DFS

The Fantasy Labs Mailbag, Volume 1

This first edition of the Fantasy Labs mailbag will be co-founder Jonathan Bales and I answering the questions you sent in via Twitter. We’ll keep the #FLmailbag hashtag the same each week, so feel free to keep on sending in your questions and I’ll pool the best ones every week.

Okay, enough rambling. Let’s mailbag (or Twitterbag?)!

Bryan: I look at a variety of things when gauging the opposing team or matchup for a player. One is “true DvP,” a topic I’ve written about a couple times recently (here’s a link to the original and updated articles). Basically, that stat shows how many points a team allows to a certain position after adjusting for the salary of that position.

Other stats I look at is a team’s defensive rating (points allowed per 100 possessions, so it adjusts for pace) and where they allow points. For example, if a team is hemorrhaging 3-pointers to their opposition and are facing Stephen Curry and Klay Thompson, that is obviously notable. Or if a team doesn’t have a great rim protector and thus lets up a lot of close-range shots, that would bump up players who are either good in the post (think Jonas Valanciunas) or even guys who can get to the rim well (think Alec Burks).

Other than that, keep in mind the Vegas implied totals for each team – that’s not exactly an “opposing team stat,” but it does help define what matchup you’re really looking at for the day. It is also a fairly good potential indicator of pace. Finally, paceD is a stat in our models that accounts for that as well, and I think is really important when a high-paced team is playing a low-paced one.

Bales: I think cash games are suggested to new players because they’re relatively safe. Whereas GPPs are super volatile and you can go on long cold streaks, cash games tend to provide a steadier return (even if it’s negative). If you’re using cash games to learn how to approach the game, I think that’s fine.

However, the idea that you can build up a bankroll through cash games is wrong if you aren’t starting with a healthy bankroll already. Just do the math. You shouldn’t realistically play more than, say, 10% of your bankroll per night. If you have even a 15% ROI (which is unrealistic in cash games, especially at larger stakes), a bankroll of $1,000 would mean you’d return $15 per night – not exactly life-changing money.
Further, I believe the biggest long-term edge is in tournaments, which are less capable of being “solved.” So while I’m fine with new players competing in all league types to see what works for them and to just learn the basics, you aren’t going to build a massive bankroll in cash games unless you have money to start or a whole lot of time.

Bryan: Unfortunately, there is not a simple answer. The problem with accurately projecting ownership – like if you wanted to try to model this with an algorithm – is just how many factors there really are that go into ownership and how constantly they change. Think about things that could affect it – matchup, price, groupthink, slate size, positional scarcity, the price difference between the top and mid-tier guys, injuries, safety/volatility, biases, narratives, models, etc.. I could go on for a while, but the point is that the factors are numerous and thus make it hard to project ownership.

I don’t think it’s impossible though, just hard. If you want to create an algorithm that gets within two percentage points of every player, that’s probably not super realistic right now. But if you want to just approximate it, there are ways and a lot of them I listed above. Value is the biggest indicator of ownership, in my opinion – Mario Chalmers will be very high owned tonight if Mike Conley is out, because the combination of his price and projected minutes means he’s the best value on the slate, and people know that.

Other than that, matchup really matters for the elite guys – if you have four players all $9,800 or higher, their ownership will usually fall in line with their matchups. Lastly, it gets really tough in smaller slates – people are much less willing to go with a balanced lineup in small slates than in bigger ones, mostly because the room for error is smaller. If one stud goes off, you’re out of contention if you don’t have him.

Bales: It means I’m way more likely to be contrarian (and yes, fewer cash games). Ownership obviously skyrockets in short slates like Wild Card Weekend, so there’s just more value in identifying under-the-radar plays, and bypassing values to do it.

Bales: It has much more value because there’s a reduced opportunity cost associated with it. With fewer options available, the idea that you’re limiting your ceiling by stacking an offense too heavily has less merit. In a normal slate, we know there’s going to be enough big performances that you can’t just completely load up on an offense. We don’t know that in a four-game slate – we could see normal scoring from all but one team – and that increases the value of stacking an offense. It’s actually my favorite way to be “contrarian with chalk” in these situations.

Bryan: The best way is to live and breathe basketball. No, really – we have Justin Phan that works full time at this and constantly monitors news, injuries, minute loads for every player in every game, and on and on. It’s very difficult for an average player to do this on their own (mostly because of time restraints), which is a big reason why I think our models here at Fantasy Labs are so valuable. You aren’t just getting blind projections or obvious plays (Anthony Davis could do well against the 76ers!) – you’re getting minute and usage projections that are heavily researched and constantly monitored every single day.

Anyway, I’ll wrap up this chat. Thanks for participating and sending in questions – please go ahead and send some more our way for next week. Speaking of Justin, he’ll be our next mailbagger!

YOU MAY ALSO LIKE

How to Play Games (and Win at DFS, Too) DFS, The Abilene Paradox, and Taking Chances Polling the FantasyLabs Team: What is the Edge in DFS? Asset Allocation: The Secret of Bankroll Management

This first edition of the Fantasy Labs mailbag will be co-founder Jonathan Bales and I answering the questions you sent in via Twitter. We’ll keep the #FLmailbag hashtag the same each week, so feel free to keep on sending in your questions and I’ll pool the best ones every week.

Okay, enough rambling. Let’s mailbag (or Twitterbag?)!

Bryan: I look at a variety of things when gauging the opposing team or matchup for a player. One is “true DvP,” a topic I’ve written about a couple times recently (here’s a link to the original and updated articles). Basically, that stat shows how many points a team allows to a certain position after adjusting for the salary of that position.

Other stats I look at is a team’s defensive rating (points allowed per 100 possessions, so it adjusts for pace) and where they allow points. For example, if a team is hemorrhaging 3-pointers to their opposition and are facing Stephen Curry and Klay Thompson, that is obviously notable. Or if a team doesn’t have a great rim protector and thus lets up a lot of close-range shots, that would bump up players who are either good in the post (think Jonas Valanciunas) or even guys who can get to the rim well (think Alec Burks).

Other than that, keep in mind the Vegas implied totals for each team – that’s not exactly an “opposing team stat,” but it does help define what matchup you’re really looking at for the day. It is also a fairly good potential indicator of pace. Finally, paceD is a stat in our models that accounts for that as well, and I think is really important when a high-paced team is playing a low-paced one.

Bales: I think cash games are suggested to new players because they’re relatively safe. Whereas GPPs are super volatile and you can go on long cold streaks, cash games tend to provide a steadier return (even if it’s negative). If you’re using cash games to learn how to approach the game, I think that’s fine.

However, the idea that you can build up a bankroll through cash games is wrong if you aren’t starting with a healthy bankroll already. Just do the math. You shouldn’t realistically play more than, say, 10% of your bankroll per night. If you have even a 15% ROI (which is unrealistic in cash games, especially at larger stakes), a bankroll of $1,000 would mean you’d return $15 per night – not exactly life-changing money.
Further, I believe the biggest long-term edge is in tournaments, which are less capable of being “solved.” So while I’m fine with new players competing in all league types to see what works for them and to just learn the basics, you aren’t going to build a massive bankroll in cash games unless you have money to start or a whole lot of time.

Bryan: Unfortunately, there is not a simple answer. The problem with accurately projecting ownership – like if you wanted to try to model this with an algorithm – is just how many factors there really are that go into ownership and how constantly they change. Think about things that could affect it – matchup, price, groupthink, slate size, positional scarcity, the price difference between the top and mid-tier guys, injuries, safety/volatility, biases, narratives, models, etc.. I could go on for a while, but the point is that the factors are numerous and thus make it hard to project ownership.

I don’t think it’s impossible though, just hard. If you want to create an algorithm that gets within two percentage points of every player, that’s probably not super realistic right now. But if you want to just approximate it, there are ways and a lot of them I listed above. Value is the biggest indicator of ownership, in my opinion – Mario Chalmers will be very high owned tonight if Mike Conley is out, because the combination of his price and projected minutes means he’s the best value on the slate, and people know that.

Other than that, matchup really matters for the elite guys – if you have four players all $9,800 or higher, their ownership will usually fall in line with their matchups. Lastly, it gets really tough in smaller slates – people are much less willing to go with a balanced lineup in small slates than in bigger ones, mostly because the room for error is smaller. If one stud goes off, you’re out of contention if you don’t have him.

Bales: It means I’m way more likely to be contrarian (and yes, fewer cash games). Ownership obviously skyrockets in short slates like Wild Card Weekend, so there’s just more value in identifying under-the-radar plays, and bypassing values to do it.

Bales: It has much more value because there’s a reduced opportunity cost associated with it. With fewer options available, the idea that you’re limiting your ceiling by stacking an offense too heavily has less merit. In a normal slate, we know there’s going to be enough big performances that you can’t just completely load up on an offense. We don’t know that in a four-game slate – we could see normal scoring from all but one team – and that increases the value of stacking an offense. It’s actually my favorite way to be “contrarian with chalk” in these situations.

Bryan: The best way is to live and breathe basketball. No, really – we have Justin Phan that works full time at this and constantly monitors news, injuries, minute loads for every player in every game, and on and on. It’s very difficult for an average player to do this on their own (mostly because of time restraints), which is a big reason why I think our models here at Fantasy Labs are so valuable. You aren’t just getting blind projections or obvious plays (Anthony Davis could do well against the 76ers!) – you’re getting minute and usage projections that are heavily researched and constantly monitored every single day.

Anyway, I’ll wrap up this chat. Thanks for participating and sending in questions – please go ahead and send some more our way for next week. Speaking of Justin, he’ll be our next mailbagger!

YOU MAY ALSO LIKE

How to Play Games (and Win at DFS, Too) DFS, The Abilene Paradox, and Taking Chances Polling the FantasyLabs Team: What is the Edge in DFS? Asset Allocation: The Secret of Bankroll Management