# NBA FORUM

• jv21

Multiplier value is an important aspect of dfs. It’s help us asses players and gives us direction when making a lineup.
For NBA, its usually suggested you want a 5-6X value. This will give a project score of around 300. While I think this idea is good, I am leaning toward moving to a new approach as I play NBA dfs more. I feel like the following will help create better lineups, help win GPPS and not make mistakes that could of been avoidable.

Idea: Assign a target score for a salary range rather than a multiplier. (Fanduel)

30 points : salary less than \$5000
40 points : salary between \$5000-\$9000
50 points : salary \$9000+

This is a simplistic approach but to win a GPP you have to score high and have all your players hit.
thoughts?

• shockermandan

• ### Moderator

I think you’re absolutely right.

The fact of the matter is 6x from a min-salary guy (21 points) isn’t going to win a GPP on an average night. At the same time, if James Harden puts up 55 tonight as a \$10,800 SG you’re probably going to want him on your team despite only being at 5x.

I play only GPPs, and consider 24.5 points for a min-priced player (7x) hitting upside-value, and have a sliding scale where I could have a \$9k player projected at as little as 5x and still be okay with that.

• btwice80

@shockermandan said...

The fact of the matter is 6x from a min-salary guy (21 points) isnâ€™t going to win a GPP on an average night. At the same time, if James Harden puts up 55 tonight as a \$10,800 SG youâ€™re probably going to want him on your team despite only being at 5x.

IMO, these these two examples are equally likely to be on a GPP-winning lineup on an average night, assuming they have similar ownership %.

• btwice80

OP, another way to do it is add a constant to the multiplier. For FD, cash game value is roughly 4x + 5. So min priced value is 19 and \$10k value is 45. For GPP value you can bump it up to 5x + 5.

• jv21

@btwice80 said...

IMO, these these two examples are equally likely to be on a GPP-winning lineup on an average night, assuming they have similar ownership %.

Not necessarily, studs are highly priced because they are consistent and can hit the high scores. A lot of things have to happen for a low price player to be a good play and hit. Hence why they are “punts” most of the time.

In regards to your second post, yeah I have seen that method too. Seems fine. Probably better than a just a number multiplier.

• btwice80

@jv21 said...

Not necessarily, studs are highly priced because they are consistent and can hit the high scores. A lot of things have to happen for a low price player to be a good play and hit. Hence why they are â€śpuntsâ€ť most of the time.

I was not saying that a min priced punt is just as likely to score 21 as Harden is to score 55. Obviously Harden’s score is more likely. I was just saying that if the punt play does score 21, and his ownership % is similar to Harden, I think they are equally likely to be on the GPP-winning team.

• jv21

@btwice80 said...

I was not saying that a min priced punt is just as likely to score 21 as Harden is to score 55. Obviously Harden’s score is more likely. I was just saying that if the punt play does score 21, and his ownership % is similar to Harden, I think they are equally likely to be on the GPP-winning team.

Yeah I understood although I don’t fully agree with it. My point was towards the idea that replacing the 55 score from Harden is much harder than the 21 from a punt. You will probably need Harden in your lineup if he scores 55 but you probably wont need the min priced punt. Eg
You can win with a lineup with Harden and others, or
Harden with the punt but
The punt with others without Harden will probably not win
So two situations to win with Harden vs one with the punt

There are other factors if you dig deeper eg slate size etc, but that’s my initial thoughts.

• shockermandan

• ### Moderator

@btwice80 said...

For GPP value you can bump it up to 5x + 5.

I think this formula is fine too. Definitely agree that a sliding scale for value makes sense.

Also worth noting that historically, value at every position is not created equally. For example, across the salary spectrum value for PGs on winning Slam lineups in Q1 2015 (I hard-keyed these in last year because I’m a nerd) was, on average, 6.24, whereas SG 5.74.

• ColonialRampage

@jv21 said...

Multiplier value is an important aspect of dfs. It’s help us asses players and gives us direction when making a lineup.
For NBA, its usually suggested you want a 5-6X value. This will give a project score of around 300. While I think this idea is good, I am leaning toward moving to a new approach as I play NBA dfs more. I feel like the following will help create better lineups, help win GPPS and not make mistakes that could of been avoidable.

Idea: Assign a target score for a salary range rather than a multiplier. (Fanduel)

30 points : salary less than \$5000
40 points : salary between \$5000-\$9000
50 points : salary \$9000+

This is a simplistic approach but to win a GPP you have to score high and have all your players hit.
thoughts?

You may not realize it, but your method is going to bias you toward guys with salaries 4500-500 and 8500-9000 and away from guys who are 5100-6000

• jv21

@ColonialRampage said...

You may not realize it, but your method is going to bias you toward guys with salaries 4500-500 and 8500-9000 and away from guys who are 5100-6000

Yeah this was something I thought about, but haven’t made a final decision on its significance. I am fine with these players in the \$5000-\$6000 range if you like them and/or that’s how the slate works out.

• dangerb407

@ColonialRampage said...

You may not realize it, but your method is going to bias you toward guys with salaries 4500-500 and 8500-9000 and away from guys who are 5100-6000

I’ve been playing around with it and I like it overall. I think there maybe need to be a couple of tiers added for that reason but I like looking at it that way.

• dustyschmidt22

Good thread, I think we get so caught up in the oversimplified 5x/6x values when it doesn’t completely account for what typically happens for a lineup that can push for a GPP win, or to cash in a cash game…whatever you’re going for.
Im primarily a cash game player so I try to exceed 5x value from my LU as a whole, but I don’t necessarily look at it like that when it comes to individual players. I look at the players in the 9k+ players to try and hit 4.5-5.5 value, the “punt” plays to hit over 6x value, and a sliding value scale in between.

• jv21

Thinking about it little more, I’d probably move just add something like a +300 buffer for the lower tier. This is mainly to account for salary fluctuation due things like high ownership. A player going from \$4800 to \$5200, I’ll still see them as a lower tier player and expect 30 points.

• BIF

I like this approach that you proposed (target score by salary range) but I was wondering if the position of the player impacts your decision making both in this approach and the 5x/6x approach ?

What positions do you typically spend the most salary on ?
What position do you typically use the Utility/Flex on ?

I realize that a small lineup of games impacts your decisions based on availability but on a night with 10+ games, what positions are you spending the salary on ? And what positions do you value-shop on ?

• JonBales

• 514

• x3

### 2015 DraftKings FFWC Finalist

This is a great topic. I agree that X value multipliers are sort of a naive way to think about value, and target scores are preferred. However, I still think created distinct salary ranges suffers from a similar problem; we shouldn’t be making any arbitrary cutoff points when assessing value.

Why should a player who costs \$5,000 be expected to score the same as one who costs \$9,000? Actually, he shouldn’t even be expected to score the same as one who costs \$5,100. To get a true sense of value, I think you need to have one complete range of salary-based expectations, presumably based on historic scoring.

• jv21

@BIF said...

I like this approach that you proposed (target score by salary range) but I was wondering if the position of the player impacts your decision making both in this approach and the 5x/6x approach ?

What positions do you typically spend the most salary on ?
What position do you typically use the Utility/Flex on ?

I realize that a small lineup of games impacts your decisions based on availability but on a night with 10+ games, what positions are you spending the salary on ? And what positions do you value-shop on ?

Yeah I do think there is some influence by positions and what’s available.
Everyone has their own preferences that shape how they assess players.
My preferences may explain why I am undervaluing plays between \$5-6000 since it includes players I tend to not play.

Shooting Guard and Center are the spots to pay down at.
I pay up for Point Guard and Small Forward most often

The intention of the idea is that it will help identify plays easier without to much stress on salary. If you look at a slate and see a \$7000 cost guy that can hit 40, I would just lock them in. Find those guys first, then you can look for other plays and use other strategies.

• shockermandan

• ### Moderator

@Jon Bales said...

Actually, he shouldnâ€™t even be expected to score the same as one who costs \$5,100. To get a true sense of value, I think you need to have one complete range of salary-based expectations, presumably based on historic scoring.

Agree with this completely. I understand the tiers to make things easy to work with, but in a perfect world you should have a different target for every salary point. That makes a sliding scale like btwice’s 5x+5 make sense.

• jv21

@Jon Bales said...

This is a great topic. I agree that X value multipliers are sort of a naive way to think about value, and target scores are preferred. However, I still think created distinct salary ranges suffers from a similar problem; we shouldn’t be making any arbitrary cutoff points when assessing value.

Why should a player who costs \$5,000 be expected to score the same as one who costs \$9,000? Actually, he shouldn’t even be expected to score the same as one who costs \$5,100. To get a true sense of value, I think you need to have one complete range of salary-based expectations, presumably based on historic scoring.

I definitely agree. The idea isn’t fully polished. A \$5000 player shouldn’t necessary be expected to produce like a \$9000 player. This comparison is actually part of the reason why I thought of my post.
First I’ll admit, there is a lot of bias in the strategy and that this bias is partially because 25 points from a \$5000 priced player is very lacking. 40 points is a reach and I’d most likely look at other options.

Second, a reason why I decided to put them into ranges is to help remove the bias from salaries. Yes, a player is priced at \$9000 because they produce high scores. But isn’t that a representation of their production and not expectation. When considering other factors, a \$7000 priced guy can have the same expectation as a \$9000 player.

I’ll look more into this and other strategies like the constant + multiplier method.

• Sjv

As a cash player on fd, here is the scale I use. I use it as a guide to be followed loosely. But, I have found success with it.

Pts/Cost
19 \$4.9
26 \$5.9
26 \$5.9
33 \$6.6
33 \$6.6
33 \$6.6
40 \$7.5
40 \$7.5
47 \$8.5

It totals out to be 297 and has forced me to take a high-floor approach. Hopefully someone finds this useful for them as well.

• thehazyone

RG Contributor

• ### Blogger of the Month

When building NBA lineups I don’t worry about whether or not a player will hit 5x or 6x – obviously we want players who will score the most points relative to their salary but what I want when I construct a roster is a lineup that will give me the most points possible based upon the projected values I am giving players. If this means that I have a \$5K guy that’s only projected to score 22.5 points in my lineup then he’s going in the lineup if it means that the other 7-8 guys create a value higher than say a \$6K guy projected to score 33 points. What I mean by this is that in order for me to use that \$6K guy who has a better xValue, I’d have to destroy roster components that would actually make the projected roster value lower than if I had used the 4.5x guy as opposed to the 5.5x guy.

This is for GPP’s only though – I don’t play NBA cash and would probably be more concerned with each player hitting a particular value if I was.

• stlcardinals84

• 246

• ### 2018 DraftKings FBBWC Finalist

@jv21 said...

Idea: Assign a target score for a salary range rather than a multiplier. (Fanduel)

30 points : salary less than \$5000
40 points : salary between \$5000-\$9000
50 points : salary \$9000+

I think this is a good starting point, but I think you need to parse out the \$5,000 – \$9,000 a little bit more. You definitely want more from an \$8,900 guy than you want from a \$5,100 guy. If you tweaked this to represent maybe 5 or 6 different ranges instead of 3, you could get a more accurate value threshold here.

• jv21

30 points : salaries between \$4500 than \$5500

35 points : salaries between \$5600 and \$6500

40 points : salaries between \$6600 – \$8000

45 points : salaries between \$8100 – \$9000

50 points : salaries between \$9100+

After thinking about it a bit, here is a revised version. It’s a combination of the methods that’s been mention. Now instead of doing the math, you can just look at the range. For the two higher tier, I would note these should be floors instead of ceiling. Eg. If I roster John Wall , I am expecting at least 50 points.

• Sjv

With these ranges I can theoretically use 9 players at the 6600 mark and reach 360. I think these goals are a bit lofty.

• ColonialRampage

@jv21 said...

30 points : salaries between \$4500 than \$5500

35 points : salaries between \$5600 and \$6500

40 points : salaries between \$6600 – \$8000

45 points : salaries between \$8100 – \$9000

50 points : salaries between \$9100+

After thinking about it a bit, here is a revised version. It’s a combination of the methods that’s been mention. Now instead of doing the math, you can just look at the range. For the two higher tier, I would note these should be floors instead of ceiling. Eg. If I roster John Wall , I am expecting at least 50 points.

Now divide all of those ranges in half one more time, and you’re pretty much have a 6x system.

• theylive

I just stick to salary divided by two for a 5x base and work from there…….the true question I have is which stud will burn you? A \$10,000 stud has to work harder for that 6x+ than the two \$5000 guys. Generally the lower priced guys cover or don’t hurt you as bad on an off night. On the other hand that 70 burger dream date can cost you 30+ points with a bad night. I always build from the bottom up and then try and fill in what studs will do the least damage if they are off……kinda play from a fear angle I guess

• sixnine0312

What do you mean a salary divided by two ?