Analytical Musings on the Value of Points
Posted: Tue May 16, 2017 3:23 pm
Introduction
At its most rudimentary, a player's skill level can be approximated by the current and future color rankings (along with an eyeball measurement of skills in certain areas). If we take it just a single step further, we can add up all of a player's skill points in each area to see how many total skill points (including athletic points) that player has. Now, with the new training system, PBSL has arbitrarily assigned a point system point value (henceforth referred to as PSP Value) to each skill point and each athletic point. These values are as follows:
CURRENT PLAYER SKILL POINTS:
In any case, now that we quantifiably know the PSP value of a skill point or an athletic point, we can look at the overall values of all of the players in the league in another measurable way.
The Data
Please enjoy the following spreadsheet to reference while reading this article:
https://docs.google.com/spreadsheets/d/ ... sp=sharing
COLUMN DEFINITIONS
Name: The player's name, ya dingus.
Team: The team the player is on (unless he was traded recently)
Pos: The player's primary position as of the creation of the spreadsheet.
Pure_tot: The number you get if you just add up all of their current skill and athletic ratings
Adj_tot: The number you get if you add up what each individual skill or athletic point is worth according to the new system (basically, how many Point System points would it take to create that player from scratch, disregarding potential band increases)
Pure_ranking: The player's ranking in comparison to the rest of the league based on pure_tot
Adj_ranking: The player's ranking in comparison to the rest of the league based on adj_tot
Comb_ranking: The previous two rankings added together. A lower comb_ranking is better. This is how the sheet is sorted initially.
Change: The amount the player's ranking would change from a look at pure_tot to adj_tot
CAVEATS:
1. I pulled all the data for this analysis after training camp, but before player training. Any training that's been done to these players is NOT reflected in the data set.
2. Since there was no value for points in the A-band (due to inability to train), I added .3 points to the player skill points (as is consistent with the increases per band) and .4 to the athletic points (as is consistent per band except for the .6 point increase from 0-20 to 21-40, which may or may not have been intentional). An A band skill point has a PSP Value of 2.4 and an A band athletic point, a PSP Value of 3.4.
The Findings
For the most part, players stay in the same general vicinity in terms of rankings. While pure_tot puts Kevin Love as the best player in the league (which seems to be a pretty well agreed upon prognosis), Deron Williams has the highest adj_total, and even though he's not a purple currently, I think that's a fair argument that can be made, given how the Pacers would almost never miss the playoffs without him. Of course, it's not completely apples to apples when comparing a point guard to a power forward. Deron is helped by PGs needing a more versatile skill set. The average values per position are as follows:
Average pure_tot:
However, Kevin Durant and LeBron James both make ranking gains in the upper echelon of the league when viewed by their adjusted totals. It's really quite simple to explain...LBJ and KD are both insane in the athletic points department. Whereas Rudy Gay and Paul George have comparable skills, they lack the same athleticism to rise up in these rankings.
On that note, let's look at what types of players had the biggest losses and gains in terms of overall league ranking when moving from a pure to an adjusted system.
BIGGEST LOSERS
Guess what? It's mostly old people! Tim Thomas, Stephen Jackson, and Rashard Lewis are the biggest ranking losers. It's pretty much common sense. They're guys who can play 3/4 and have stayed around a while because they can do a little bit of everything. Unfortunately, they're old, so their bodies are decrepit, and their athletic points are exceedingly low.
Also of note, Nikola Jokic and Joel Embiid are ranked a lot lower under the adjusted system. Despite high potentials, these rookie players still need time to grow and develop...moving those B skill points to more valuable A skill points. Because of the discrepancy in athleticism between bigs and smalls, young bigs might get disproportionately penalized. As they grow stronger and increase stamina, their stock will rise here, but they're probably not going to get much quicker.
BIGGEST GAINERS
On the flipside, give those young Centers a few years to develop, and they'll start to look way better than their pure rankings suggest. Hassan Whiteside and Bismack Biyombo top the list, and guys like Andre Drummond and Greg Oden aren't too far behind. These players have all had time to develop the areas that a big needs to excel at while also gaining some strength and stamina over time. This shows me that an adjusted rating does a good job of valuing centers based on the skills that make for a good center. If we look at this again in a few seasons, I'm sure that guys like Jokic and Embiid will be on the flipside of the coin.
Also on this side of the coin? Athletic guards with average skills. Look at guys like Louis Williams, Sundiata Gaines, Tony Wroten, and even guys like Zach LaVine and DeMar DeRozan. While none of them (at least right now) have a skill set that blows me away, athleticism maximizes the skills these players do have. DeRozan has a decent inside game, but his shooting is subpar. Fortunately, his athletic skills give him the toolset to be a 20 PPG type player. While I'd much rather have an Andre Drummond than a Louis Williams, I think that there are definitely spots for underskilled but overathletic guards to contribute to teams.
Of course, all of these findings seem like common sense to me. Maybe a spreadsheet wasn't exactly necessary to do anything but confirm things we already assumed, but it's kind of neat to look at, and it shows things like how even players like Chris Paul, Chris Bosh, and Kevin Durant are when measured.
Final Take
So, mess around with the spreadsheet, but take all of this with a grain of salt. Athleticism is really important with these numbers because we've decided that to build a player's athleticism should cost more than it should to build a player's skills. Whether that is fair or not would require an analysis of how skill points and athletic points translate to production. However, there are so many other factors that can affect a player's value: contract, team fit, hidden ratings, etc.
The best part of the game, though, is that these things remain intangible. If we abandoned intuition for pure formulae, the magic would be lost, and we'd just unabashedly be doing math without the layer of fun.
Anyway, I hope you enjoyed this summary enough to discuss it below!
...also, gimme them 5 points
At its most rudimentary, a player's skill level can be approximated by the current and future color rankings (along with an eyeball measurement of skills in certain areas). If we take it just a single step further, we can add up all of a player's skill points in each area to see how many total skill points (including athletic points) that player has. Now, with the new training system, PBSL has arbitrarily assigned a point system point value (henceforth referred to as PSP Value) to each skill point and each athletic point. These values are as follows:
CURRENT PLAYER SKILL POINTS:
CURRENT PLAYER ATHLETIC POINTS:F = 1.2 points
D = 1.5 points
C = 1.8 points
B = 2.1 points
A = may not be trained
This new training system determines that a skill or athletic point is worth more at the elite level than it is at the ground floor. For example, if player A and player B both gain 5 points in offensive rebounding, and player A is already a great offensive rebounder while player B is a poor offensive rebounder, player A's gain is more significant. In theory, this makes sense. It should be easier to learn the basics of any particular skill than it should be to master that skill.0-20 = 1.6 points
21-40 = 2.2 points
41-60 = 2.6 points
61-80 = 3.0 points
81+ = may not be trained
In any case, now that we quantifiably know the PSP value of a skill point or an athletic point, we can look at the overall values of all of the players in the league in another measurable way.
The Data
Please enjoy the following spreadsheet to reference while reading this article:
https://docs.google.com/spreadsheets/d/ ... sp=sharing
COLUMN DEFINITIONS
Name: The player's name, ya dingus.
Team: The team the player is on (unless he was traded recently)
Pos: The player's primary position as of the creation of the spreadsheet.
Pure_tot: The number you get if you just add up all of their current skill and athletic ratings
Adj_tot: The number you get if you add up what each individual skill or athletic point is worth according to the new system (basically, how many Point System points would it take to create that player from scratch, disregarding potential band increases)
Pure_ranking: The player's ranking in comparison to the rest of the league based on pure_tot
Adj_ranking: The player's ranking in comparison to the rest of the league based on adj_tot
Comb_ranking: The previous two rankings added together. A lower comb_ranking is better. This is how the sheet is sorted initially.
Change: The amount the player's ranking would change from a look at pure_tot to adj_tot
CAVEATS:
1. I pulled all the data for this analysis after training camp, but before player training. Any training that's been done to these players is NOT reflected in the data set.
2. Since there was no value for points in the A-band (due to inability to train), I added .3 points to the player skill points (as is consistent with the increases per band) and .4 to the athletic points (as is consistent per band except for the .6 point increase from 0-20 to 21-40, which may or may not have been intentional). An A band skill point has a PSP Value of 2.4 and an A band athletic point, a PSP Value of 3.4.
The Findings
For the most part, players stay in the same general vicinity in terms of rankings. While pure_tot puts Kevin Love as the best player in the league (which seems to be a pretty well agreed upon prognosis), Deron Williams has the highest adj_total, and even though he's not a purple currently, I think that's a fair argument that can be made, given how the Pacers would almost never miss the playoffs without him. Of course, it's not completely apples to apples when comparing a point guard to a power forward. Deron is helped by PGs needing a more versatile skill set. The average values per position are as follows:
Average pure_tot:
Average adj_tot:+------+----------+
| pos | AVG(tot) |
+------+----------+
| C | 752.7353 |
| SG | 757.8481 |
| PG | 820.3824 |
| PF | 768.9221 |
| SF | 794.2917 |
+------+----------+
Perhaps the most interesting takeaways from looking at how the adjusted totals affect player rankings on the elite level are how small forwards are affected. Paul George and Rudy Gay are both P/P players, and at the small forward spot, which means doing a lot of things very well without necessarily specializing in a particular area (for example, a solid C really just needs to get them boards, play good D, and maybe have a solid inside game). Thusly, weighting A-band points higher than F-band points should theoretically keep small forwards on a pretty even plane. However, both these current purple players drop off in adjusted rankings.+------+-------------+
| pos | AVG(tot) |
+------+-------------+
| C | 1293.444118 |
| SG | 1299.916456 |
| PG | 1446.304412 |
| PF | 1304.575325 |
| SF | 1356.016667 |
+------+-------------+
However, Kevin Durant and LeBron James both make ranking gains in the upper echelon of the league when viewed by their adjusted totals. It's really quite simple to explain...LBJ and KD are both insane in the athletic points department. Whereas Rudy Gay and Paul George have comparable skills, they lack the same athleticism to rise up in these rankings.
On that note, let's look at what types of players had the biggest losses and gains in terms of overall league ranking when moving from a pure to an adjusted system.
BIGGEST LOSERS
Guess what? It's mostly old people! Tim Thomas, Stephen Jackson, and Rashard Lewis are the biggest ranking losers. It's pretty much common sense. They're guys who can play 3/4 and have stayed around a while because they can do a little bit of everything. Unfortunately, they're old, so their bodies are decrepit, and their athletic points are exceedingly low.
Also of note, Nikola Jokic and Joel Embiid are ranked a lot lower under the adjusted system. Despite high potentials, these rookie players still need time to grow and develop...moving those B skill points to more valuable A skill points. Because of the discrepancy in athleticism between bigs and smalls, young bigs might get disproportionately penalized. As they grow stronger and increase stamina, their stock will rise here, but they're probably not going to get much quicker.
BIGGEST GAINERS
On the flipside, give those young Centers a few years to develop, and they'll start to look way better than their pure rankings suggest. Hassan Whiteside and Bismack Biyombo top the list, and guys like Andre Drummond and Greg Oden aren't too far behind. These players have all had time to develop the areas that a big needs to excel at while also gaining some strength and stamina over time. This shows me that an adjusted rating does a good job of valuing centers based on the skills that make for a good center. If we look at this again in a few seasons, I'm sure that guys like Jokic and Embiid will be on the flipside of the coin.
Also on this side of the coin? Athletic guards with average skills. Look at guys like Louis Williams, Sundiata Gaines, Tony Wroten, and even guys like Zach LaVine and DeMar DeRozan. While none of them (at least right now) have a skill set that blows me away, athleticism maximizes the skills these players do have. DeRozan has a decent inside game, but his shooting is subpar. Fortunately, his athletic skills give him the toolset to be a 20 PPG type player. While I'd much rather have an Andre Drummond than a Louis Williams, I think that there are definitely spots for underskilled but overathletic guards to contribute to teams.
Of course, all of these findings seem like common sense to me. Maybe a spreadsheet wasn't exactly necessary to do anything but confirm things we already assumed, but it's kind of neat to look at, and it shows things like how even players like Chris Paul, Chris Bosh, and Kevin Durant are when measured.
Final Take
So, mess around with the spreadsheet, but take all of this with a grain of salt. Athleticism is really important with these numbers because we've decided that to build a player's athleticism should cost more than it should to build a player's skills. Whether that is fair or not would require an analysis of how skill points and athletic points translate to production. However, there are so many other factors that can affect a player's value: contract, team fit, hidden ratings, etc.
The best part of the game, though, is that these things remain intangible. If we abandoned intuition for pure formulae, the magic would be lost, and we'd just unabashedly be doing math without the layer of fun.
Anyway, I hope you enjoyed this summary enough to discuss it below!
...also, gimme them 5 points