In last week’s article we covered the important of variance on choosing your players to maximize upside, an integral factor in creating a roster that maximizes ceiling potential. Beyond that, it is essential to understand the GPP marketplace and how GPP projections should differ from the ones you use for cash.
The GPP Marketplace
This is a topic that doesn’t get addressed specifically too often, and it’s a simple concept. The marketplace is similar to Wall Street, where a bunch of players are making bets on various players, and as a result certain stocks are in high demand while others are ice cold. This marketplace differs daily, as players can go from zero demand to perceived must haves based on injury news, and there are certainly a number of bettors making their decisions based on matchups and a player’s recent performance.
Our job in this marketplace is to identify the players who have actual value but whose stocks will get discarded based on the market generally choosing to steer clear of them because of a couple bad games, or the fact that there are better options for their purchasing dollar. In order to find these plays, I try to project ceiling performances and then compare different options relative to how I predict each player will be owned to some degree. It is best to show using an example:
Player 1 ($8,000 salary): Projected at 40 points, 37 minutes, expected ownership 25%
Player 2 ($8,000 salary): Projected at 38 points, 33 minutes, expected ownership 10%
The reason player 1 will be more popular is he will return a slightly better return per dollar based on the projections people are using. However, I would generally project these two players differently, as player 1 almost never exceeds 37 minutes, and player 2 will sometimes play 36 minutes, but sometimes will get 30, making him more volatile than the former player.
Therefore, instead of projecting player 2 at 33 minutes, I will modify that number to 36, thereby giving him a projection of (38/33)*36 = 41.45 points, thus making him an actual better play based on ceiling projections. Note that this does not even take into account PPM variance, where it’s possible that player 2 is more volatile per minute, and his upside could even be larger as a result. Finding a 10% owned player that is actually higher upside than the “higher projected” 25% owned player is a guaranteed way to boost your ROI immediately.
The first player I looked for today is Kemba Walker, and the most common way to project him would be to 1) estimate minutes and 2) estimate PPM, then multiply them. At 34 minutes * 1.15 PPM (reasonable estimates) he gets a projection of 39.1, making him a good play at $7700 on DraftKings.
However, I am looking for Kemba’s ceiling, so that will require some data and some analysis of the situation. Charlotte is in do-or-die mode going forth, and Kemba has been playing terribly lately, so he hasn’t seen big minutes. In some of Kemba’s bigger games he’s played 38-40 minutes, so it’s fair to say that if he’s on, 38 is reasonable. He also has an SD/AVG of about .25, so in order to get a ceiling type value we now use the following equation:
(38m)*(1.15)*(1.25) = 54.625, using ceiling minutes, base PPM, and the 1.25 coefficient (1+SD/AVG).
Compare this to a player who is likely to be popular today such as Isaiah Thomas, who we could give 33 minutes at 1.4 PPM (seems reasonable as a 15 point favorite vs PHX). His base projection is 46.2 and a very good play there, but there are some issues here. The first is that Boston is fairly healthy and this game could easily wind up in a blowout where IT2 doesn’t see 33 minutes. It’s certainly possible he could go more, but it’s so much more likely he does not that I won’t adjust my ceiling number. IT2’s SD/AVG on the season is 16.6%, so my ceiling formula is (33m)*(1.4)*(1.166) = 53.86, which actually rates him a worse play than Kemba.
This example is simply to illustrate how, using reasonable assumptions to make ceiling projections, a likely low owned player (Kemba) may actually be a much better GPP play than the likely higher owned IT2.
Just remember that if you are using ceiling minutes for projections, be careful as it can be like playing with fire. It is important to know what the current roster looks like and how close the point spread is and if it’s a critical game. Let’s look at some various players on the slate and see how we might be able to adjust our minutes from a common projection site into ceiling minutes.
Malcolm Brogdon: 29 min projection (conservative 32 min ceiling): In his past 10 games as a starter, Brogdon’s minutes have been all over the map from around 20-41 minutes. In a critical game vs ATL I expect Brogdon to have a big ceiling on minutes if he establishes himself as a hot hand early over some of the other guards (Dellavedova, Jason Terry). This is the exact type of player you want when his stock is low, since if he has a couple great 35 minute games, you’ll start to see projection sites raise his minutes projection, and it will be impossible to get him at a low ownership.
Giannis Antetokounmpo: 36 min (38 min): Given their urgency to make the playoffs, it’s not unrealistic to see Giannis get 40 minutes in games where he’s not in foul trouble. If it weren’t for this specific situation I would look to fade him as he should garner decent ownership coming off a great game. However, it can’t be denied that he has all kinds of upside here.
Rajon Rondo: 33 min (35 min): Another player in a “must win” scenario. Rondo could easily get 38 minutes in a strong game, or get phased out of the rotation in a dud. He’s generally a preferable play when he’s lost some momentum due to his volatile nature.
Kentavious Caldwell-Pope 33 min (36 min): DET is still on the playoff bubble, and KCP has been primarily quite bad of late. However, in games where he’s performed well, minutes in the upper 30’s/low 40’s have been there for him.
Nicolas Batum 33 min (36 min): It’s unclear why his minutes have been so down lately, but I don’t think there exist zero runouts where he doesn’t have this kind of minute upside.
Khris Middleton 32 min (34 min): I think that Kidd would prefer not to play him giant minutes, but he is a good candidate to be thrust into heavy minutes in certain close game situations.
DeMarcus Cousins 35 min (37 min): His major issue has been foul troubles, but in close games where he’s played well, 38-39 min is a real possibility. He comes with some momentum that could increase his ownership, but if you’re going to eat chalk, let it be with some good minute upside.
Dario Saric 32 min (35 min): This play is contingent on Jahlil Okafor not playing. Saric has volatile minutes due to game flow and foul trouble, but if he’s on I see no reason why he can’t get minutes.
Ivica Zubac 28 min (32 min): The Lakers plan is to play him 30-35 minutes down the stretch, so he just needs to play well and stay out of foul trouble to achieve this.
Alan Williams 25 min (28 min): One of the most volatile players you can roster, Williams is a foul magnet, and as a result can disappear for large stretches. On the flip side, starting C Alex Len is no stranger to foul trouble as well. Phoenix’s C situation is based on a combination of hot hand and foul trouble, so there will always low floors and high ceilings there.
Don’t be afraid to change your projections to account for minutes ceilings, just make sure that the numbers you use are grounded in reality!