Zhiba NBA

Chapter 143 (140 II) Development of Advanced Data
On the flight home, Yuan Fei told everyone about Jason Bass' invitation in a half-concealed manner.After hearing the excitement, everyone cursed Bass one after another.

"This Jason actually wants to collude with the small boss and disrupt the fair environment of the alliance. It's an unforgivable crime!" Bosh was indignant.

"However, the Bass family is full of talents. Is he really so sure that he can inherit the Lakers?" Billups thinks farther, and his angle of consideration is a bit unique.

"We don't have to worry about who the Lakers will take over other people's affairs." Yuan Fei said with a smile: "No matter what, we are the defending champions now, and the Lakers are just outside the playoffs. They should be worried, not us."

"Jason's words are really arrogant, but I like it." Yao Ming also laughed.

"This guy is still a big fan of Kobe Bryant. He thinks Bryant is better than me? Interesting." McGrady also laughed, and what Bass said aroused his fighting spirit.

"Bass thinks that Bryant will become a figure like Jordan, and building a team around him can create a dynasty. Of course, I don't think so, and our Rockets don't agree." Of course, Yuan Fei will give McGrady support at this time.

"Jason, how do you compare superstars like Bryant, Tracy or Iverson?" Bosh asked suddenly.

"Well, this question is broad. If you talk about historical status, you must first consider honor. Tracy, who won the MVP in one year, and the championship and FMVP Grand Slam superstar will definitely give a lot of points. After all, Jordan will only be able to do this in the future. He and O'Neal Duncan, and those two can hit the top ten positions in history. If Tracy suddenly announces his retirement today, I think he can almost rank in the top forty in history." Yuan Fei thought for a while and replied.

"No, what I want to ask is, can the data reflect the value of players on the court? You can't say whoever scores more is better?" Bosh added.

"Oh, then you have to know what the data does first. The ideal data can reflect all the behaviors that occur on the court. However, due to the limited statistical ability and analysis ability, the current data is obviously not omnipotent, and there are many data that cannot be used for specific statistics. But fortunately, the data analysis of basketball is still developing and improving, and there will always be some better statistical standards." Usually translated as efficiency value), this statistic can better show the overall strength of a player."

"What is advanced data? What is per?" Bosh was confused.

"Advanced stats (Advanced.stats) are some data integrated on the basis of basic data. Like the rebound rate I told you last time, it is the advanced data evolved on the basis of the number of rebounds. It can better reflect the player's ability to grab rebounds than the number of rebounds." Yuan Fei talked about basketball theory and talked endlessly, "The efficiency value is used to reflect the average strength of the player compared to this era. The calculation method of this statistic is very complicated. Positive data such as offense, offensive rebounds, defensive rebounds, and steals also include negative data such as strikes and turnovers, and finally these are unified into one value. The league average is 15, which means that those above 15 are good players, those above 20 are considered stars, and those above 25 are almost superstars.”

"This thing is good, simple and easy to understand. Then who is in the front of our team? It should be Tracy?" Bosh asked: "How many are Bryant and Iverson? Which of them is taller?"

"You can't access the Internet on the plane. I'll check the statistics when I get back and I'll tell you." Yuan Fei said embarrassingly: "Another good statistic is called Win.Shares (WS for short). This statistic also calculates various basic data, and finally simulates each person's contribution to the team's winning.

"Then it's troublesome. It's hard to get into the playoffs with 39 wins." Bosh frowned.

"Of course it can't be completely calculated like this. Because Yao has always been on the field and has the ball, he has played very good statistics. If Yao can't play, Dapu will be the top starter. Because Dapu's offensive skills are not as good as Yao's. Even if your playing time remains the same, you will get more ball possessions, which will also lead to better PER and WS." Yuan Fei analyzed: "Of course the efficiency value and victory contribution value are not perfect. They are statistics made in a specific tactical environment. "

"It turns out that these advanced data are not particularly easy to use." Bosh seemed a little disappointed, "Then what obvious shortcomings do they have?"

"Let's talk about the efficiency value first. The biggest problem is that it is too much related to the possession of the ball, especially the right to shoot. In the calculation formula, it has a relatively light penalty for hitting irons and a high bonus for shooting shots. This leads to some players with average shooting ability but a lot of shots have an advantage in this statistic, while players with accurate shots but few shots are at a disadvantage. In other words, this thing is more suitable for those so-called data brushes. The disadvantage is that the efficiency value cannot reflect the defensive ability. There are some role players who are good at defense but rarely shoot. Not only can the PER not reach the average of 15, but it will even drop below 10. But in fact, they are all very good talents in the league."

"No way."

"There are more minor problems. For example, the bonus ratio of assists is 2/3. I don't know why the assists are set at 2/3. A bad pass is barely made due to the receiver's skill, and a good pass leads to the receiver. Easy dunk, both have the same bonus. This is the bottleneck of current data analysis. There is no way to specifically quantify the contribution of multiple people in a goal." Yuanfei continued: "For example, the contribution ratio after a block is multiplied by the backcourt rebound rate. This idea is good. , but not objective enough. After a block, not every ball becomes a rebound. For example, if you block the opponent out of bounds once, although the opponent still has the ball, but there is not much attack time left, it is actually equivalent to defense. Therefore, the efficiency value is a bit low for block statistics.”

"No, I'm confused." Bosh said, "Can WS be simpler?"

"Sorry, the algorithm of victory contribution is more complicated than the efficiency value. Its origin is upgraded from the algorithm of Bill James, a data expert in the baseball world. The main disadvantage of WS is that several calculations are based on the overall data of the team. This does not take into account the distinction between the main force and the substitute. After all, there is a difference between having good teammates and poor teammates. Relatively speaking, it is more beneficial to calculate WS with very healthy stars who start in strong teams."

"Okay, I completely give up learning data analysis from you. The more you study this stuff, the more difficult it is." Bosh said with a bitter face.

"For you, just learn how to use it. There is no need to delve into how the data comes from and what the algorithm looks like. Just find it." Yuan Fei thought for a while and said, "Actually, I advocate that players should learn a little data, so that they can improve according to their own conditions. Human energy is limited, and you should devote your limited energy to more effective training programs."

Bosh's eyes lit up, "Okay!"

　This chapter is purely hard work, please leave a message if you have reading difficulties.

　　In addition, there is a wrong PER calculation formula circulating on the Internet: appearance efficiency = ((points + rebounds + assists + steals + blocks) - ((number of shots - number of hits) + (number of free throws - number of free throws) + turnovers)) / number of appearances.This statistic is called the Player Efficiency Index, the calculation method is simple and rude, and it has no gold content.In terms of practical value, it is less than 1% of PER.Golfers please ignore it.

　　
　
(End of this chapter)