NFL FORUM

  • mtdurham

    Hey guys I’m pretty familiar with DFS, game theory, and the NFL itself… but this will be my first year playing.

    Would be interested in getting a thread going where we share ideas to improve. I will be glad to contribute as well once I have a week or two under my belt and feel I have something of value to add.

    But if you had a chance to know 1-2 tidbits going into your first NFL DFS season that you did not know at the time, what would htey be???

    This can be anything from how to approach stacks, fades, roster construction, projections, ownership plays, game theory, whatever….

    Just what is something specific to NFL that you feel would have helped you improve a lot had you given it proper attention, weighting, and consideration earlier on in your DFS NFL career?

  • blake024

    Start digesting and studying each slate of games early in the week and have your lineup(s) drafted by Friday or Saturday. Never change your lineup on Sunday AM unless a player you have has been ruled out for the game. Trust your gut always.

  • escot4

    • x2

      $2M Prize Winner

    • 2016 DraftKings FFWC Champion

    Try to avoid confirmation bias. Some would say the casual fan has it easier than ever now with so many great sources of stats and information readily available. Most people won’t use them correctly though, instead most who go to a site like playerprofiler.com will go in thinking “I like player X” and then instantly lock onto all the stats that support their original notion, while easily shrugging off any data that goes against it.

  • CleverGroom

    It’s important to think about things holistically. I’m sure this applies to other sports as well, but it may be a bit extreme in the NFL.

    A recent example is how I’ve been approaching DSTs in the preseason. Yesterday, Chiefs DST was 70% owned in the KC at CHI Showdown Quarter Jukebox. I had 15%. Chiefs DST finished with 0 points.

    Why was I right, when the public was wrong? I’d argue that it’s because the public didn’t think about the problem holistically.

    The argument for playing KC was quite simple: we knew CHI was resting their starters. KC’s defense would get an entire game against CHI’s offensive backups. They’d even get some complimentary help from their offensive starters, who were expected to play into the 3rd quarter against CHI’s defensive backups.

    It’s a smash spot, right? You jam KC DST in. Especially if you’re coming off Thursday night’s 0-5 defense-fest! Folks played about 25% of each DST in that game, if I recall correctly, and they got crushed. No CLE DST, no cash.

    So with all of that fresh in the public’s minds, they rostered 75% KC DST and waited for Chase Daniel and Tyler Bray to toss up the picks. Easy game.

    The thing is that the Bears’ second-string offense isn’t at much disadvantage, if any, playing against the Chiefs’ starting defense. KC lost most of their good players during this offseason and have never been especially intimidating on the road (9.4 PPG at home last year, vs. 5.8 PPG on the road).

    The drop-off in defensive productivity was playing to see in KC’s first two preseason games. Despite playing four different QBs in ATL, each worse than the last, and virtually an entire game against HOU’s backups in preseason week 1, KC came into the week with a pathetic 7 total pressures (1 sack, 6 QB hits) and 2 turnovers generated. The only team in the league less successful at generating preseason pressures was Detroit. 30 other teams performed better than KC’s defense through the first two weeks.

    Perhaps even more importantly, KC was missing key pieces, including both starting safeties. Now if you know one thing about Eric Berry, what is it? The dude erases TEs. It’s literally his job. If you know only one thing about Matt Nagy’s offensive scheme, what do you suppose it is? Built around passes to the TE. Feels like a mismatch to me.

    For CHI’s part, they have the advantage of playing at home. Their backup QBs are both above-average understudies. Javon Wims, Daniel Brown, Ben Braunecker (note the two TEs), Tanner Gentry, and Taquan Mizzell had all excelled in previous games. The CHI offense had allowed just 13 pressures through weeks 1 and 2. This isn’t the profile of a unit you want to exploit.

    No one of these factors is enough. You have to take all of it in, and synthesize all of the information into a judgment. To do otherwise is to risk fixating on one factor—CHI is playing their backups for 4 quarters!—and missing the forest for the trees.

    When you do that, the mistakes compound themselves. Not only did the public roster a 0-point brick at a 75% clip, they also failed to invest sufficiently in CHI offensive pieces. Wims, Daniel, and Kevin White put up absolutely monstrous preseason games. You had no shot in GPPs with the KC onslaught chalk that the public landed on. You were wrong if you went into that game with player exposures that said Daniel and Bray had a 0% chance of success, and you probably lost.

    There’s some key concepts buried in here. For example, DST scoring is about pressuring the QB. But where does that come from, exactly?

    It’s quite simple, really.

    You need to know how often a team will be passing, which means you need to know the score of the game. You need to know how effective a defense is at generating pressure. You need to know how well the OL will pass protect. You need to know whether the QB has good pocket presence, whether he’ll be able to escape pressure, whether he’ll panic and arm-punt when things get dicey. You need to know whether the DC is actually going to send pressure (especially tricky in vanilla-schemed preseason games), or if he’s of the Matt Patricia, bend-don’t-break school. You need to know whether the players are willing to gamble for TOs, like Marcus Peters. You need to know whether the QB will simply ignore receivers drawing dangerous DB matchups, like Aaron Rodgers or Tom Brady will, or if he’s a young fool like Cooper Rush and will happily throw directly at Patrick Peterson. You probably need to know a great many more things that I still don’t know myself!

    Take another example: we all know opportunities are important, but how do you qualify and quantify them? A player might get a ton of snaps and still not run routes. A player might run a ton of routes but still not get targets. Two players may run the same number of routes and get the same number of targets, but if one guy’s a slot receiver or a flanker and the other guy’s an X wide receiver, their expected fantasy production could be completely different!

    So even a simple axiom like “DST scoring is about pressuring the QB,” or “opportunities are the best way to predict scoring,” is complicated. It depends on comprehensive knowledge and analysis. I think if there’s a sport-specific bugaboo for NFL, that’s it.

    What I struggle with is that I’m not sure all of this complexity is reducible to a deterministic model. I don’t use projections except on an ad hoc, case-by-case basis. My so-called “process” is built almost entirely out of my own knowledge and idiosyncratic intuitions.

    When I sit down to start each slate, my first step is to write down every player I think I’ll be interested in, from memory. There’s no spreadsheet for me to sort by projected DKFP and say, “Ah ha! It’s Kendall Wright week!”

    The good news is that nobody can replicate my process. The bad news is that it’s almost certainly a batshit-crazy house of cards.

    So yeah. Football is, like…complicated. Lean into it. Embrace ambiguity, and embrace conflicting data. Find as many data points as you possibly can, then average them instead of relying on a single factor.

  • CleverGroom

    BONUS TIP

    Look at each game on each slate and decide how it plays out.

    I’m not talking about using Vegas lines and simply deciding who wins. In fact, I’d recommend looking at those only after you’ve sat down with the slate and handicapped each game yourself. Vegas lines should be a quality assurance step in your process—not the process itself. All of your opponents already have free access to that information, just like you.

    Besides that, Vegas lines are…quirky. DAL and NE always get a few extra points, because the public likes them. CLE probably gets a bunch of juice this year too.

    Vegas is not interested in being accurate. Vegas is interested in making money. They do that not by gambling like the rest of us, but by splitting their liability equally between both outcomes and pocketing the transaction fees. Vegas is only wrong when a -7 favorite loses if they’re caught paying out more to the underdog bettors than they kept from the sheeple who bet the favorite.

    Look at each game and ask yourself, who going to succeed early? How will that success play out in terms of each offensive and defensive unit? How do matchups play out on each level (OL vs. DL, WR vs. DB, ST vs. ST, &c.)? Who wins, and how do they get there?

    Once you think you’ve hit on your likeliest game script, challenge yourself. You think JAX gets out to an early lead against NYG? Cool story, bro. How could things go differently? What happens if NYG scores on their opening drive, and now Blake Bortles has to open up the playbook? How will that change things?

    If your “red team” thought experiment leads you to a dramatically different conclusion—and I think it should, in this case—then what likelihood would you assign to each outcome? Say you started off thinking JAX wins that game 90% of the time, and now you can see a 25% path for NYG to steal it. How does that change your ownership on Odell Beckham Jr. at the lowest price he’ll have all season?

    I honestly don’t know whether this applies equally to other sports, but it’s crucial in NFL. Your rosters were DOA if you went into PHI at CLE, KC at CHI, CIN at BUF, or ARI at DAL with the wrong theory of the game. Most people did!

    There’s an enormous edge to be gained from working through each of these games, unit by unit, minute by minute, and figuring out how each moving part influences the others. Everything is interdependent. If one team has so much as a punter injury, you gotta ask yourself how that influences the game as a whole.

    Otherwise you’re just the dope who sees the World Champion Philadelphia Eagles as underdogs against Hue Jackson’s 1-31 Cleveland Browns, shrugs, and hammers the points with PHI. Bad look once Nick Foles turns back into a pumpkin.

  • CleverGroom

    General DFS tip I’m beating around the bush on: your ownership of a player is a bet on the likelihood that said player will be a top-scorer and/or top value on the slate.

    We extremize these predictions by necessity. I’ll have 0% of most QBs on larger slates, even if I don’t think Ryan Tannehill has literally a 0% chance at being the top QB play. There’s no way around that, but as you go, ask yourself whether you think your ownership of each player accurately prices in his chance at failure, or success.

    I’ve been hearing a lot of folks talking about 100% exposure to some plays in Discord this weekend, and I think that’s a head-scratcher. I told everybody who’d listen that ARI DST was the best play, followed by CIN, and yet I played both. Why not stick to my guns? I could have cashed close to 100% of my GPP entries if I’d only played my top seed.

    The answer is that I’m not always right. Neither are you. Warren Sharp has a career record of 65% accuracy in his NFL calls, and he wears that stat like a Medal of Honor. He should. It’s heckin’ incredible.

    Every play can bust. Be honest with yourself, or be ROADKILL.

  • gaelicgirl

    Damn, CG. That was fascinating, and now I understand why I fail so spectacularly at football.

    Thank you!

  • CleverGroom

    @gaelicgirl said...

    Damn, CG. That was fascinating, and now I understand why I fail so spectacularly at football.

    I’m just trying to share some of the things I’ve learned from two years of failure. If it was easy to beat the rake, we’d all…no…wait…

    I’ve done some maths, and it turns out we’ll never all beat the rake.

  • sochoice

    • 2017 DraftKings FBWC Finalist

    • 2017 FanDuel WFFC Champion

    @mtdurham said...

    Hey guys I’m pretty familiar with DFS, game theory, and the NFL itself… but this will be my first year playing.

    Would be interested in getting a thread going where we share ideas to improve. I will be glad to contribute as well once I have a week or two under my belt and feel I have something of value to add.

    But if you had a chance to know 1-2 tidbits going into your first NFL DFS season that you did not know at the time, what would htey be???

    This can be anything from how to approach stacks, fades, roster construction, projections, ownership plays, game theory, whatever….

    Just what is something specific to NFL that you feel would have helped you improve a lot had you given it proper attention, weighting, and consideration earlier on in your DFS NFL career?

    You will lose a lot. Probably most weeks. Keep this in mind, stay the course and hold on for those one or two great weeks.

  • KindGuy

    CG, those were some amazing reads. You should really start up your own blog this upcoming NFL season. Just something like a stream of consciousness before each slate or even a recap of the week would be awesome to read from you. Your writing is top notch!

  • mtdurham

    @blake024 said...

    Start digesting and studying each slate of games early in the week and have your lineup(s) drafted by Friday or Saturday. Never change your lineup on Sunday AM unless a player you have has been ruled out for the game. Trust your gut always.

    Agree with this… ive got burned in other sports by tinkering… so i def will be doing my best to avoid this

  • mtdurham

    @escot4 said...

    Try to avoid confirmation bias. Some would say the casual fan has it easier than ever now with so many great sources of stats and information readily available. Most people won’t use them correctly though, instead most who go to a site like playerprofiler.com will go in thinking “I like player X” and then instantly lock onto all the stats that support their original notion, while easily shrugging off any data that goes against it.

    Good advice… my professional/educational background is in behavioral analysis with regards to financial markets so eliminating/mitigating bias is always at the top of my list. Even still its probably the hardest thing to do, we’re all human after all.

    Well except papagates, he’s a bot lol

  • mtdurham

    @CleverGroom said...

    It’s important to think about things holistically. I’m sure this applies to other sports as well, but it may be a bit extreme in the NFL.

    A recent example is how I’ve been approaching DSTs in the preseason. Yesterday, Chiefs DST was 70% owned in the KC at CHI Showdown Quarter Jukebox. I had 15%. Chiefs DST finished with 0 points.

    Why was I right, when the public was wrong? I’d argue that it’s because the public didn’t think about the problem holistically.

    The argument for playing KC was quite simple: we knew CHI was resting their starters. KC’s defense would get an entire game against CHI’s offensive backups. They’d even get some complimentary help from their offensive starters, who were expected to play into the 3rd quarter against CHI’s defensive backups.

    It’s a smash spot, right? You jam KC DST in. Especially if you’re coming off Thursday night’s 0-5 defense-fest! Folks played about 25% of each DST in that game, if I recall correctly, and they got crushed. No CLE DST, no cash.

    So with all of that fresh in the public’s minds, they rostered 75% KC DST and waited for Chase Daniel and Tyler Bray to toss up the picks. Easy game.

    The thing is that the Bears’ second-string offense isn’t at much disadvantage, if any, playing against the Chiefs’ starting defense. KC lost most of their good players during this offseason and have never been especially intimidating on the road (9.4 PPG at home last year, vs. 5.8 PPG on the road).

    The drop-off in defensive productivity was playing to see in KC’s first two preseason games. Despite playing four different QBs in ATL, each worse than the last, and virtually an entire game against HOU’s backups in preseason week 1, KC came into the week with a pathetic 7 total pressures (1 sack, 6 QB hits) and 2 turnovers generated. The only team in the league less successful at generating preseason pressures was Detroit. 30 other teams performed better than KC’s defense through the first two weeks.

    Perhaps even more importantly, KC was missing key pieces, including both starting safeties. Now if you know one thing about Eric Berry, what is it? The dude erases TEs. It’s literally his job. If you know only one thing about Matt Nagy’s offensive scheme, what do you suppose it is? Built around passes to the TE. Feels like a mismatch to me.

    For CHI’s part, they have the advantage of playing at home. Their backup QBs are both above-average understudies. Javon Wims, Daniel Brown, Ben Braunecker (note the two TEs), Tanner Gentry, and Taquan Mizzell had all excelled in previous games. The CHI offense had allowed just 13 pressures through weeks 1 and 2. This isn’t the profile of a unit you want to exploit.

    No one of these factors is enough. You have to take all of it in, and synthesize all of the information into a judgment. To do otherwise is to risk fixating on one factor—CHI is playing their backups for 4 quarters!—and missing the forest for the trees.

    When you do that, the mistakes compound themselves. Not only did the public roster a 0-point brick at a 75% clip, they also failed to invest sufficiently in CHI offensive pieces. Wims, Daniel, and Kevin White put up absolutely monstrous preseason games. You had no shot in GPPs with the KC onslaught chalk that the public landed on. You were wrong if you went into that game with player exposures that said Daniel and Bray had a 0% chance of success, and you probably lost.

    There’s some key concepts buried in here. For example, DST scoring is about pressuring the QB. But where does that come from, exactly?

    It’s quite simple, really.

    You need to know how often a team will be passing, which means you need to know the score of the game. You need to know how effective a defense is at generating pressure. You need to know how well the OL will pass protect. You need to know whether the QB has good pocket presence, whether he’ll be able to escape pressure, whether he’ll panic and arm-punt when things get dicey. You need to know whether the DC is actually going to send pressure (especially tricky in vanilla-schemed preseason games), or if he’s of the Matt Patricia, bend-don’t-break school. You need to know whether the players are willing to gamble for TOs, like Marcus Peters. You need to know whether the QB will simply ignore receivers drawing dangerous DB matchups, like Aaron Rodgers or Tom Brady will, or if he’s a young fool like Cooper Rush and will happily throw directly at Patrick Peterson. You probably need to know a great many more things that I still don’t know myself!

    Take another example: we all know opportunities are important, but how do you qualify and quantify them? A player might get a ton of snaps and still not run routes. A player might run a ton of routes but still not get targets. Two players may run the same number of routes and get the same number of targets, but if one guy’s a slot receiver or a flanker and the other guy’s an X wide receiver, their expected fantasy production could be completely different!

    So even a simple axiom like “DST scoring is about pressuring the QB,” or “opportunities are the best way to predict scoring,” is complicated. It depends on comprehensive knowledge and analysis. I think if there’s a sport-specific bugaboo for NFL, that’s it.

    What I struggle with is that I’m not sure all of this complexity is reducible to a deterministic model. I don’t use projections except on an ad hoc, case-by-case basis. My so-called “process” is built almost entirely out of my own knowledge and idiosyncratic intuitions.

    When I sit down to start each slate, my first step is to write down every player I think I’ll be interested in, from memory. There’s no spreadsheet for me to sort by projected DKFP and say, “Ah ha! It’s Kendall Wright week!”

    The good news is that nobody can replicate my process. The bad news is that it’s almost certainly a batshit-crazy house of cards.

    So yeah. Football is, like…complicated. Lean into it. Embrace ambiguity, and embrace conflicting data. Find as many data points as you possibly can, then average them instead of relying on a single factor.

    Great post. Being new to football i dont really know all the “axioms”? What are a few others so i can at least start in the right direction before i start digging in further

  • mtdurham

    @CleverGroom said...

    BONUS TIP

    Look at each game on each slate and decide how it plays out.

    I’m not talking about using Vegas lines and simply deciding who wins. In fact, I’d recommend looking at those only after you’ve sat down with the slate and handicapped each game yourself. Vegas lines should be a quality assurance step in your process—not the process itself. All of your opponents already have free access to that information, just like you.

    Besides that, Vegas lines are…quirky. DAL and NE always get a few extra points, because the public likes them. CLE probably gets a bunch of juice this year too.

    Vegas is not interested in being accurate. Vegas is interested in making money. They do that not by gambling like the rest of us, but by splitting their liability equally between both outcomes and pocketing the transaction fees. Vegas is only wrong when a -7 favorite loses if they’re caught paying out more to the underdog bettors than they kept from the sheeple who bet the favorite.

    Look at each game and ask yourself, who going to succeed early? How will that success play out in terms of each offensive and defensive unit? How do matchups play out on each level (OL vs. DL, WR vs. DB, ST vs. ST, &c.)? Who wins, and how do they get there?

    Once you think you’ve hit on your likeliest game script, challenge yourself. You think JAX gets out to an early lead against NYG? Cool story, bro. How could things go differently? What happens if NYG scores on their opening drive, and now Blake Bortles has to open up the playbook? How will that change things?

    If your “red team” thought experiment leads you to a dramatically different conclusion—and I think it should, in this case—then what likelihood would you assign to each outcome? Say you started off thinking JAX wins that game 90% of the time, and now you can see a 25% path for NYG to steal it. How does that change your ownership on Odell Beckham Jr. at the lowest price he’ll have all season?

    I honestly don’t know whether this applies equally to other sports, but it’s crucial in NFL. Your rosters were DOA if you went into PHI at CLE, KC at CHI, CIN at BUF, or ARI at DAL with the wrong theory of the game. Most people did!

    There’s an enormous edge to be gained from working through each of these games, unit by unit, minute by minute, and figuring out how each moving part influences the others. Everything is interdependent. If one team has so much as a punter injury, you gotta ask yourself how that influences the game as a whole.

    Otherwise you’re just the dope who sees the World Champion Philadelphia Eagles as underdogs against Hue Jackson’s 1-31 Cleveland Browns, shrugs, and hammers the points with PHI. Bad look once Nick Foles turns back into a pumpkin.

    I think i understand what you’re saying…. like if you’re playing fantasy football its easy to say… team X is favored by -7.5 they’ll probably get an early lead and hand the ball off to player Y all game… but thats fantasy football where you’re just trying to play the guy with the highest likelihood of putting up numbers… thats prob appropriate “cash game strategy” on draftkings

    but for GPP that’s a good way to end up with chalky guys at higher price points.. whereas if the gamescript happens to flip on an early special teams play all the sudden the cheap RB on the other team is a nice play at a lot price point/ownership and the average QB who was only popular because he was expected to be throwing from behind all game is a terrible guy to own

  • mtdurham

    What are you guys feelings on backup RB’s ? Are they playable at all as an injury replacement / rip off a few long runs/vullture TDs… or is this too much of a hail mary?

  • Stewburtx8

    • 2012 FanDuel WFBC Finalist

    @mtdurham said...

    What are you guys feelings on backup RB’s ? Are they playable at all as an injury replacement / rip off a few long runs/vullture TDs… or is this too much of a hail mary?

    “True” backups are rarely going to be usable unless the bell cow starter gets hurt. So someone like James Conner is not going to be usable unless Le’Veon Bell gets hurt.

    But a lot of teams use a 2 or 3 back approach. In those instances there sometimes can be value in the passing down back (especially on DraftKings), especially if the game script calls for them to be playing from behind and throwing more. For example, Dion Lewis vs Derrick Henry on Tennessee. Other possibilities include Duke Johnson, Tarik Cohen, James White, Chris Thompson, Theo Riddick, etc.

  • mtdurham

    @Stewburtx8 said...

    “True” backups are rarely going to be usable unless the bell cow starter gets hurt. So someone like James Conner is not going to be usable unless Le’Veon Bell gets hurt.

    But a lot of teams use a 2 or 3 back approach. In those instances there sometimes can be value in the passing down back (especially on DraftKings), especially if the game script calls for them to be playing from behind and throwing more. For example, Dion Lewis vs Derrick Henry on Tennessee. Other possibilities include Duke Johnson, Tarik Cohen, James White, Chris Thompson, Theo Riddick, etc.

    Yes that is what i meant…. a “backup” but a guy who is already part of the gameplan… basically someone you’re relying on to break off some gamebreakers or see an increased role due to injury

    Waht is typically a GPP winning score? Obv it varies a bit …. jsut a range so i can plan accordingly would be great…
    .

  • Trappist1

    200-240 points will win most GPP’s. Lower scoring game weeks, you can see 185-200 range be enough. I recommend bookmarking and reading this weekly recaps from the last NFL season, lots of golden nuggets there on how the ‘pros’ tackle NFL GPP’s. https://www.blog.sportsdatadirect.com/author/matteohoch/

  • mtdurham

    @Trappist1 said...

    200-240 points will win most GPP’s. Lower scoring game weeks, you can see 185-200 range be enough. I recommend bookmarking and reading this weekly recaps from the last NFL season, lots of golden nuggets there on how the ‘pros’ tackle NFL GPP’s. https://www.blog.sportsdatadirect.com/author/matteohoch/

    great site, im still devouring it… if anyone has anything remotely similar to this id love to see it..

    im starting to realize im at abig disadvantage to the guys who have been playing this a few years lol

  • CleverGroom

    @mtdurham said...

    Good advice… my professional/educational background is in behavioral analysis so eliminating/mitigating bias is always at the top of my list. Even still its probably the hardest thing to do, we’re all human after all.

    That’s a great background to bring into this. It’s not a coincidence that you see a lot of successful DFS players coming from complimentary fields, like professional poker.

    Have you read Superforecasting by Philip E. Tetlock and Dan Gardner? I’d recommend that book to anybody who wants to learn more about bias and making predictions.

  • CleverGroom

    @mtdurham said...

    Great post. Being new to football i dont really know all the “axioms”? What are a few others so i can at least start in the right direction before i start digging in further

    That’s a big question, and some of the answers could be misleading.

    A good place to start would be 4for4’s DFS Strategy archives. They’ve got a ton of amazing, free content from greats like T.J. Hernandez and Chris Raybon. They’ll give you the in-depth skinny on concepts like why we generally want to pay down at QB, or how you should approach stacking.

  • Trappist1

    @CleverGroom said...

    That’s a great background to bring into this. It’s not a coincidence that you see a lot of successful DFS players coming from complimentary fields, like professional poker.

    Have you read Superforecasting by Philip E. Tetlock and Dan Gardner? I’d recommend that book to anybody who wants to learn more about bias and making predictions.

    I am a book work and information junkie by nature. So I second the Tetlock & Gardner book too. Then these come to mind as good reads too.

    James Surowiecki, The wisdom of the crowds – https://www.publishersweekly.com/978-0-385-50386-0

    Infotopia, Cass Sunstein – https://global.oup.com/academic/product/infotopia-9780195189285

    The Drunkard’s Walk, Leonard Mlodinow – https://www.penguin.co.uk/books/55764/the-drunkard-s-walk

    Thinking, Fast and Slow, Daniel Kahneman – https://www.penguinrandomhouse.ca/books/89308/thinking-fast-and-slow

  • CleverGroom

    @mtdurham said...

    I think i understand what you’re saying…. like if you’re playing fantasy football its easy to say… team X is favored by -7.5 they’ll probably get an early lead and hand the ball off to player Y all game… but thats fantasy football where you’re just trying to play the guy with the highest likelihood of putting up numbers… thats prob appropriate “cash game strategy” on draftkings

    but for GPP that’s a good way to end up with chalky guys at higher price points.. whereas if the gamescript happens to flip on an early special teams play all the sudden the cheap RB on the other team is a nice play at a lot price point/ownership and the average QB who was only popular because he was expected to be throwing from behind all game is a terrible guy to own

    I’d agree for the most part. A couple points.

    First, it’s a myth that we want QBs in negative game script, because they’ll be throwing to catch up. How did they get into negative game script? By not scoring points early, right? And what happens when you’re trailing, and your only way to score fast enough is to throw? The defense puts in their pass rushers. They pin their ears back. They substitute fresh bodies whenever you stop the clock with an incompletion. They pick you off when you have no choice but to throw a 50/50 ball. They rack up pressures and increase their chances of stripping the QB.

    Conversely, if you start hot and hold a lead throughout the game, the defense is forced to respect the run. They can’t give up efficient running plays that keep the clock running. You start seeing 8- or 9-man boxes, and that leads to cover-0 in the secondary. A QB may not pass as often under those circumstances, but when he does, he will be more successful.

    This explains a key axiom for selecting a QB: Weight Efficiency More Than Volume. Read more from Raybon in this article.

    Note that the “throwing to catch up” myth is extremely popular and easy to exploit. People won’t play Joe Flacco in week 1 partly because, “he’ll have a lead all game. Why would he throw?”

    That’s stupid. Part those fools from their money (doesn’t have to be with Flacco, necessarily, but he’s a poster child for exploitable bias).

    Second, ownership matters, but being right matters more. If everybody owns Le’Veon Bell and I don’t, that’s not good or bad in the abstract. It depends on the process that led me to the fade.

    I faded Bell week 1 of last year in Cleveland in all formats, if I recall correctly. He was terrible, apparently because he hadn’t, you know, touched a football in the better part of a year. He joined the team so late that he wasn’t even a major part of the game plan (16 opportunities, compared to 37 at CLE the year prior!). I daresay the CLE defense outperformed the public’s expectations as well.

    I played David Johnson instead, and sadly, he promptly broke his arm. Even so, Bell finished with 7.7 DKFP and DJ finished with 14.1.

    Bell was 58% owned in cash and DJ was 57%. I gained ground on the field by fading the right guy and playing a guy who did better, even if 14.1 wasn’t nearly what I’d hoped for.

    My point is that you don’t have to follow the public just to block a bad, over-owned play, even in cash. Trust and refine your process. Be explicit and honest with yourself about how likely each outcome is. If that leads you to a different understanding of a game or a player, don’t be afraid to stake out a position.

  • CleverGroom

    @mtdurham said...

    Yes that is what i meant…. a “backup” but a guy who is already part of the gameplan… basically someone you’re relying on to break off some gamebreakers or see an increased role due to injury

    Waht is typically a GPP winning score? Obv it varies a bit …. jsut a range so i can plan accordingly would be great…
    .

    The RBs question is very situational. Part of the problem is that coaches usually don’t game plan rationally around the same stats that we’re looking at. Just because the Titans play a team with LBs and safeties who can be exploited in coverage doesn’t mean that Matt LaFleur will give extra snaps and opportunities to Dion Lewis over Derrick Henry.

    Part of the game is learning the coaches’ tendencies. Kyle Shanahan is brilliant at exploiting RB mismatches in the passing game. Bill Belichick and Josh McDaniels can reinvent their entire offense on a weekly basis, but they usually have so many versatile RBs that we have to wait for injury situations if we want clarity. Doug Pederson and Frank Reich are committed to Running Back By Committee (RBBC), which usually renders all of their RBs unplayable outside of GPPs.

    One thing I’ll note is that there’s an edge in watching preseason for usage. One thing I’d hoped for, coming into this season, was more Corey Grant in a satellite/3rd-down back role. The good news/bad news is that we’re seeing more passing work to Leonard Fournette instead. I might’ve taken some shots on Grant early, but now my takeaway is that Fournette could be a better play than he showed last year.

    Ezekiel Elliott is another example. Going into the offseason program, we heard that Tavon Austin would be used as a satellite back. Instead, he’s lined up almost exclusively at WR. We haven’t seen Zeke in any preseason games, but the word from DAL is that he’s going to see more passing work this year. That’ll be incredible if it’s true.

    As for winning scores, the 4for4 articles I linked earlier will give you some food for thought. Your target scores should vary by salary and by position, and they’re not linear. Getting 3x on a $3,000 WR is very different from getting 3x on a $9,800 RB (hint: the first one sucks).

  • mtdurham

    @CleverGroom said...

    That’s a great background to bring into this. It’s not a coincidence that you see a lot of successful DFS players coming from complimentary fields, like professional poker.

    Have you read Superforecasting by Philip E. Tetlock and Dan Gardner? I’d recommend that book to anybody who wants to learn more about bias and making predictions.

    I havent. Unfortunately i have a lot of self study i barely have time to do with all the time i waste on DFS so i dont pick up a lot of outside reading lol…

  • mtdurham

    @CleverGroom said...

    That’s a big question, and some of the answers could be misleading.

    A good place to start would be 4for4’s DFS Strategy archives. They’ve got a ton of amazing, free content from greats like T.J. Hernandez and Chris Raybon. They’ll give you the in-depth skinny on concepts like why we generally want to pay down at QB, or how you should approach stacking.

    I played a lot of basebal this season… so far QB reminds me of pitching… you need to nail it to have any shot so it makes sense to find lower price guys with upside. You’ve really gotta leave some meat on the bone so takng a guy whose $7k and only gets you 20 is a killer..

    TE reminds me of catcher… a bit unpredictable with only a few good options so you gotta spread it around a little so you have some exposure a lot of different places.

    Defense is similar to TE/Catcher as well but perhaps a bit more critical to get right

  • X Unread Thread
  • X Thread with New Replies*
  • *Jumps to your first unread reply

RotoGrinders.com is the home of the daily fantasy sports community. Our content, rankings, member blogs, promotions and forum discussion all cater to the players that like to create a new fantasy team every day of the week.

If you or someone you know has a gambling problem, crisis counseling and referral services can be accessed by calling 1-800-GAMBLER (1-800-426-2537) (IL). Gambling problem? Call 1-800-GAMBLER (NJ/WV/PA/MI), 1-800-9-WITH-IT (IN), 1-800-522-4700 (CO), 1-800-BETS OFF (IA), 1-888-532-3500 (VA) or call/text TN REDLINE 1-800-889-9789 (TN).