# NASCAR FORUM

• bhdevault

• ### Blogger of the Month

Sun, March 25th
2:00 PM ET

MONSTER ENERGY NASCAR CUP SERIES AT MARTINSVILLE
Martinsville Speedway

Admin Note: The mod team is working to keep these weekly threads more on topic – Nascar strategy talk for this week’s contest. Post referring to last week’s contest will be moved to the correct thread. Any off topic posts or posts containing full lineups will be deleted..

• Cooper08

• ### Blogger of the Month

@gje627 said...

Thanks… But again, if rankings don’t reflect skill how is doing well winning a competition?

In my example above, on a per night basis, it’s almost equivalent (not mathematically, of course) to a basketball team having 150 players on the court at a time versus your one (1) player….

Again, not a perfect mathematical equivalent, but using hyperbole here to make a point….

Again, I agree but it’s the only thing out there and yes it is not a perfect system by any means.

• gje627

As an addendum, and not to raise a big stink in the NASCAR forum, I’ve read the formula used for rankings, and it doesn’t make sense to me (but I may be too stupid to understand).

Here it is….

Each individual contest result is plugged into the following formula (the bold, rather than quotes, marks the exact wording of the Rotogrinders rankings FAQ page that includes the formula and the explanation):

Points = (Prizepool^.6) / (PlaceFinished^.6)

This is an exponential distribution that awards more points at the top of the curve where place finished is high, and successively fewer points at the base of the curve where place finished is low.

Mathematically, this makes sense…. But there are underlying assumptions that are not considered, nor seemingly adjusted for. Specifically, it assumes that all contestants have the same number of entries in each contest, all contestants are playing against the same number of total entries, all contestants are equally diversifying the sports they play, it’s not in the formula, but there is no mathematically correct adjustment elsewhere for this, only an arbitrary award in bonus points at increments of 5% for the more sports a person plays regularly….

And yes, of course, the latter adjustment only applies to TPOY rankings….

I guess the point I’m making is that it really makes no sense, and rather than being helpful could actually do harm as a contestant focusing on these results may be encouraged to get reckless with their cash and enter what they can’t afford simply to up their “public ranking” for bragging-rights when the whole ranking system is fundamentally flawed….

Again, really don’t care, but just trying to figure out if this is really an appropriate way to judge the relative worth of fantasy contestants….

Maybe there’s a better way?

P.S.: Often times I write posts like this hoping to garner the attention from someone at Rotogrinders in hopes of a reply; in a similar (but different way) that I commented on a user’s soft-porn avatar in the NASCAR Las Vegas Forum or I asked whatever happened to the new Rotogrinders data guru because I was thinking he could bolster what’s included on the NASCAR page….

Sadly, though, my questions seem to have remained under the radar (:

• stevietpfl

Morning Grind co-host, Lead NASCAR Analyst

• ### 2015 FAWBC Finalist

For everyone that was wondering about fastest laps, official NASCAR report is out from Martinsville.

• anilprao88

• 26

• Ranked #20

• ### 2018 DraftKings FBWC Finalist

Thanks Stevie

• JimKronlund

@gje627 said...

As an addendum, and not to raise a big stink in the NASCAR forum, I’ve read the formula used for rankings, and it doesn’t make sense to me (but I may be too stupid to understand).

Here it is….

Each individual contest result is plugged into the following formula (the bold, rather than quotes, marks the exact wording of the Rotogrinders rankings FAQ page that includes the formula and the explanation):

Points = (Prizepool^.6) / (PlaceFinished^.6)

This is an exponential distribution that awards more points at the top of the curve where place finished is high, and successively fewer points at the base of the curve where place finished is low.

Mathematically, this makes sense…. But there are underlying assumptions that are not considered, nor seemingly adjusted for. Specifically, it assumes that all contestants have the same number of entries in each contest, all contestants are playing against the same number of total entries, all contestants are equally diversifying the sports they play, it’s not in the formula, but there is no mathematically correct adjustment elsewhere for this, only an arbitrary award in bonus points at increments of 5% for the more sports a person plays regularly….

And yes, of course, the latter adjustment only applies to TPOY rankings….

I guess the point I’m making is that it really makes no sense, and rather than being helpful could actually do harm as a contestant focusing on these results may be encouraged to get reckless with their cash and enter what they can’t afford simply to up their “public ranking” for bragging-rights when the whole ranking system is fundamentally flawed….

Again, really don’t care, but just trying to figure out if this is really an appropriate way to judge the relative worth of fantasy contestants….

Maybe there’s a better way?

P.S.: Often times I write posts like this hoping to garner the attention from someone at Rotogrinders in hopes of a reply; in a similar (but different way) that I commented on a user’s soft-porn avatar in the NASCAR Las Vegas Forum or I asked whatever happened to the new Rotogrinders data guru because I was thinking he could bolster what’s included on the NASCAR page….

Sadly, though, my questions seem to have remained under the radar (:

gje627 Your observations have merit. Just some points to add; Only contests that are above \$5K are counted. Like I posted earlier, maxing 150 lu’s will skew the results. Just like the contests with entry fees that are high \$ and total entrants are low will skew the results. Total amount played will skew the results. Rankings have no correlation to profitability.
That said, there are other markers within a persons profile that will shed more light. Within a profile you can see their “Most recent activity”. At the bottom of that you can hit the “See all”, and view everything. You also can see their lowest scores and the frequencies of them getting a low score. At the bottom of that you can again hit the “See all”, and view everything.
At the very bottom of their profile it shows information on their past; “Top 20% Scores”, “Average Entrants”, “Average Buy-In”, “Average Tiered Buy-In”. That is even segregated by the last 7/30/365 days.
Using the Rankings and the rest of all the information readily available, will help give a pretty good indication for what someone is looking for. It certainly is not perfect but it never could be. It is an ongoing cost to RotoGrinders, and I think they have done an exemplary job in making this free service.
Now about your other point; “why is everyone so concerned about rankings?”. It is certainly not everyone. For those that do use and like it are not necessarily “concerned” about rankings. Their reasons for wanting to know them are completely subjective and personal. Some are simply curious, some like to use it as a tool to see how they are doing comparatively to others. For some maybe it for a sense of pride. There are many other reasons, and I do not judge anyone who wants to use it or not. Nor do I judge if anyone wishes to hide it for whatever reason that is for them.

In regards to your PS point. A moderators job is to look for the posts that are not allowed in RG TOS. A post like, “I asked whatever happened to the new Rotogrinders data guru because I was thinking he could bolster what’s included on the NASCAR page….”, has a real likelihood of it being as you said; “remained under the radar”. That is because it is a post that is within 100’s of other posts a day we oversee and we can quickly tell it does not violate any RG’s TOS parameters.
It is not a moderators job to facilitate answers for question like these. These types of questions or suggestions should be made on the “Contact-us” page. https://rotogrinders.com/contact-us. If it is about a moderating issue it should be emailed to moderation@rotogrinders.com. Or as said to you before. Please give me a PM and I will do what I can.

In either case gje627, if I do spot a question to RotoGrinders like your question above. I will always try to do something about it or I will forward it to the person that may be able to help with that. But with the total amount of forums and posts that I (and other moderators in other forums) must oversee. I do not always pick up on posts that are inauspicious and can miss the suggestion or question. I will promise you that in the future I will try to do a better job in picking posts like that out because in the end all we both want is what is best for everyone.

Lastly, about the questionable Avatar. I know now the one that you are talking about. Although I do not consider it to be in the “soft-porn” category. I do feel it is the “poor taste” category, and I will confer with Bill and see if it should be pulled. I am guessing that we probably will.
We try to not over moderate and we are working harder to eradicate posts that are disrespectful or degrading other members.

I honestly value all your input, and I value your thoughts above. It opens up dialogue and allows for answers and discussions about topics that I am sure many others have questions about as well.

Cheers,
JimK

Edit: We agree that Avatar does not belong. Thanks.

• Riley

RG Co-Founder

@gje627 said...

But there are underlying assumptions that are not considered, nor seemingly adjusted for. Specifically, it assumes that all contestants have the same number of entries in each contest, all contestants are playing against the same number of total entries, all contestants are equally diversifying the sports they play, it’s not in the formula…

I just wanted to hop in and address at least two of the above points which are misconceived (a. that the number of contest entries is not accounted for, and b. that differences in diversity per player are not accounted for).

Prize pool is one of the two variables that goes directly into the rankings formula, and prize pool is often itself calculated by BUY-IN * ENTRANTS. Perhaps if the formula on the FAQ were written as

Points = ((BuyIn * Entrants))^.6) / (PlaceFinished^.6)

it would be more obvious that the number of people in any particular contest IS taken into account in the current rankings formula.

You are correct that we intentionally do not penalize people for entering a contest multiple times. This is a choice we have made. We want to celebrate the best players without discouraging play. We do not want a rankings system where players have a chance of being ranked higher by stopping play or by entering fewer lineups. It is worth noting, however, that only your best finish in any contest where you have multi-entries counts towards the rankings.

To address the diversity question, we did have a diversity formula that based your percent bonus on the volume of each sport you played. We found, however, that most members were put off by the complication of that system, and we invoked the simpler 5% per additional sport bonus for that reason. Note that the 5% increase is only applied to the points awarded to the sport in question. For example, if you play NFL, MLB, and NASCAR and that is the order of rankings points you’ve earned for each sport respectively, you get a 5% increase on your MLB points and a 10% increase to your NASCAR points (so in effect, the fact that people play a different volume for their secondary, tertiary, etc. sports IS taken account into the current rankings formula).

• gje627

@Riley said...

I just wanted to hop in and address at least two of the above points which are misconceived (a. that the number of contest entries is not accounted for, and b. that differences in diversity per player are not accounted for).

Prize pool is one of the two variables that goes directly into the rankings formula, and prize pool is often itself calculated by BUY-IN * ENTRANTS. Perhaps if the formula on the FAQ were written as

Points = ((BuyIn * Entrants))^.6) / (PlaceFinished^.6)

it would be more obvious that the number of people in any particular contest IS taken into account in the current rankings formula.

You are correct that we intentionally do not penalize people for entering a contest multiple times. This is a choice we have made. We want to celebrate the best players without discouraging play. We do not want a rankings system where players have a chance of being ranked higher by stopping play or by entering fewer lineups. It is worth noting, however, that only your best finish in any contest where you have multi-entries counts towards the rankings.

To address the diversity question, we did have a diversity formula that based your percent bonus on the volume of each sport you played. We found, however, that most members were put off by the complication of that system, and we invoked the simpler 5% per additional sport bonus for that reason. Note that the 5% increase is only applied to the points awarded to the sport in question. For example, if you play NFL, MLB, and NASCAR and that is the order of rankings points you’ve earned for each sport respectively, you get a 5% increase on your MLB points and a 10% increase to your NASCAR points (so in effect, the fact that people play a different volume for their secondary, tertiary, etc. sports IS taken account into the current rankings formula).

This reply is to @JimKronlund and @Riley.

First, and foremost, I really didn’t mean to create such a kerfuffle here in the NASCAR forum…. I have a serious character flaw in that I am incapable of short posts, and generally offer too much detail in said posts, which may lead some to believe that I am angry or intentionally trying to be disagreeable… Which I assure you I am not.

Second, I very much appreciate both of your detailed replies to my questions/concerns. Since I can be slow to pick things up, I read both of your posts several times and each time I found a new nugget of valuable information and commentary. It’s obvious that you both spent significant time and energy in articulating your thoughts, and for that, I sincerely thank both of you.

Regarding the change to the presentation of the player ranking formula on the FAQ page, I agree that would be a good change as it would more clearly explain precisely what is being considered in the player ranking calculations. The only additional suggestion I would have here, is that in the first part of the potentially revised formula, an additional calculation should be made; that being eliminating the contest rake. I think this is important since not all contests have the same rake, and there is also a disparity in rake among fantasy sites; so in this regard it would be the same as removing vig from a bookmaker’s odds to determine the exact odds the bookmaker has set for a sporting event.

This of course would be more than a technical correction in how the formula is displayed on the FAQ page, in that it would change the result of the calculation. While I haven’t given it much thought, the new formula could be as follows (or something similar):

Points = (((BuyIn * Entrants) – Rake))^.6) / (PlaceFinished^.6)

Thus, by removing the rake you are applying an exactly accurate dollar amount to the prize pool for each contest.

Additionally, I do understand that the information you receive from the sites hamstrings your ability to come up with a perfect rankings system. I should have accounted for this in my remarks and made note that this makes your task of ranking fantasy players perfectly, impossible.

Regarding my comments about more and better data on the NASCAR page and the “soft porn” “poor taste” avatar, I greatly appreciate your help with those as well…. With respect to the data, my only thought was that if there was more data available on your NASCAR pages it would discourage me from wasting so much space with what I post each week in the NASCAR thread. :)

As far as the avatar, More and more women seem to be playing fantasy sports and I personally think that cleaning-up the forums a little to eliminate the occasional sexist or offensive remarks, etc., is important so everyone here feels welcome. Sorry if this is too politically correct !!

Again, thank you both for taking the time to reply to my late night rant !!

I sincerely appreciate it.

• Riley

RG Co-Founder

Yes, you are very correct that “rake” should be a part of that prize pool calculation; and perhaps also overlay if we want wanted to be completely clear. In that case, we should even more clearly write:

Points = (((BuyIn * MaxEntrants) – Rake))^.6) / (PlaceFinished^.6)

At times in the past we have done a prize pool estimated based on actual number of entrants (because of the limited information we had available), and in those cases in the past a \$10, \$100K tournament with 10,000 entries awarded more points than a \$10, \$100K tournament with 9,000 entries.

We now successfully collect the actual prize pool from all the sites that make up the rankings, so luckily we no longer have that additional confusion regarding rankings points.

• Toddalan2286

Truex vs Harvick Vs Busch @ Texas

Who will be the dominator(s) at Texas? Truex has the 1.5 mile record from 2017 and dominated Auto club, Harvick got off to a quick start but has been relatively quiet since Phoenix (Harvick did run down Truex last year for the victory last year at Texas) and Busch has stayed up front most of the season and has been on 4 perfect lineups out of 6 races this season.

• gje627

@Riley said...

Points = (((BuyIn * MaxEntrants) – Rake))^.6) / (PlaceFinished^.6)

Hmm. I need to think a little here… If the newly calculated MaxEntrants is intended to adjust for overlay, I don’t think it does because you would also need to factor in another calculation, that being ActualEntrants (i.e.the total overlay formula).

In this regard, since overlay doesn’t reduce the actual prize pool (as overlay simply means there are fewer contestants vying for the same prize pool, the left side of the formula) but instead finishing position (right side of the formula) , or more accurately relative finishing position, it would also need to placed on the right side of the formula.

Additionally, including overlay necessarily changes the formula substantially. Right now, the formula only considers absolute finishing position, regardless of the actual number of contest entries.

Thus, overlay, or lack of overlay isn’t important or even considered. By including overlay you would be implicitly saying the total number of actual contestants is now important; which means finishing position relative to the total number of contestants is now important….

I think?…. Correct me if you think I’m wrong but I don’t think using MaxEntrants without actually including ActualEntrants, regardless of which side of the formula it is on, does anything to adjust for overlay…..

EDIT: My revised formula used actual entries and your revision of my revised formula uses max entries…. Mine did not intend to factor in overlay, but your formula did.

Thus, to include overlay and using either one in isolation, nothing is done to change the actual finishing position on the right side of the formula…. In other words, according to either revised formula, finishing 10th in a contest with 60,000 actual entries is no different than finishing 10th in that contest where the max entries equals 70,000, according to the right side of the formula. Either the way, the right side of the formula still indicates a finishing position of 10. This is what I mean by absolute finishing position as opposed to relative finishing position….Hopefully, this clarifies my point and does not create further confusion.

• gje627

Okay, last thought… at least until someone smarter than me intervenes….

Instead of fumbling through calculating overlay for each contest, the same can be achieved by simply using percentiles (finishing position of actual number of entrants). Thus, no overlay calculation would be necessary but relative finishing position would now be included.

Depending on the precision of the percentiles, i.e how many decimal places they are carried it out, if any, this would also have the effect of smoothing out contest results where a number of contestants tie at certain positions. For example, a person finishing 11th in a contest with a 10-way tie for first, would be rewarded a little extra for actually having the 2nd best overall lineup in the contest….

For simplicity and accuracy, percentiles could simply be in increments of 1 or 2 units, i.e. upper:

1, 2, 3, 4, … 100; or

2, 4, 6, 8 … 100.

Certainly, though if Rotogrinders wanted to reward a player a little extra for finishing 1st alone, or in the top 10, this is also wouldn’t be that difficult under this system. For example, upper:

0.25, 0.5, 075, 1, 2, 3, 4, 6, 8 … 100.

Of course, the exact graduated percentile increments would take some evaluation before final numbers are set.

As far as using percentiles with exponents, the upper 10th percentile would simply be calculated as 10^.6., thus using percentiles as opposed to actual finishing position would still have mathematical integrity.

Thus, the new formula would be:

Points = (((BuyIn * MaxEntrants) – Rake))^.6) / (MaxPercentileFinish^.6)

Of course, for transparency, rather than Rake the full rake formula could be used on the FAQ page….

If this were adopted, all my concerns would have been addressed, except one, which of course is comparing contestants who max enter versus those who don’t.

• KindGuy

Someone with a greater knowledge than me in statistics (i.e. gje627) can probably explain this to me, but why can’t the rankings just be based on average or median percentile? Am I missing something?

Apologies if I come off as a noob……

• gje627

@elementasrat said...

Someone with a greater knowledge than me in statistics (i.e. gje627) can probably explain this to me, but why can’t the rankings just be based on average or median percentile? Am I missing something?

Apologies if I come off as a noob……

No apologies needed…. That’s pretty much what I propose. :)

However, the difficulties for Rotogrinders are threefold, (though, not to speak on their behalf):

1. Correctly applying the relative importance of contest winnings;
2. Correctly applying the relative importance of finishing position; and,
3. All this given that for #1, #2, and other data Rotogrinders has access to imperfect/incomplete information.

• KindGuy

@gje627 said...

No apologies needed…. That’s pretty much what I propose. :)

However, the difficulties for Rotogrinders are threefold, (though, not to speak on their behalf):

1. Correctly applying the relative importance of contest winnings;
2. Correctly applying the relative importance of finishing position; and,
3. All this given that for #1, #2, and other data Rotogrinders has access to imperfect/incomplete information.

I’m not really a fan of weighting results based on entry fee. Yes, I understand that in general, the skill level of contests increase as you go higher up in stakes. But, I think the current system favors high rollers too much. Just weigh them all the same. I think the difference in cashlines at a given finishing percentile is usually marginal and as a result, it isn’t worth applying an arbitrary weight.

• gje627

@elementasrat said...

But, I think the current system favors high rollers too much. Just weigh them all the same. I think the difference in cashlines at a given finishing percentile is usually marginal and as a result, it isn’t worth applying an arbitrary weight.

I agree with your first point, and as I stated in my first post that’s the primary reason I started my series of questions. Regarding your “arbitrary weight” comment, I also agree insofar as it applies to the divsersity factor added for a contestant playing multiple sports. I also stated that in my original comments.

Regarding arbitrary weights in general, proper weighting or proper re-coding, can be mathematically accurate as they are only dependent on what factors a system deems most important. In this case, those factors are dollar amounts won and finishing position. So again, “arbitrary” weights are bad, but mathematically correct weighting can be an excellent tool to accurately evaluate performance.

EDIT: I didn’t propose weighting based on entry fee or total prize pool, though I think the latter would be appropriate to weight.

• JimKronlund

@elementasrat said...

I understand that in general, the skill level of contests increase as you go higher up in stakes.

I am not sure that is true. Bank roll does not necessarily equate equally to skill level. The higher stakes contests have fewer entries and lower winning score averages.

• Riley

RG Co-Founder

>>> “why can’t the rankings just be based on average or median percentile?”

We do not want to create a situation where someone stops playing because they want to retain their high ranking status.

In the situation where someone binks 1st or 2nd place in a big tournament as one of their first scores (or even after a weighted “best of” set of scores), he would be incentivized to STOP playing DFS in the interest of maintaining his high ranking. We do not want to create this situation, and as such, our Rankings Formula is currently volume heavy (players with higher volume have a better chance of being ranked higher), and this is to encourage more play rather than discourage it.

Milwaukee couldn’t miss a 3-pointer and couldn’t get called for a foul on a block attempt in Game 3, that won’t happen again in Game 4.