NASCAR FORUM

Comments

  • joephoto

    MIchigan International Speedway.
    Anybody watching today ?

  • gje627

    Wow ….

    Loop data already in.

    Anyone here want to see the top 50 or so optimal on DK?

    Anyone here want me to post the final fantasy point numbers on DK? On FD?

    Only asking since so I wont waste my time if nobody is interested.

    Sadly, the NASCAR forum is the absolute dead zone ….

  • Toddalan2286

    Yeah we’ve missed you around here. I’m a sponge for data so anything you post I’m interested in.

  • gje627

    Okay, did the top 50 ….

    The DK optimal was a whopping 360.75 DK fantasy points, far above the winning $50K lineup; though the winning lineup was the 3rd best possible.

    The optimal is as follow:

    Ryan Preece
    Austin Dillon
    Daniel Suarez
    Kyle Larson
    Denny Hamlin
    Martin Truex Jr

    360.75
    $49,900

  • gje627

    @Toddalan2286 said...

    Yeah we’ve missed you around here. I’m a sponge for data so anything you post I’m interested in.

    Baby pictures?

    I was sooo cute !!! :)

  • rush1228

    @gje627 said...

    Baby pictures?

    I was sooo cute !!! :)

    lol

  • gje627

    FYI:

    Excel file in my Dropbox.

    First tab is top 500 optimal DK lineups including total salary and fantasy points.

    Second tab is the DK fantasy point scores, per driver, ranked and by scoring category.

    EDIT: Link to my Google Drive instead. Will add shortly.

    EDIT 2: Sorry, back to Dropbox. I screwed-up the file extension on the first try. You can download as an Excel file now.

    EDIT 3: For those not familiar with Dropbox … Click the three dots (periods) in the upper right, then click download to download the Excel file.

    Gone, but not forgotten … this is the story of Johnny Rotten

    EDIT 4: I’ll leave the link above live for 24 hours and then I will delete, so if interested please download as soon as possible.

    EDIT 5: Deleted because I have a gazillion files and I need to keep ‘em clean. Another music trivia question (YAY … see last week) …. I deleted the link and replaced it with some lyrics … A) What is the name of the song? B) Who wrote the song (2 fellas)? C) Who sang the song? D) WTF does this song have to do with Kurt Cobain (hint, too sad, but at least we still have Dave Grohl).

  • Toddalan2286

    @gje627 said...

    Wow ….

    Loop data already in.

    Anyone here want to see the top 50 or so optimal on DK?

    Anyone here want me to post the final fantasy point numbers on DK? On FD?

    Only asking since so I wont waste my time if nobody is interested.

    Sadly, the NASCAR forum is the absolute dead zone ….

    In addition to my last post. I remember you providing more predictive stats besides which driver to play etc. I don’t see many if any industry experts discussing those.

  • Toddalan2286

    @gje627 said...

    Baby pictures?

    I was sooo cute !!! :)

    Lol, sure!

  • Moszman

    Ouch, late Brad pit stop knocked me from $15.20 to $0. Missed cash by .1 points.

  • gje627

    @Toddalan2286 said...

    In addition to my last post. I remember you providing more predictive stats besides which driver to play etc. I don’t see many if any industry experts discussing those.

    Sure. I can post some things next week, or if you want something specific just PM me.

    This new package, has my math/stats/data-driven approach in disarray, more or less.

    I will say, as big a critic as I have been of the “new package” this year, particularly 550 HP at most tracks.

    Today’s race wasn’t too bad.

  • gje627

    @rush1228 said...

    lol

    Okay, maybe not….

    As a baby I was FUGGLY

    Amazing how I turned-out so hot !!!

  • gje627

    @Moszman said...

    Ouch, late Brad pit stop knocked me from $15.20 to $0. Missed cash by .1 points.

    Ouch, that sucks ….

    I got lucky today.

    Three high scoring lineups and then I cashed with the majority of my other lineups.

    Tough to profit in NASCAR GPPs.

  • depalma13

    $150 in $0 out and I had Preece and Ty Dillon, but alas Johnson killed me.

  • Toddalan2286

    @gje627 said...

    Sure. I can post some things next week, or if you want something specific just PM me.

    This new package, has my math/stats/data-driven approach in disarray, more or less.

    I will say, as big a critic as I have been of the “new package” this year, particularly 550 HP at most tracks.

    Today’s race wasn’t too bad.

    Yeah I’ve had trouble adapting some things to the new package but I have identified a few patterns that worked well for making me comfortable fading some high owned drivers today, and hopefully going forward. You turned me on to really looking at start position performance (not driver) last year and there are some definite trends this year that have started to pop.

  • gje627

    @Toddalan2286 said...

    Yeah I’ve had trouble adapting some things to the new package but I have identified a few patterns that worked well for making me comfortable fading some high owned drivers today, and hopefully going forward. You turned me on to really looking at start position performance (not driver) last year and there are some definite trends this year that have started to pop.

    Please share ?

    Not surprisingly, as I argued last year, even though starting starting position is, at most tracks, only a medium to low-high predictor of finishing position ….

    It is the single-best predictor of fantasy points.

    And, after all, that’s what we’re all mostly concerned with here.

    EDIT: This is where some simple chaos mathematics really becomes the best technique for predicting fantasy scores; more so for base fantasy points (finishing position + place differential) in contrast to dominator points (fast laps + laps led) … though it is still highly useful for the latter as well.

  • Toddalan2286

    I don’t have specific the data in front of me now, but I’ll explain my model. I created a model 2 seasons ago using various indicators/flags. (In addition to my own projections etc) 1 set of indicators consists of 4 different colors, with each driver being assigned a color through a series of formulas using loop data that Is entered in the model after each race, as well as other data like fantasy points, salary, start position and other calculated metrics I use, etc. the other set of indicators uses data like practice speed, fantasy performance and start position. I track all data from these indicators in regards to the optimal lineup each week where I’ve developed a database of how each indicator performs.

    The colors help me determine the build for the week by tracking quantities of each indicator in the optimal lineup based on how many of each in the starting grid. today’s most likely color code I predicted and was building around was 1-red 2-green 2-yellow and 1-black, below are the colors my model had for each optimal driver:

    Truex-red
    Preece-green
    Hamlin-green
    Suarez-yellow
    Larson-yellow
    Dillon-black

    Unfortunately I built mostly around bowyer (yellow) instead if Larson and Harvick/Keselowski (reds) instead of Truex today. Have to make choices only building 20 lineups.

    As for the fades, today I was able to feel confident in fading chase Elliott, Erik Jones, Ricky Stenhouse (his salary vs start made him interesting) Kurt Busch, aric Almirola, ty Dillon and yes Daniel Hemric (amongst others) I also have an indicator that pointed to Harvick being optimal today. He wasn’t optimal but obviously won the race.,Some of those may seem obvious but it’s nice to have a little confidence in the strategy. Also it requires a manual comb through tied to a process but I’m already creating my fade list on Monday and once the lineup and practice data is entered into the model everything populates. I’m a six sigma black belt by trade so I like data, and I like a process that I can continuously tweak and improve by measuring it’s performance. Plus I’m a bit of an excel nerd.

    I’m getting frustrated by researching the “experts” their logic is large in part based on having 90% plus of the field in their pool and 150 lineups and I’ve allowed their picks to distract from my process leading to bad decisions.

    That’s a peak into some of my secrets. I will tell you start positions 13-18 this year have been consistently 0-1 in terms of optimal drivers, that of course coming to an end today. I’ll blame Harvicks loose wheel…

  • Toddalan2286

    @Toddalan2286 said...

    I don’t have specific the data in front of me now, but I’ll explain my model. I created a model 2 seasons ago using various indicators/flags. (In addition to my own projections etc) 1 set of indicators consists of 4 different colors, with each driver being assigned a color through a series of formulas using loop data that Is entered in the model after each race, as well as other data like fantasy points, salary, start position and other calculated metrics I use, etc. the other set of indicators uses data like practice speed, fantasy performance and start position. I track all data from these indicators in regards to the optimal lineup each week where I’ve developed a database of how each indicator performs.

    The colors help me determine the build for the week by tracking quantities of each indicator in the optimal lineup based on how many of each in the starting grid. today’s most likely color code I predicted and was building around was 1-red 2-green 2-yellow and 1-black, below are the colors my model had for each optimal driver:

    Truex-red
    Preece-green
    Hamlin-green
    Suarez-yellow
    Larson-yellow
    Dillon-black

    Unfortunately I built mostly around bowyer (yellow) instead if Larson and Harvick/Keselowski (reds) instead of Truex today. Have to make choices only building 20 lineups.

    As for the fades, today I was able to feel confident in fading chase Elliott, Erik Jones, Ricky Stenhouse (his salary vs start made him interesting) Kurt Busch, aric Almirola, ty Dillon and yes Daniel Hemric (amongst others) I also have an indicator that pointed to Harvick being optimal today. He wasn’t optimal but obviously won the race.,Some of those may seem obvious but it’s nice to have a little confidence in the strategy. Also it requires a manual comb through tied to a process but I’m already creating my fade list on Monday and once the lineup and practice data is entered into the model everything populates. I’m a six sigma black belt by trade so I like data, and I like a process that I can continuously tweak and improve by measuring it’s performance. Plus I’m a bit of an excel nerd.

    I’m getting frustrated by researching the “experts” their logic is large in part based on having 90% plus of the field in their pool and 150 lineups and I’ve allowed their picks to distract from my process leading to bad decisions.

    That’s a peak into some of my secrets. I will tell you start positions 13-18 this year have been consistently 0-1 in terms of optimal drivers, that of course coming to an end today. I’ll blame Harvicks loose wheel…

    Oh and Kyle Busch was a full fade today too.

  • gje627

    @Toddalan2286 said...

    I’m getting frustrated by researching the “experts” their logic is large in part based on having 90% plus of the field in their pool and 150 lineups and I’ve allowed their picks to distract from my process leading to bad decisions.

    WOW !!!

    Your process is fascinating !!!

    At least to me it is ….

    A couple of thoughts:

    1. I think your process is excellent, but I would add some tweaks if I were you were (Harvick’s loose wheel, i.e. factoring variance into your projections can be super helpful). At some point, I can explain my method for doing this, in super-simple fashion.

    2. When I talk about chaos mathematics my primary (only) motivation is NOT to identify drivers, but how drivers as a group can collectively become a 6-driver lineup … since 1 or 2 drivers will not equate to a winning lineup. To me, this is what makes NASCAR fantasy so intriguing … it is truly the only ranked fantasy sport. Not to bore you, but this where chaos mathematics becomes so critical; that being when one driver moves-up a spot by definition another driver moves down a spot … and this for up to 40 drivers under current NASCAR field rules. For example, often fantasy players identify their two highest scoring drivers before a race. Assuming these drivers are expensive they “lock-em-in”. Also assuming these drivers start say … outside the top 10, they are pretty limited salary-wise with the other 4 drivers they can roster. Let’s assume that these 2 studs start 14th and 17th. What happens next? They stuff in a couple of other high-priced drivers starting 1-14 and fill the remaining 2 positions with “value drivers” and brag that they got 4 of 6 drivers correct but are totally pissed that they didn’t even cash !!! The problem with a solution? Never-ever play drivers together who have a high likelihood of cannibalizing points from one another !!! More often than not, this means never rostering drivers starting at immediately starting together (i.e., 1 and 2, 2 and 3, 19 and 20 … etc.), but this gets far more fun when we are talking about 2 to 5 and 23 to 27 … again for example.

    3. I’m stupid BUT I’m strange. I’m an attorney by profession BUT I have tons of weird, weird hobbies. Less relevant are bird-watching, 17th and 18th century Russian peasant literature, linguistics, and rare book collecting and restoration. More relevant are mathematics (particularly discrete), combinatorics, chaos mathematics, number theory, matrix theory, set theory, and programming (C++, Python, Perl, Lisp, R, Processing …. blah, blah, blah.). Point being: I make lineups without ever knowing who the drivers are that are in all my lineups !!! I only know who they are when I am completely finished (my last stage in making lineups, is running a program I wrote …. also write my own builders/optimizers …. that differentiates a group of lineups together based on a prior calculation of variance. Again, I’m too stupid to do these things on my own so I rely wholly on the greatness of prior mathematicians !!!

    4. I quoted one section of your reply … that being a somewhat critical opinion of the NASCAR “touts” …. Here I’m going to have to say I disagree as each week I try to read as much expert commentary as I can … even if the so-called expert has only earned his/her stripes by self-proclamation. Why? Even in a large heap of rubbish, there may be one or two good points that are worth considering, whether about a particular driver, race script, pit strategy, etc. I.e. my rule of thumb is NEVER, EVER discount an opinion as long as it is SUPPORTED BY REASONABLE EVIDENCE that LOGICALLY CONNECTS THE ARGUMENT TO THE CONCLUSION. Yes, they may be wrong sometimes, but remember we’re playing the numbers here (probability theory) and what happens outside the norm is beyond everyone’s ability to precisely predict and, after all, it’s up to each one of us to separate the wheat from the chaff … so bad on us for relying on advice that makes no sense.

    Damn, that was gibberish.

    The long and the short of it is, my recommendation is:

    1. Play groups of drivers (lineups) not individual drivers.

    2. As a corollary, don’t rely on any advice that simply identifies the best drivers, and then stuff as many of though drivers in lineup as possible, regardless of salary and regardless of starting position and regardless of race script.

    Okay, that makes more sense.

    (c)

  • starstx

    @gje627 said...

    Congrats !!!

    This new package is killing me, so I didn’t play that much either (only 17 lineups in the $50K).

    I was wondering what you thought about the new rules package for DFS purposes. I have not done well this year either but been slowly improving and adjusting my model. Have there always been this many failed inspections? I do not recall this happening near as often last season. There have been multiple races with a slew of drivers failing and at least a couple almost every race.

  • Toddalan2286

    @gje627 said...

    WOW !!!

    Your process is fascinating !!!

    At least to me it is ….

    A couple of thoughts:

    1. I think your process is excellent, but I would add some tweaks if I were you were (Harvick’s loose wheel, i.e. factoring variance into your projections can be super helpful). At some point, I can explain my method for doing this, in super-simple fashion.

    2. When I talk about chaos mathematics my primary (only) motivation is NOT to identify drivers, but how drivers as a group can collectively become a 6-driver lineup … since 1 or 2 drivers will not equate to a winning lineup. To me, this is what makes NASCAR fantasy so intriguing … it is truly the only ranked fantasy sport. Not to bore you, but this where chaos mathematics becomes so critical; that being when one driver moves-up a spot by definition another driver moves down a spot … and this for up to 40 drivers under current NASCAR field rules. For example, often fantasy players identify their two highest scoring drivers before a race. Assuming these drivers are expensive they “lock-em-in”. Also assuming these drivers start say … outside the top 10, they are pretty limited salary-wise with the other 4 drivers they can roster. Let’s assume that these 2 studs start 14th and 17th. What happens next? They stuff in a couple of other high-priced drivers starting 1-14 and fill the remaining 2 positions with “value drivers” and brag that they got 4 of 6 drivers correct but are totally pissed that they didn’t even cash !!! The problem with a solution? Never-ever play drivers together who have a high likelihood of cannibalizing points from one another !!! More often than not, this means never rostering drivers starting at immediately starting together (i.e., 1 and 2, 2 and 3, 19 and 20 … etc.), but this gets far more fun when we are talking about 2 to 5 and 23 to 27 … again for example.

    3. I’m stupid BUT I’m strange. I’m an attorney by profession BUT I have tons of weird, weird hobbies. Less relevant are bird-watching, 17th and 18th century Russian peasant literature, linguistics, and rare book collecting and restoration. More relevant are mathematics (particularly discrete), combinatorics, chaos mathematics, number theory, matrix theory, set theory, and programming (C++, Python, Perl, Lisp, R, Processing …. blah, blah, blah.). Point being: I make lineups without ever knowing who the drivers are that are in all my lineups !!! I only know who they are when I am completely finished (my last stage in making lineups, is running a program I wrote …. also write my own builders/optimizers …. that differentiates a group of lineups together based on a prior calculation of variance. Again, I’m too stupid to do these things on my own so I rely wholly on the greatness of prior mathematicians !!!

    4. I quoted one section of your reply … that being a somewhat critical opinion of the NASCAR “touts” …. Here I’m going to have to say I disagree as each week I try to read as much expert commentary as I can … even if the so-called expert has only earned his/her stripes by self-proclamation. Why? Even in a large heap of rubbish, there may be one or two good points that are worth considering, whether about a particular driver, race script, pit strategy, etc. I.e. my rule of thumb is NEVER, EVER discount an opinion as long as it is SUPPORTED BY REASONABLE EVIDENCE that LOGICALLY CONNECTS THE ARGUMENT TO THE CONCLUSION. Yes, they may be wrong sometimes, but remember we’re playing the numbers here (probability theory) and what happens outside the norm is beyond everyone’s ability to precisely predict and, after all, it’s up to each one of us to separate the wheat from the chaff … so bad on us for relying on advice that makes no sense.

    Damn, that was gibberish.

    The long and the short of it is, my recommendation is:

    1. Play groups of drivers (lineups) not individual drivers.

    2. As a corollary, don’t rely on any advice that simply identifies the best drivers, and then stuff as many of though drivers in lineup as possible, regardless of salary and regardless of starting position and regardless of race script.

    Okay, that makes more sense.

    In regards to the touts comment, with most (I’m generalizing) the way sone of them win or do well is because they hedge (play their low owned mass entry drivers in other contests or vice versa) and they don’t reduce their player pools to the extent that a player like myself or others in here has to. It’s an ownership game and I’m bias because I think projecting and worrying about ownership is a waste of time, has nothing to do with driver performance. I also don’t see any accountability, id love to see a regression analysis or correlation done on the tout rankings or projections vs the actuals each week. Show me you’re performing (METRICS not your winnings)and why I should pay you hundreds of dollars for your product. Just my opinion. Can you pull some valuable information, yes of course that’s why I’m constantly listening, but there are definitely improvements that can be made. We might have to agree to disagree on that one.

    In regards to the model tweak for predicting Harvick, I’m very curious what your thoughts are on that and what you mean by using variance in projections? Standard deviation of outcomes?

  • Toddalan2286

    @starstx said...

    I was wondering what you thought about the new rules package for DFS purposes. I have not done well this year either but been slowly improving and adjusting my model. Have there always been this many failed inspections? I do not recall this happening near as often last season. There have been multiple races with a slew of drivers failing and at least a couple almost every race.

    I could be wrong and don’t have data to support this but I feel like there has been less or about the same amount of failures this year compared to last. I think there have been more impound races this year so that makes things tough but I can only think of a few races where there were substantial failures. Kansas and Richmond come to mind. There was Dover where Bowman were sent to the rear (kept starting spot). Of course this weekend with Childress was odd. Again, that’s my gut feel after reading this, someone might have data to say otherwise.

    I think switching the qualifying protocol a 1/4 into the season made things difficult. I also think the 550 package is more weather sensitive due to the momentum needed and the fine line between wide open and having to lift when track is slick. Main reason Denny faded and Harvick cane in late in yesterday’s race, track cooled off and benefited the cars set up for speed vs handling.

  • gje627

    If anyone is still interested in the data I posted for yesterday’s race, I will leave the link above live for another 60 minutes so download now …. I’m not keeping a copy for myself, as I store the same information in other formats so it will not be available after 60 minutes.

    Fair warning ….

    DELETED

    ….

    Also, I owe some people replies here. I will reply but it may be a day or two … sorry ….

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

Subforum Index

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), 1-800-9-WITH-IT (IN), 1-800-522-4700 (CO) or 1-800-BETS OFF (IA). 21+. NJ/PA/WV/IN/IA/CO/IL only.