather, especially on slates where there are bigger concerns, you can check out the other video. But this is a part of staying on top of the news is knowing what the weather conditions are.

For the slate in question, let's go ahead and start building some lineups. The first thing we want to do is to get our contest file uploaded. So, we can go over to the entry editor. Get our contest file and get that pulled into SaberSim here. And we'll see all of our entries that we entered in the previous video.

If you're on the ultimate plan, this is a good opportunity to set up your contest sims. I've gone back and forth on how I'm actually handling my contest sims a few times in the past. We do have some work coming in the very near future to streamline the process of making sure that you're using the right contest sims for the right contest and getting the right entries in there.

But for now As a way to simplify this and keep things generally pretty easy, what I've been doing is creating one contest sim for what I think is a representative 20 max and 150 max contest. So I could probably use either the 4 seamer here or the mini max. I'm going to go with the mini max. And add that contest SIM.

And then I'm also going to add a contest SIM for a representative single entry or three max. And I'm going to use the daily dollar for this one here, rather than running individual contest SIMs for every contest I'm playing and entering all those lineups accordingly, which is basically what our new updates is going to make a lot easier.

I think it's easier to just have two contest SIMs that both represent contests of that type and run them that way. If you're not on the ultimate plan, don't worry, we'll go ahead and just keep moving forward from this point. So we go back to the build here. The main process when it comes to actually building your lineups is going to be build, review, revise.

So I'm going to let our play by play game simulations of every game on the slate handle a lot of the work for me here. We put a lot of work into our models, put a lot of work into our simulations. They are counting essentially all of the important factors here when it comes to projecting games. And every lineup we build on SaberSim is the best possible lineup for a given way that the slate may play out.

So I'll show you how. You can apply your takes here in just a second, your research and things like that. But I really strongly recommend you keep this process up front as simple as possible and leave a lot of the rules and tweaks and optimizer rules that you may have gotten used to on other tools at the door because you really don't need them here and they can potentially hurt the quality of your lineups and at worst they will make your process take longer.

A lot of this is about keeping things simple so you can get good lineups into all of these different contests that we're going to be playing. So the next step here. For me is going to be to just get some lineups built here. So. I'm going to build a full pool of 5, 000 lineups. I'm on the ultimate plan here, and I'm going to set up this build for that 150 max, 10 to 50 K the sliders.

I just like to pick something like a representative contest in between everything I'm playing here. The 20 max to 10 to 50 K would be fine. The 150 max 10 to 50 K is what I tend to prefer because that sim diversity slider is a little bit higher here. And I'm going to specify that I'm looking for 218 lineups, but I'm not going to make any other changes apart from the two that I have to make to abide by DraftKings rules here.

I'm going to let SaberSim handle setting up my build entirely and do my little value adds to clean up the build on the other side. So when we click build again, what's going to happen here is SaberSim is essentially going to simulate this slate here 5, 000 times giving us the best GPP lineup for each of those simulations.

So we'll have a 5, 000 lineup pool from which to go through and find the 218 that are going to go into our contests and we will get them entered into here. So we'll give this just a minute to finish building and then pick this up on the other side. All right. Now that our build is complete, we are ready for the next steps here, which is ultimately going to be reviewing and revising and filling and entering our contests here.

Now, this is a little bit of a different process. If you're on the SaberSim standard or the pro plans versus if you're on the SaberSim ultimate plans. So I'm going to start by explaining how I would do this on ultimate, because that's more representative of what my process looks like every night. And then once we're done with that.

I'll back up and explain the differences for how we would do this on standard or pro. That's just mostly a subtle difference in the way that your lineups are sorted. The process is generally the same, but it is a little bit different here. So on ultimate, the next step here is going to be to run our contest sim.

So in our contest sim settings, we will see here, That we have the two contest sims that we had configured before, and we can just click run contest sim. And now we're going to take each lineup in our 5, 000 lineup pool and compete it in essentially what SaberSim expects the minimax and the daily dollar to look like tonight.

So again, we're playing just those two contests, but we're saying. Lineups that are good in the minimax, or what we expect the minimax to look like, are also likely to be good in the other 20 and 150 max contests we're playing. Lineups that are good, or likely to be good, in the daily dollar are probably good in all the single entry and 3 maxes that we're playing.

So now that our contest sim is complete, I like to set up my build one step further here by filtering out the lineups that are unprofitable in either of those two contest sims. So I'm just going to go ahead here and remove those by saying only show me lineups where the ROI is greater than zero for both of these contest sims.

And this is going to trim out a chunk of our pool, leaving with us with a 2100 lineup pool from which to work with here. So, the next step, the way I handle this here, is to group up my single entries and my, uh, 150 max, and group up my single entries and 3 maxes into one category, and the 20 maxes and 150 maxes into another category.

So, for my 150 maxes and 20 maxes, I have 150, 170, 190, 210 total lineups. So I'm gonna adjust this down to 210. And the next step here is I want to make sure that these lineups are diversified. This comes back to optimizing your portfolio. Part of the reason of playing all these different lineups is I want to have a bunch of different shots on goal as possible.

I want to have a portfolio of lineups. Where I'm giving myself a chance at first place for the widest range of possible outcomes. If all of my lineups are very similar, or even some of my lineups are very similar to one another, it doesn't really have that effect. I want to be playing lineups that are, are very different from each other.

And the best way to do that is with the minuniques tool. So the next step I would do at this point is to adjust minuniques to basically as high as it can go, where I stop, where I'm not sacrificing a ton of ROI. And what I've found is the best way to do that. Oh, I'm sorry. One thing that we need to do first here is after we have done that filter is we need to sort by the risk adjusted return on investment for the contest sim.

That is an important step here. Uh, I forgot about that. So we need to sort our lineups by the best risk adjusted return on investment here for this contest. From there I recommend using minuniques and setting it to one less than the maximum number you can while you're still getting the number of lineups you need back.

So I need 210 lineups to fill my 150 maxes and my 20 maxes. So if we crank this up and just see what happens at 5, we'll see we're still getting 210 lineups back where every lineup is 5 players different from each other. If we go up to 6, We get only 155. So five is our maximum number we could set min uniques at, but I don't want to diversify at all costs.

I want to balance upside and diversification. So I would likely settle in here at four min uniques. And you see that's gonna have the immediate impact of really flattening out your exposures and making your exposures look nice here. All of these lineups are profitable in the contest sim here, and our worst lineup in our pool is only, let's see.

Is only lineup 364 out of our entire pool. So our lineup, our worst possible lineup is still grading out very well, but all of our lineups are four players different from each other, making sure that we're playing a portfolio here where the lineups are very different and we're covering a wide range of outcomes.

So at this point, this is where you can start to add in your own takes. If you had no opinions about what makes a good baseball lineup or what makes a good baseball lineup on this particular slate, you could probably just call it right here and enter these lineups and be in a good spot. But this is really the, the steps, this is the base process that you should be following, and at this point applying your particular takes.

So I'll show you a couple things that I tend to do some review on here and look at a bit and then talk about what you can do if you have more player takes or things like that. I always like to take a quick look over at my stacks and clean up my stack types a lot, a little bit here. You'll see you're getting generally the types of stacks that you would expect.

That is the correlation slider and the sims at work. But I often like to do a little bit of cleanup here. So I will just go through and actually handpick the handful of stacks that I'm actually generally willing to play, which looks more often than not something like this. So I'm just gonna do a little bit of cleanup to remove some of these smaller stacks from my pool.

They can certainly win, but I prefer to play more correlated lineups. The other thing I'll do is just do a quick glance through and take a look at what my exposure looks like to different teams on the slate here. See if I have any strong leans or opinions about the right way to approach a slate. On this one, I really don't.

I'm pretty comfortable playing with what the sims are giving me for the most part. But I'll take a quick glance through and make some adjustments. If you are more opinionated about what stacks you're getting or what teams you're getting, the way I would use the data that you're seeing here is as the baseline of how much exposure you should be getting to a certain team.

So let's say you had a strong opinion about a certain stack. Maybe you liked Baltimore a lot on this particular slate and you thought they were under owned. When you're asking yourself, what should I do to get more Baltimore? How much Baltimore should I get? You can use the curve of your lineup exposure that you've gotten to this point as a guideline.

So what Saberson wants to do is slam the Dodgers and then have the next highest exposed team be the Mets at 25%. Then the Tigers at 21 percent and so on here. So if you wanted to get more Baltimore here, ask yourself, how much do you like that team? Are they your number one favorite team on the entire slate?

Are they a team that you like in your top five? Where do they grade out actually relative to the other teams on the slate? Let's say they were a top three team overall for you. You wanted to make sure your top one of your top three stacks weren't Baltimore on this slate. Set their min exposure to get them to that point, but no further.

So maybe we set Baltimore here to about 25 percent and we click apply. And Saberson will go in and find the best 210 lineups here, where 25 percent of them are coming from Baltimore stacks. And we've had the effect of making them a top 3 stack for us, but we're doing it in the most profitable way possible, while not sacrificing our diversification or anything like that.

If we reset this here, just to go back to what the baselines look like before, you can do the same thing on the player's side. And when you're viewing it from a player's side, I would recommend using the leverage column, which is going to take your exposure to a player minus the ownership projection for a player.

Just show you how over or under the field you are on particular players. So we look at the pictures, for example, you see our highest exposure leverage is Cole Reagan's here. And then Tariq scuba at 20 and 17%. And then we get to pictures where we're even with the field on and on the flip side, we would have pictures that we are fading.

We look at the pictures overall with the highest to lowest projections here. Let's say you had a similar stand you wanted to take on a particular picture. Let's say you liked Freddie Peralta tonight as maybe a potential high strikeout upside guy, and you wanted to make him one of your higher own pitchers.

Sabers him says, the highest we want to get over the field on any given pitcher tonight is by 20%, and then we have Tariq SKU ball at 17%. So how highly exposed should we get to Freddie Peralta to take that stand effectively here? We should get about 20% over the field on him. We expect him to be owned in 17% of lineups here.

So let's set that minimum exposure to 37. Click Apply. And we'll quickly see that Freddy Peralta becomes our highest leveraged pitcher here on this slate. But we're not going overboard with that. We haven't locked him into our lineups. We're merely meeting the exact same leverage that Sabresim said was the highest leverage it makes sense to be over on a pitcher tonight.

So the leverage column and using your exposures after you have gone through the process of building, running the contest sim, and diversifying your lineups here can be a really useful tool to take those stands, but do so in an intelligent way. You can certainly add value with research based on data or reading or watching videos from other sharp players.

You can certainly do those kinds of things, but you want to make sure it fits into the context. overall approach we're taking to this portfolio here and is not completely throwing away what the sims think about the slate. What we've done in this case, for example, is we've taken the most profitable 37 percent Freddy Peralta lineups from our pool and put them into our entries file if we were to enter them here.

So I've gone ahead and reset that again now because I'm not actually going to take that stand on this slate here and I'm going to just get things back to how they a second ago. So this is exactly what I would do here if I was just building it into these right now, uh, and I essentially am for my 20 max and 150 max contest.

So let's go ahead and get these entered into those contests. Then we'll do the single entry and three max. And then I'll talk about just briefly how the process changes. If you're on the starter or the pro plan. So we'll click save to my contests here. And these contests, these lineups are only intended to go into the single entry or excuse me, the 20 max and 150 max.

What I recommend doing here is sorting, checking only those contests where these are actually supposed to go. Then sorting by entry fee. So the best lineup goes into your most expensive contests and then filling with unique rank. So when we do that, we're going to put these 210 lineups into these contests here in order of best lineup first into most expensive contest first.

So that's done. So the next step here now is to do the same thing for our single entry and three max. So for this, all we need to do is change the sorting method to change to risk adjusted ROI for the single entry and three max. Change the number of lineups to eight in this case. Now we want to adjust our diversification.

We're going to be able to get a lot more diversified in these high ROI, single entry and three max contests. And we want to take advantage of that. The next thing I'm going to do is crank up these MinUniques and see how high we can get MinUniques to go. So if we get minuniques all the way up to 8, can we do it?

We can. Can we get minuniques all the way up to 9? Can we get minuniques all the way up to 10? So 9 is our maximum value. So I'm going to use 8 minuniques here as my baseline. Now that is 8 players different per lineup here. But you see, our worst lineup from our portfolio is still only lineup 193 from our pool.

This is still a great lineup. It grades out well. If you are uncomfortable with digging that deep into your pool, you can always take this down another notch. Even seven minuniques, which is maybe what I would actually end up doing here to be quite honest, is still going to be very diversified. We're getting seven players different across all of our single entry and three maxes here.

Now you can approach this from the same standpoint here. And you'll find that in single entry and three maxes, you're able, listening to leverage column and listening to the stack exposures, to be more comfortable taking some bigger stands. And that makes sense here. We're only playing a handful of lineups.

It's okay in this case to be a little more concentrated on your top place. And I think that's a good lesson here in general. If you have strong takes about the slate, the better place to apply those takes is in your single entry and three max, then in your 20 max and 150 maxes, or put maybe in a better way here, the contest where you can be more confident in taking those stands is in the single entry and three max type contests.

Now again, I'm not going to have a lot of those takes here for my particular lineups. All I'm going to do is review my stack types again. In this case, you see we're even more concentrated on some of my favorite stack types. So I'm pretty happy with these overall. And I would probably be ready here to go ahead and enter these, this time only selecting the single entry and three max contests where these are going to end up.

So we'll go ahead and get those entered. And just to make sure I actually get these in time here, I'm going to click download selected entries at this point and get these uploaded onto DraftKings so that I at least have my lineups in place here. Now let's talk about how the process is a little bit different on starter or pro.

It's not really dramatically different at all. But we won't be running a contest sim here and calculating the direct ROIs of our lineups. So if I reset things to how they looked when the build first finished up, we'll remove the filter and we'll go back to SaberScore and set this back to 218 lineups.

So SaberScore is a custom algorithm that we developed by backtesting using contest sims to identify how much projection, upside, and ownership matters in the build. Generally quantifying the strength of a lineup and sorting by saber score here. We are still seeing positive ROI lineups show up as the best saber score lineups from our pool.

We're just not specifically identifying the mathematically best lineups, but you're still going to be playing at very good lineups. If you're playing mostly the low stakes contests, I think you're generally going to be able to get by perfectly fine with Saber score, and you probably don't need Ultimate until you've got that first 1k win, and you're really starting to take DFS a little more seriously.

It's just going to be a little more general about identifying the best lineups from your pool. Because saberscore doesn't change dynamically on a per contest basis, the only difference in the way that I would manage my lineups here is I would just group everything together and don't set a filter. So I would do all 218 lineups together using saberscore here, still go through the process of diversifying with minuniques, so we'll see what's our maximum number of minuniques here.

In this case, our maximum number of mini uniques is six, so we'll go back down to five and we have 218 lineups. We're still gonna get very nicely diversified in our pool, and we can still use things like the leverage column, our exposures to take our stands and take our decisions. You can see that Sabre score comes to a little bit of a different, uh, decision about how to approach this slate, playing a little more heavily into the Coors Field game, getting some leverage with Philadelphia and Tampa Bay instead of going to the Dodgers side of things here.

Overall, it's still a very strong approach here. And in terms of process, it, it's not dramatically different, except you're just grouping everything together here and not setting those filters to remove the, the. negative ROI lineups from your pool using the contest sim. Now that is basically it. That is literally how I'm going to enter my lineups for this slate.

That is how I build and process my lineups from most slates I play. I'll go through, trust the sims when it comes to building out my pool, run my contest sim, diversify with minuniques, make any minor adjustments that I can see generally to stacks, or if I have a stronger opinion about a particular team or player on the slate, I'll make some light adjustments to exposures, but that's pretty much it here.

I did also enter a showdown contest for this Baltimore and Toronto showdown that happens at Locke here in the previous episode. So I wanted to bounce over to that showdown contest to just walk through the process again. So you can see it for a different slate style and really just see how similar it also is here for the showdowns.

Again, when I'm playing all of these different entries, like I recommend you to do to create this strong profile here, strong portfolio, you're generally not going to Have a ton of time to spend on all of these different slates, building out a really strong opinion or taking a lot of stands or, um, doing a lot of research here.

We're going to just let the Sims handle a lot of that busy work and maybe make a small, a few small takes here or there. So let's go ahead and start the process over on this Baltimore and Toronto slate. So we'll get our entries file uploaded in. And then for this particular contest, these are both 20 maxes.

They're both very similar size and structure. So I'm just going to create one contest SIM for the solo shot here and use that as my contest SIM baseline. Now I'll go ahead here and start my build and make the two updates to projections here. And I'm going to adjust this to 38 lineups, which is actually how many lineups I need to fill this year.

Now when I'm building on my own, I do typically bump up pool size up to 5, 000. It will take a few minutes longer. So for the purposes of this video, I'm going to leave pool size at 2000, which is set that way just to make builds run a little bit faster. But if you have the time to build out the full 5, 000 lineup pool, you certainly can.

I'm just going to let Sabresim build me 2, 000 here, which will be plenty to work with for this slate. Now again, what is happening in the background here? We're simulating the game 2, 000 times in this case, building the optimal lineup for each of those 2, 000 sims. We'll be able to identify the best ones from among them to play here.

And again, the simulations are carrying a lot of the heavy lifting here. They're going to help build out our pool of profitable lineups very quickly. These strong foundation of lineups here. And then the contest Sam or saber score is going to help us identify the best possible ones to play. So we'll give this just a minute here to finish up.

And then next step here again is going to be run contest sim. If you're on ultimate, you just, or if you're not on ultimate, you just skip that step and keep going to the next step, which we'll get to in a second here. But for me, I want to run the contest sim, identify the most profitable lineups to play.

Okay. So our contest sim is complete. I'm going to, in this case, once again, sort by risk adjusted return on investment here and filter out the unprofitable lineups from our pool. Okay. The next step here, once again, is to diversify. So let's see if we can find. that maximum number of min uniques. And the maximum in this case is three min uniques, so I'm going to set that at two.

And now we have this nicely diversified portfolio of lineups here on this particular slate. So we can go through again and either we could certainly enter these lineups at this point if we're happy with the results here. But I like to do once again a little bit of review here to make sure I'm just happy with what I'm seeing.

One thing that I have noticed on SaberSim is that we do tend to favor the underdog stacks pretty heavily here on these MLB showdowns. I'm good with that because they tend to come in a little under owned because average projections don't like them very much. But one thing I don't love doing is getting really overweight here with the pitcher.

That just, it seems to be a very easy way to get duplicated quite a bit, even in the small stake stuff. So let's use this as an actual example here. What's the most, what's the highest fade that I'm taking on this slate here? It is negative 26 percent on this picture here. It's negative 10 percent on the next batter.

So Saber Simison, you can take a big fade on this slate. If you want to, we are going to full fade at Trevor Rogers here. That's a big stand. So it's okay to take a big fade. Okay. I'm gonna actually just go ahead and fade this other starting pitcher here. I just personally like to get exposure to the hitters a little bit more in these slates for MLB showdowns.

Especially in a higher total game or a decently high total game. I just don't like the approach of playing the underdog pitcher very much. I win, but I'm gonna take a big stand at least to just illustrate a point here. And we see if we look at our highest negative leverage pitchers here. That player is not even our highest negative leverage player overall on the slate.

It may feel like it was an extreme stand, but Sabersim is telling us, no, it's okay to do that. We're taking some other big stands of our own, just listening to the Sims here. So it was a good opportunity to show that here. And that also has the effect of really balancing out a lot of our other exposures here.

The one thing I'm also going to take a look at here is just quickly see if I'm playing any relief pitchers in my portfolio. They can pop up from time to time here. A lot of times just to save time, I will just X them out of my build. The probably better solution to doing that correctly would be to actually research that pitcher and see if they've pitched recently and if they have a chance to come out of the bullpen in this game.

But a lot of times I think it's easier to just uncheck them. So that is what I have tended to do. This is looking pretty good here. We've got another portfolio that I, I like the way it's come together. So we'll go ahead, save these to my contests, get the best ones into my more expensive contests here, download and get uploaded into DraftKings.

So you can see very quickly here, over 20 minutes or so while I was narrating this live, we were able to fill in almost 250 different entries for this particular slate. Very comfortable with the lineups that we've entered. They seem to be very well constructed. I did have the opportunity to add in some of my limited takes and opinions about these slates here, and we've made sure that they're very diversified.

Each one of these lineups representing a different angle on their respective slates here. Now, a couple notes here before we wrap up, I will be back to record another video for the same slate talking about how to pay attention to the news during late swap to make sure that we're getting the confirmed starting lineups here for these teams on the night games.

And we'll be able to, on the fourth episode of this series here, I guess the fifth episode of this series here, we'll actually be able to review our results in these contests and see how we played. Right. At this point, I think it's very easy to be like, does this actually work? A lot of times I get questions when I record videos like this that are, well, how did you actually do on this slate?

I'll show you a way to check that and verify that your process is actually working on the next video that I do or on the final video that I do in this series. But in the meantime, this one. Make sure to check out the next video in the series that I'll walk through my late swap for this very main slate here.

And otherwise, good luck.

‍