This forum is archived and no longer active. You can visit us in our Discord Server here!

Topic: Lineup Rewind

1 Topic by CeeGee
Hey fellas,

First of all, I want to thank you for what you've created here... I think you're on the verge of something big...
I've done some number crunching with the NBA Lineup Rewind feature and I'm going to post some of the data I've found here below...

*note* the results are nearly 100% "default" based on the lineups provided by the Lineup Optimizer... The only changes I made at all to the suggested to lineups were due to injury or late scratches...

So, what I did was take a look back at the results from the past few days... I generated 100 lineups for each day, using the projection / ceiling / floor options... Essentially, I created 300 unique lineups each night... The results are below (with questions to follow)

I used the $1 Shot #2 as a reference for these results ($2 shot with 57,000 entries used on 12/3 and 12/4
*note* some lineups finished close to the bubble of the cash line for a minimum return, but, others finished much higher... For argument sake, I'll use the min cash amount for a total winnings number

12/3/14 - $2 Shot 57,000 entries
$2 Shot -- Pro LU -- -- min cash = 290.9 -- 97/100 LU Cashed --
Min $ = $5.00, so, Total = $5.00(97) = $485 - $200 buy in = $185


$2 Shot -- Ceiling LU -- min cash = 290.9 -- 53/100 LU Cashed --
Min $ = $5.00, so, Total = $5.00(53) = $265 - $200 buy in = $65


$2 Shot -- Floor LU -- min cash = 290.9 -- 41/100 LU Cashed --
Min $ = $5.00, so, Total = $5.00(41) = $205 - $200 buy in = $5

Daily totals: Cashed 191/300 (63.67%) for a Profit of $355
(again, this is assuming all lineups were MIN CASH which some were much higher)


12/4/14 - $2 Shot 57,000 entries
$2 Shot -- Pro LU -- min cash score = 255.5 -- 11/100 LU Cashed -- Min $ = $5.00, so, Total = $5.00(11) = $55 - $200 buy in = (-)$145
$2 Shot -- Ceiling LU -- min cash score = 255.5 -- 11/100 LU Cashed -- Min $ = $5.00, so, Total = $5.00(53) = $265 - $200 buy in = (-)$145
$2 Shot -- Floor LU -- min cash score = 255.5 -- 75/100 LU Cashed -- Min $ = $5.00, so, Total = $5.00(75) = $375 - $200 buy in = $175
Daily totals: Cashed 97/300 (32.33%) for a Profit of (-)$115
(again, this is assuming all lineups were MIN CASH which some were much higher)


12/6/14 - $1 Swat #2 13,000 entries
$1 Swat -- Pro LU -- min cash score = 292.4 -- 21/100 LU Cashed -- Min $ = $3.00, so, Total = $3.00(21) = $63 - $100 buy in = (-)$37
$1 Swat -- Ceiling LU -- min cash score = 292.4 -- 30/100 LU Cashed -- Min $ = $3.00, so, Total = $3.00(30) = $90 - $100 buy in = (-)$10
$1 Swat -- Floor LU -- min cash score = 292.4 -- 86/100 LU Cashed -- Min $ = $3.00, so, Total = $3.00(86) = $258 - $100 buy in = $158
Daily totals: Cashed 137/300 (45.67%) for a Profit of $111
(again, this is assuming all lineups were MIN CASH which some were much higher)


12/8/14 - $1 Swat #2 13,000 entries
$1 Swat -- Pro LU -- min cash score = 294.8 -- 37/100 LU Cashed -- Min $ = $3.00, so, Total = $3.00(37) = $111 - $100 buy in = $11
$1 Swat -- Ceiling LU -- min cash score = 294.8 -- 68/100 LU Cashed -- Min $ = $3.00, so, Total = $3.00(30) = $204 - $100 buy in = $104
$1 Swat -- Floor LU -- min cash score = 294.8 -- 0/100 LU Cashed -- Min $ = $3.00, so, Total = $3.00(0) = $0 - $100 buy in = (-)$100
Daily totals: Cashed 105/300 (35%) for a Profit of $15
(again, this is assuming all lineups were MIN CASH which some were much higher)


12/9/14 - $1 Swat #2 13,000 entries
$1 Swat -- Pro LU -- min cash score = 262.2 -- 24/100 LU Cashed -- Min $ = $3.00, so, Total = $3.00(24) = $72 - $100 buy in = (-)$28
$1 Swat -- Ceiling LU -- min cash score = 262.2 -- 0/100 LU Cashed -- Min $ = $3.00, so, Total = $3.00(0) = $0 - $100 buy in = (-)$100
$1 Swat -- Floor LU -- min cash score = 262.2 -- 87/100 LU Cashed -- Min $ = $3.00, so, Total = $3.00(87) = $261 - $100 buy in = $161
Daily totals: Cashed 111/300 (37%) for a Profit of $33
(again, this is assuming all lineups were MIN CASH which some were much higher)

Lineup Totals from Above (sorry I did not play on 12/5/14

Pro LU: 190/500 (38%) = $86.00
Ceiling LU: 162/500 (32.4%) =(-)$86.00
Floor LU: 289/500 (57.8%) = $399.00
Total = 641/1500 (42.73%) = $399 profit

So, as you can see, the Floor LU generated by this Optimizer is pretty solid... I think if it can cash in this GPP anywhere close to 50% than it should do pretty well on a daily basis in Cash Games (50/50, H2H and DU)

Now, here is my question, and this is really important...
I run the optimizer and set my lineups, before lock of course, usually around 5:30pm after the most recent Fantasy Cruncher update... Then the day plays itself out as it does every night for us grinders...
Then the following morning, I will log onto the Fantasy Cruncher and run the previous nights lineups to see how they "would have done" if I had played all 300...

My question is... The next morning, will the same 100 lineups be generated in the LINEUP REWIND that were Generated BEFORE Lock the Day Before??? Assuming I make the same exclusions as I made prior to lock, then the LINEUP REWIND should in fact give me the same set of lineups correct?

Basically, I'm trying to figure out if these numbers are anywhere close to accurate? Or, if I just wasted a few hours of my time haha

sorry about the formatting
Interesting analysis! The projected lineup will not change in Lineup Rewind. BUT what might change our floor/cieling numbers. These numbers are currently ties into the main database. If a player has a really good game his ceiling will increase. If a player has an especially bad game his floor may decrease.
Good to know.

In theory, do you think the Floor / Ceiling results would change dramatically from Pre Lock to Post Lock Lineup Rewind?

For what its worth... I tried this out today, I used the 1st 10 Lineups from each the Pro/Ceiling/Floor and currently I am cashing about 7/30 so far... I think, If I could have used all 300 unique Lineups I would be in better shape...

My issue there, of course, would be having the time to create 300 unique Lineups lol and trying to track any changes that are sure to happen last minute before lock
However, good news though, if you're telling me the Default Projected Lineup will not change...

If you figure a 38% cash rate, in a GPP that pays out the Top 18% I think it would be safe to assume that you could do fairly well, using the Projected Lineup daily in 50/50 and DU games with some good success...

From my numbers, the Floor LU cashes much better, but, since the floor LU could change in the LU Rewind I'm not sure how much it would differ from the default Pre Lock "Floor" LU... I would imagine it would be less than the 57.8% that I have figured above huh?
The question then becomes... Do you use, 5 or 10 of the Projected Lineups in your 50/50s? Do you settle on 1 of them? Do you use all 100 of them spread out in all your cash games?

Game selection has always been my down fall I feel...
Thank you for such an in depth post! I am taking a few minutes to digest it all. I also saw the formatting issues and made some changes to our css so your post will display properly.
Thanks Dave for adjusting the Formatting.

@oreo, in response to this quote earlier -

Interesting analysis! The projected lineup will not change in Lineup Rewind. BUT what might change our floor/cieling numbers. These numbers are currently ties into the main database. If a player has a really good game his ceiling will increase. If a player has an especially bad game his floor may decrease.


I would ask if there was a way, that you know of, for the Cruncher to spit out the same 100 LU for Ceiling / Floor as it did prior to lock ?

Just like you said would happen, some of the lineups from what I'm seeing are slightly different this morning.

If not, not biggie really, and I'll continue to do my studies using mainly the 1st "Projected" Lineup but it would be nice if I could do the same with the Ceiling / Floor

-Cg
Thanks Dave for adjusting the Formatting.

@oreo, in response to this quote earlier -

Interesting analysis! The projected lineup will not change in Lineup Rewind. BUT what might change our floor/cieling numbers. These numbers are currently ties into the main database. If a player has a really good game his ceiling will increase. If a player has an especially bad game his floor may decrease.


I would ask if there was a way, that you know of, for the Cruncher to spit out the same 100 LU for Ceiling / Floor as it did prior to lock ?

Just like you said would happen, some of the lineups from what I'm seeing are slightly different this morning.

If not, not biggie really, and I'll continue to do my studies using mainly the 1st "Projected" Lineup but it would be nice if I could do the same with the Ceiling / Floor

-Cg[/quote]


This issue was pointed out to me on Reddit, I actually hadn't realized it prior. The fix is simple, not 100% it will be simple to implement, but i think it shouldnt be too difficult. The lineup rewind needs to only gather the data from the DB of days prior to that day. Sounds simple enough, I will see about getting that fixed. Also, I will be working on getting the actual numbers to display on the saved lineups after the day has recorded stats. So you could in theory save all 100 lineups and then go back and review that way as well.
12-10-14 LU Rewind

If you read above, you'll get a better idea of what I'm doing... Going forward, I will only give results for the PROJECTION Lineup, and not the Ceiling / Floor Lineups... I don't have the time to go through the Ceiling / Floor Lineups to figure out what / who / which players changed from Pre Lock Set to the Post Lock "LU Rewind" set.

So, every day, I'll post my "Projected LU" Results

I will hypothetically use all 100 LU's in the $1 Swat... The Min Payout is $3.00... On most days, I will assume that every LU that "cashed" only cashed in at the Min $3.00 even though some will probably finish much higher.

Without further ado, here are the results from 12-10-14

Excluded Players
Tim Duncan
Manu Ginobili
Tony Parker
Mirza Teletovic
Donald Sloan

$1 Swat #2 -- 11,494 entries -- 2,175 cash (18.9%)
*note* i like this GPP because of the 3x min cash line compared to the 2.5x min cash of some of the other GPPs

38/100 Lineups Cashed, so, 38 x $3.00(min) = $114 for a profit of $14 (estimate)

I've got a little extra time this morning, so I'm going to list the cashing lineup scores, and where they finished in the Tourney with the REAL payout...

Scores

259.7 -- 1,665 -- $3.00
277.2 -- 474 -- $4.00
257.6 -- 1,869 -- $3.00
270.3 -- 815 -- $3.00
288.7 -- 186 -- $5.00
260.7 -- 1,572 -- $3.00
273.5 -- 631 -- $3.00
266 -- 1,120 -- $3.00
258.8 -- 1,749 -- $3.00
262.5 -- 1,416 -- $3.00
280.5 -- 368 -- $4.00
282.8 -- 300 -- $5.00
256.9 -- 1,939 -- $3.00
262.1 -- 1,444 -- $3.00
*297.8 (high) -- 67 -- $9.00
281.8 -- 328 -- $4.50
267 -- 1,017 -- $3.00
268 -- 950 -- $3.00
255.6 -- 2,079 -- $3.00
274.6 -- 564 -- $4.00
268 -- 950 -- $3.00
259.5 -- 1,692 -- $3.00
270.2 -- 825 -- $3.00
275.9 -- 533 -- $4.00
273.6 -- 626 -- $4.00
261.3 -- 1,520 -- $3.00
281.5 -- 336 -- $4.00
281.8 -- 328 -- $4.50
264.4 -- 1,253 -- $3.00
255.1 -- 2,143 -- $3.00
271 -- 769 -- $3.00
261.1 -- 1,544 -- $3.00
280.9 -- 352 -- $4.00
258.3 -- 1,807 -- $3.00
257.4 -- 1,882 -- $3.00
292.5 -- 128 -- $7.00
260.1 -- 1,636 -- $3.00
261.7 -- 1,489 -- $3.00

Total = $138, (-)$100 buy in for a profit of $38 bucks...

Not a HUGE amount of money, which is mainly because of the small $1 GPP, but, I think a 38% cash rate in a GPP paying out the Top 18% is pretty damn good... Also, if someone has the BR to use these

Here are a few other Games from yesterday and how these 100 lineups would have done...

$35K Wed NBA Layup -- Min Cash = 261.7 -- $5.00 Buy In -- 25/100 Cashed
$35K Wed NBA Rebound -- Min Cash = 264.9 -- $10.00 Buy in -- 21/100 Cashed
*note* not as good of a return here as the min cash payouts is only 2.5x of the Buy in vs 3x of the $1 Swat

$5 Double Up -- Min Cash = 254.2 -- 38/100 Cashed
$10 Double Up -- Min Cash = 255.4 -- 37/100 Cashed
$25 Double Up -- Min Cash = 255.6 -- 37/100 Cashed


Keep up the AMAZING work fellas, and Ill keep posting daily results like this!
-Cg
Thanks Dave for adjusting the Formatting.

@oreo, in response to this quote earlier -

Interesting analysis! The projected lineup will not change in Lineup Rewind. BUT what might change our floor/cieling numbers. These numbers are currently ties into the main database. If a player has a really good game his ceiling will increase. If a player has an especially bad game his floor may decrease.


I would ask if there was a way, that you know of, for the Cruncher to spit out the same 100 LU for Ceiling / Floor as it did prior to lock ?

Just like you said would happen, some of the lineups from what I'm seeing are slightly different this morning.

If not, not biggie really, and I'll continue to do my studies using mainly the 1st "Projected" Lineup but it would be nice if I could do the same with the Ceiling / Floor

-Cg[/quote]


This issue was pointed out to me on Reddit, I actually hadn't realized it prior. The fix is simple, not 100% it will be simple to implement, but i think it shouldnt be too difficult. The lineup rewind needs to only gather the data from the DB of days prior to that day. Sounds simple enough, I will see about getting that fixed. Also, I will be working on getting the actual numbers to display on the saved lineups after the day has recorded stats. So you could in theory save all 100 lineups and then go back and review that way as well.[/quote]

@Dave, that would be amazing... I think you've all proven that you really know what your doing with these projections... I am very pleased with the results that I've seen in my short time here... If I could track the Top 100 of each lineup and compare them over time, that would be simply amazing.

Thanks again for all your hard work fellas
-Cg
Super interesting, thanks for the in depth analysis CeeGee. Top 100 proj, returned a profit is the bottom line in the Swat. You wouldn't happen to have an average of the 100 would you?
Super interesting, thanks for the in depth analysis CeeGee. Top 100 proj, returned a profit is the bottom line in the Swat. You wouldn't happen to have an average of the 100 would you?


Thanks man. Right now I do not have an average for the 100 Lineups. I only wrote down the 38 lineups that cashed last night...

Since I started tracking this however, while only making minimal edits and exclusions to the Default Projections (exclusions only made for injury / scratches) the Default Pro LU has cashed 228/600 times 38% in the Swat for a total profit of $100 since 12/3/14

The "Floor" LU has cashed at a ridiculous 355/600 for 59.17% but again, I'm not sure how much the "Floor" LU Changes from the Pre Lock to the Post Lock LU the next day... Dave said he was going to work on getting the 3 Pro / Ceiling / Floor to only give the results from the Pre LU Rosters so stay tuned on that...
I have not even waded through the replies yet but I would agree if you are not locking any players the floor lineup might be the best but if you lock in 4 players like I do to each lineup I think you will get better results using the ceiling... in fact I think you will get GPP big time cash lineups.
I have not even waded through the replies yet but I would agree if you are not locking any players the floor lineup might be the best but if you lock in 4 players like I do to each lineup I think you will get better results using the ceiling... in fact I think you will get GPP big time cash lineups.


Can you share some results you've had using your "lock 4" system?
I have only been playing with the rewind using it coming up with amazing results... I work vigorous hours at a plant.. I been working on this after work.. I will have some time off over the next few days to see what I can do.
oops... sorry this is still notgonnastop.... I have created two ids to log on with so I can play with using site projections and my own projections.
All I am trying to say if you are just going to be a numbskull and not have any creativity and just play with the tool and not put in your own input ... WELL you are going to have numbskull results.
oops... sorry this is still notgonnastop.... I have created two ids to log on with so I can play with using site projections and my own projections.


You bring up a good idea, I am going to make it easier to switch quickly on the site between your projections and FC projections.
Thanks Dave for adjusting the Formatting.

@oreo, in response to this quote earlier -



I would ask if there was a way, that you know of, for the Cruncher to spit out the same 100 LU for Ceiling / Floor as it did prior to lock ?

Just like you said would happen, some of the lineups from what I'm seeing are slightly different this morning.

If not, not biggie really, and I'll continue to do my studies using mainly the 1st "Projected" Lineup but it would be nice if I could do the same with the Ceiling / Floor

-Cg



This issue was pointed out to me on Reddit, I actually hadn't realized it prior. The fix is simple, not 100% it will be simple to implement, but i think it shouldnt be too difficult. The lineup rewind needs to only gather the data from the DB of days prior to that day. Sounds simple enough, I will see about getting that fixed. Also, I will be working on getting the actual numbers to display on the saved lineups after the day has recorded stats. So you could in theory save all 100 lineups and then go back and review that way as well.[/quote]

@Dave, that would be amazing... I think you've all proven that you really know what your doing with these projections... I am very pleased with the results that I've seen in my short time here... If I could track the Top 100 of each lineup and compare them over time, that would be simply amazing.

Thanks again for all your hard work fellas
-Cg[/quote]


The first part is done. The lineup rewind now only uses the previous time periods when it grabs floors/ceilings, so the numbers you see in the lineup cruncher for a given time will be the same as in lineup rewind for that time.
Dave!

That is amazing news! Now, I can really see how these lineups "would have" finished each day.

Simply amazing, thank you.
Dave,

Is this live now, or starting today?

Meaning, Could I go back to NFL, Week 10,11,12 and so on and See how the Lineups Played out?
Dave,

Is this live now, or starting today?

Meaning, Could I go back to NFL, Week 10,11,12 and so on and See how the Lineups Played out?


This is effective for all previous days/weeks as well, if you go back to week 5 for NFL for example, the ceilings will be the highest score from prior to week 5.

**note - we were just starting prior to week 5 so weeks-1 through 4, may not work.
I have 5 teams in the top 100 in the 80k Sunday Snap 34,43,46,50,77th places. All using the cruncher with my 4 locks Le'Veon Bell, Dez,ODB, and Antonio Brown.
https://www.fanduel.com/contest/8415470/scoring/lineup/95189172/
Nice work man, however, it looks like YOU did the "hard" work on that LU haha

Regardless, I LOVE the Fantasy Cruncher! It is apart of my daily research now and I am very happy with it
Hey Dave,

for the new "Highest Average" feature, does it throw up the Lineups from Pre Lock like you set up for the Projected / Ceiling / Floor LU ?
Hi CeeGee, i really appreciate the analysis! I've just started using fantasy cruncher and i'm very impressed with it. Would it be possible to post the results of how your lines "would have" done?
Hi CeeGee, i really appreciate the analysis! I've just started using fantasy cruncher and i'm very impressed with it. Would it be possible to post the results of how your lines "would have" done?


Sure thing.

I'll try to post some daily results... Including players I excluded and players I locked to get the results.
Date: 12-16-14

Filter = Ceiling

Players Locked: John Wall, Russell Westbrook

Players Excluded: Derrick Favors, Mirza Teletovic, Dante Cunningham, Tony Allen, Justin Hamilton, Luke Babbit
*note* these player exclusions were mainly based on injury and my own intuition about the players "true" ceiling.






Min Cash in the $1 Swat was 276.8

Speaking with perfect 20/20 vision, if I could have figured out this set of parameters Pre Lock, and bought into the Swat 100 times,
I would have Cashed 87 lineups and Finished with 5 or 6 in the Top 50...
CeeGee, impressive stuff and thank you for sharing! It seems like your intuition is right on the money ;) It seems unavoidable to me that the tool has to be manually tweaked to achieve above average results - but this is to be expected.

I'm relatively new to DFS so i hope you can shed some insight on some questions i have.

I see that you locked the two PG positions - i see this quite often for top performers (300+ score). Is there a particular reason for this, when there are other positions that have similar projections/consistency ratings? Do you take vegas lines into consideration when determining your locks (over/under, ML, spread)?

Also, Is there a particular reason you decided to go with the ceiling line up instead of floor or projection? I went with projected line up for yesterday night's games and would have been in the money for all my $1 50/50 line ups if i hadn't played Paul Piecrce (lack of research on my end). But i still ended up winning 50% of the games with an average of about 270 so can't complain :)

thank you again for sharing!
12/18/14

$2 Shot - 71,839 entries
Top 13,735 cash
Min Cash = 281.4
Default Projection LU (updated at 7:37pm) with no manual changes

Nice work on the Projections today fellas!

4 LU out of the Top 10 would have cashed, and 3 of them scored 300 or more and would have cashed higher than the Min $5 cash. 3 Lineups scored over 300.

43/100 Lineups would have cashed using the Default Lineups after the most recent 7:37pm update. 29/100 Lineups finished at 300 or more

The highest score in the top 100 was 332.10 which would have placed 169th for a $30 payout

112/18/14
$2 Shot - 71,839 entries
Top 13,735 cash
Min Cash = 281.4
Default Projection LU (updated at 7:37pm) with 2 locks

Locks / Though Process

Steph Curry - Before lineups locked I thought most people would be leaning towards Westbrook because his Fantasy Points are a bit more consistent than Curry's. But, we all know Curry can go for 60 any night, and he did last night...
Curry = 61.9 and 28% owned
Westbrook = 46.6 and 44.8% owned

Jimmy Butler - Whenever Derrick Rose is out I try to play Jimmy Butler. He's usually a lock for 35 Fantasy Points, but when Rose is out he can get 45-50 because he handles the ball more. Last night he exploded for 59.5 in a surprisingly close game vs the Knicks.

So, in hindsight, its always easy to say
"lets lock in the two highest scoring players and see what the Fantasy Cruncher would have spit out"
This is why I wanted to explain these 2 locks

The Resulting lineups from those 2 locks are extremely impressive and would have made someone allot of Money had they been able to enter all 100 of them in the $2 Shot... Sadly, I did not do this as it would take a few hours to create 100 Lineups on Fanduel

96 / 100 of the Lineups would have cashed. Most of them would have cashed easily as 75 / 100 scored over 300

Below is the best LU with those settings, again, Only Curry / Butler lock with no player exclusions. So Default Projections other than those 2 locks.

This Lineup would have finished 20th for a payout of $150

Well damn! And to think I avoided yesterdays games because of a short slate (was busy too). Thanks for the analysis CeeGee, thats pretty cool to see, and its nice to have this data.
Thanks CeeGee, this is very helpful. I had James Harden and Westbrook locked yesterday and i completely missed the min as James Harden was about 25 pts short of his projection. I feel like his performance slumped due to the back to back game and with overtime too, there might be some fatigue. anyways, lesson learned. i'm trying to get a better hang of which players to lock. i think even a westbrook/davis lock would have done much better.
Thanks CeeGee, this is very helpful. I had James Harden and Westbrook locked yesterday and i completely missed the min as James Harden was about 25 pts short of his projection. I feel like his performance slumped due to the back to back game and with overtime too, there might be some fatigue. anyways, lesson learned. i'm trying to get a better hang of which players to lock. i think even a westbrook/davis lock would have done much better.



I went with the Westbrook/Davis lock and just had an OK day. I also took a risk on Kevin Durant and him rolling his ankle really limited my upside.
did anyone try to actually enter 100 lineups? oreo i see your lineups placing really high all the time. Do you spend a lot of time on projections?