Application of Tactics in Technical and Tactical Analysis of Table Tennis Mixed Doubles Based on Artificial Intelligence Graph Theory Model.

Objective. To construct the graph theory model and grey correlation model of table tennis mixed doubles technology and tactics so as to provide a new research method for the analysis of table tennis mixed doubles technology and tactics. The method of graphic theory and video observation is used to study the top tactical indicators of the game, and the method of graphic theory and video observation is used to analyze the tactical data of men and women, which has a certain contribution to the world level of table tennis. The grey correlation analysis can be better applied to the technical and tactical analysis of table tennis mixed doubles. The analysis of the contribution rate of men and women in eight rounds shows that there is little difference in the net average score between male and female athletes, and female athletes are slightly higher than male athletes. The net average score of the serving round is much higher than that of the receiving round, which shows that the serving round has certain advantages for the server, and the winner is often better able to grasp the opportunity of the serving round and get higher scores. Male athletes should strengthen physical training and adapt to a higher level of confrontation.

1. Introduction

Graph is a mathematical model to describe the relationship between objects. It is an ideal tool to describe semistructured data. It has unique mathematical theory and mathematical thought. Many problems in real life can be described by diagrams, such as network flow, resource allocation, circuit optimization, web page sorting, search, process arrangement, and so on. Graph is also an important means to describe many data structures. For example, tree structure is an essential model of computer operating system and many application systems. Graph theory is a branch of applied mathematics. It is a mathematical model reflecting the relationship between some elements (generally binary relationship). Graph theory is an important theory and method to deal with semistructured data. With the development of big data processing technology, the related applications of semistructured data are more and more widely used. The concept and achievements of graph theory and its application are very extensive. There are not only many application problems from production practice, but also many theoretical problems from other disciplines.

Grey correlation analysis can measure the relative strength of a project affected by other factors in a grey system. The factors between the two systems change with time or different objects, which is called correlation degree. The higher the synchronous change degree of the two factors, the higher the correlation degree, and the lower the synchronous change degree of the two factors, the lower the correlation degree. Therefore, the grey correlation analysis method provides a quantitative measure for measuring the correlation degree between factors according to the similarity and difference of the development trend between factors. The basic idea of grey correlation analysis is to judge whether the relationship is close according to the similarity of sequence curve geometry. The closer the curve is, the greater the correlation degree between the corresponding sequences is, and vice versa. At present, the application of graph theory and grey correlation model in the field of sports is rare.

Table tennis doubles is not an item that every association attaches great importance to in the world championships, and its gold content is low, but since the mixed doubles projects become after the Tokyo Olympic Games official competition projects, each country and society for table tennis doubles projects investment increase gradually. They are looking forward to achieving excellent results in this event and breaking the monopoly of the Chinese table tennis team on the Olympic gold medals. In the just-concluded Tokyo Olympic Games, the mixed doubles players Xu Xin/Liu Shiwen lost to Japan in the final composite Jun Mizutani/Mima Ito, unfortunately lost in the history of the first gold medal in the Olympic table tennis doubles, under the background of the Chinese scholars in Paris period preparing for the Olympic Games should be strengthened to research on the laws of the mixed doubles table tennis project winning. It provides theoretical reference and practical basis for coaches and athletes in China.

At present, there are many research achievements on evaluation and diagnosis methods of table tennis technology, from the Chinese classic three-stage index evaluation model to the use of computer and mathematical models to diagnose table tennis skills and tactics [1-14], all of which have made great contributions to the development of table tennis skills and tactics analysis. However, most of the existing research results are based on the technical and tactical analysis methods of singles matches, and there are few research results on the technical and tactical analysis of doubles and mixed doubles. A few research results on the technical and tactical analysis of mixed doubles master's theses are based on the descriptive analysis of some players' technical and tactical characteristics, and the conclusions are relatively simple. Athletes tactics in table tennis match, evaluation by level is a major factor that depends on the cohesion of the game level. Since the current research of table tennis tactics cohesion is very few, it is of great importance to research in this paper, from the current high levels of mixed doubles technique analysis, the first attempt to the method of using graph theory model, combined with longitudinal data. In order to provide more scientific reference for the training and competition of table tennis mixed doubles, this paper analyzes the skills and tactics of table tennis mixed doubles competition and constructs the linking model of skills and tactics of table tennis mixed doubles.

2. Data Sources

The data in this study are longitudinal data, which refer to the data obtained from multiple observations of the same case at different times. Throughout the data, it is very important in the analysis of ball techniques and tactics. There are two core reasons: (1) The information provided by the data is richer than that provided by ordinary data; (2) it can meet the needs of causal inference. Because the same pair of table tennis players were tracked and observed, the analysis was controlled in time order, which could be more effective in the analysis of causality. This distinction between difference and change is very helpful for the exploration of causality and makes descriptive analysis and exploratory analysis of data more reliable.

In this paper, the object of study for the world's top two mixed doubles player technology and tactics through data is our country mixed doubles combination Xu Xin/Liu Shiwen and Japanese mixed doubles combination Jun Mizutani/Ito beauty cheng for nearly three years of four games in each round, a total of 18 games, a total of 345 were analyzed, and a round. The four matches are the 2019 Korea Open Semifinal, 2019 Sweden Open Final, 2019 International Table Tennis Federation Final, and 2020 Germany Open Final. Cross data for 20 world high-level mixed doubles competition were gathered.

3. Research Methods

3.1. Selection of Indicators

In table tennis doubles and mixed doubles, the service lines are fixed, and all four players use forehand service. The service methods and hitting points are numbered, as shown in Table 1.

Table 1

Service mode and number of hitting points.

Serial number	Serve way	Hit point	Hit point
a	Side spin	A	Diagonal line
b	Reverse spin	B	Center line
c	Screw spin	C	Straight line

In order to study the table tennis technology, this study adopts the mixed double table tennis technology, as shown in Table 2. The technical and tactical numbers can be given by using the contents of Tables 1 and 2. If Ito picks the middle road with his backhand, it can be numbered 6B. For example, Xu Xin uses the forehand pull technique to return to the position of the opponent's line, numbered xu 7C, and so on.

Table 2

Technical application number.

Number	Skill	Number	Skill
1	Forehand drop	8	Backhand loop
2	Backhand drop	9	Backhand bring
3	Forehand deep ball	10	Forehand bring
4	Backhand deep ball	11	Forehand block
5	Forehand flick	12	Backhand block
6	Backhand flick	13	Forehand sideways
7	Forehand loop

4. Results of the Study

4.1. Analysis of 4-Field Longitudinal Data

4.1.1. Analysis of Service Data

This paper provides the data of the world top two table tennis, Chinese mixed doubles combination Xu (male)/Liu (female) and Japanese mixed doubles Jun (male)/Ito (female) 4 games data. The process of each ball of rounds was recorded; the game data marked the serve players and each ball hit process. The four matches are 2019 mixed doubles finals (5 games), German Open (4 games), Swedish Open (5 games), and Korea Open (4 games), with a total of 18 games and 345 rounds. Among them, Chinese players scored in 179 rounds, accounting for 51.88%; Japanese players scored in 166 rounds, accounting for 48.12%. In order to study the influence of service players on the outcome, this paper counts the number of different tees in 345 rounds, including 170 Chinese, 174 Japanese serves, one missing service data, and only 344 serves, as shown in Table 3.

Table 3

Statistics of service times for different players.

Country	Sex	Surname and personal name	Number of serves
China	Female	Liu Shiwen	93
	Male	Xu Xin	77
Japan	Female	Ito	97
	Male	Jun	77

From Table 3, we can see that male players serve 154 times, female 190 times, Chinese Liu Shiwen (female) 93 times, Chinese Xu Xin (male) 77 times; Japanese Ito (female) 97 times and Japanese Jun (male) 77 times. In the mixed doubles match of table tennis, the number of serves is not equal, which may be related to the competition strategy, player strength, and so on. In general, women serve more times than men, which may be related to their greater agility.

4.1.2. Analysis of Athlete Strike Data

This paper counted all different players in 345 rounds, all hitting techniques appeared, as shown in Table 4. As can be seen from Table 4, among the 13 different hitting techniques, table tennis hitting techniques are 7 and 8, forehand (attack) and backhand (attack), the least 4 and 10, backhand long and forehand fast, and the rest are in the middle.

Table 4

Frequency of the 13 different striking techniques.

Number	Strike technology	Frequency	Number	Strike technology	Frequency
1	The forehand is short	133	8	The backhand pull	167
2	The backhand is short	78	9	Back hand fast tear	22
3	Forehand split long	46	10	Forehand fast belt	8
4	The backhand splits long	9	11	Forehand defense (block)	60
5	Forehand pick hit	85	12	Backhand defense (block)	64
6	Back twist (pick)	71	13	Side body pull	120
7	Forehand pull (attack)	206

4.1.3. Map of Athlete Strike Mode

In this paper, taking Chinese players as an example, the graph theory method is used to summarize all the hitting methods of several players. Figure 1 and Figure 2 are the hitting methods of Chinese players Liu Shiwen and Xu Xin.

Figure 1

Matrix diagram of all striking modes of Liu Shiwen.

Figure 2

Matrix diagram of all striking modes of Xu Xin.

As we can see from Figure 1, in the 345 rounds of competition, Liu Shiwen hit the most in 8A (35), 1A (30), and 7A (30), namely, backhand pull (attack), forehand (attack), and forehand swing short combination oblique strike, followed by 3A (16), namely, forehand split long combination diagonal strike.

As we can see from Figure 2, in the 345 rounds of competition, Xu Xin hit the most in 7A (41), namely, the forehand pull (attack) combination oblique strike, followed by 13A (24) and 5 (23), namely, the side and forehand pick combination oblique strike, and the rest are in the middle.

To sum up, Liu Shiwen of China prefers backhand pull (attack), forehand pull (attack), and forehand swing short combination cross strokes, while Xu Xin of China prefers forehand pull combination cross strokes.

As can be seen from Figure 3 and Figure 4, in 345 rounds of matches, Ito hit 8A the most (35 times) and 7A (34 times), namely, backhand pull and forehand pull combination cross strokes, This is followed by 2A (22 times) and 8C (22 times), namely, backhand swing short combination slash and backhand pull (attack) combination straight line hit, other hits in the center.

Figure 3

Matrix diagram of all striking modes of ITO.

Figure 4

Matrix diagram of all striking modes of JUN.

Chinese player Liu prefers backhand (attack), forehand (attack), and forehand swing short combination slash strike, Chinese player Xu Xin prefers forehand (attack) combination slash strike, Japanese player Ito prefers backhand (attack) and forehand (attack) combination slash strike, and Japanese Jun prefers forehand (attack) combination slash strike.

5. Application of Grey Correlation Degree Analysis Model

This paper makes statistics on the final and semifinals of the world high-level table tennis competitions in the past five years, summarizes the eight rounds unique to the mixed doubles competitions, and constructs a model with the model analysis method of grey correlation degree.

In order to select the appropriate evaluation index, this article first treats male, male, male in 20 games. Grey association analysis for the scores and loss data of male, male, female, female, female (GCA). Among them, the male serving board order data includes 5 times, respectively: first board, third board, fifth board, male 7 and female 7; male receiving board sequence data includes 5 times, respectively: second board, fourth board, sixth board, male and female six; female receiving plate order data includes 5 times, respectively: second board, fourth board, sixth board, male and female six.

First, the comparison sequence (evaluation object) and the reference sequence (evaluation criteria) are defined, and the corresponding data in the final of the 55th World Championships in 2019 are taken as the reference sequence x(k); the other 19 matches for the comparison sequence are x(k). On this basis, the absolute difference sequence between the reference sequence and the comparison sequence for ∆(k): ∆i(k) = |x(k) − x(k)| (k = 1, 2,…, n; i = 1, 2,…, m) among ∆(k),(k = 1, 2,…, n; i = 1, 2,…, m) is the sequence of absolute differences. By definition, compare the sequence x(k) with the reference sequence x(k). Association coefficient of the case, ζ(t) is calculated as follows:

The corresponding grey correlation for different variables R(x, x) is

Step 1 .

Normalize the data by maximum minimum normalization:

Step 2 .

Compute the comparison sequence, x(k). With the parameter sequence, x grey correlation coefficient of (k) is

Among them, the min|x(k) − x(k)| is the minimum value of the absolute difference sequence. It is called resolution coefficient. The value range is [0, 1], and the resolution coefficient value is 0.5.

Step 3 .

Draw the matrix heat map of the grey correlation coefficient.

The first index “male male” score data as the reference sequence, after calculating each index with the first.

After the grey correlation coefficients of the indicators, a matrix heat map of the grey correlation coefficients of the final score loss of the 20 games was drawn by using the heat map function in the MATLAB software, as shown in Figure 5.

Figure 5

Matrix heat map of grey correlation coefficient with “male” data as the reference sequence.

As we can see from Figure 5, among the winners of the 20 games, the male, male, male, female, female, female and female are basically positively related to the scoring and losing data of men. Among them, the sixth match is very special, namely, The 2018 Asian Games Final Chinese Wang Chuqin (male)/Sun Yingsha (female) against China's Lin Gaoyuan (male)/Wang Manyu (female). In this game, other figures and men's points and lost figures were basically weak positive correlation. Data from the first match “55 World Championships 2019” are used as the reference sequence x(k), after calculating the grey correlation coefficient between each match and the first match, a matrix heatmap of the grey correlation coefficient of the winner of the first match data as the reference sequence was drawn by using the heatmap function in the MATLAB software, as shown in Figure 6.

Figure 6

Matrix heat map corresponding to the grey correlation coefficient with the first match data as the reference sequence.

From Figure 6, we can see that the grey correlation coefficients are quite different between the different indicators, and, among which, the 15th index, namely: the grey correlation coefficient between the female score index and other indicators is small, and overall, the correlation is weak. Except that in game 9, the grey correlation coefficient was large at 0.8674. The performance of other rounds is not very obvious. For example, in the women's round, the correlation degree is quite different from that of other rounds. In the men's round, the score rate of the winner is significantly higher than that of the loser. In other rounds, the difference is not particularly large, but there are rules to follow.

Step 4 .

Calculate the grey correlation degree.

Grey associations were calculated from the grey association coefficient for each indicator, as shown in Table 5 and Table 6.

Table 5

Use of the first indicator “male male” data as the reference sequence.

R 1-1	R 1-2	R 1-3	R 1-4	R 1-5	R 1-6	R 1-7	R 1-8
1.000	0.724	0.720	0.730	0.751	0.771	0.764	0.754
R 1-9	R1-10	R1-11	R1-12	R1-13	R1-14	R1-15	R1-16
0.774	0.727	0.761	0.758	0.751	0.743	0.743	0.721

Table 6

The grey correlation degree corresponding to the first match data as the reference sequence is taken.

R 1-1	R 1-2	R 1-3	R 1-4	R 1-5	R 1-6	R 1-7	R 1-8	R 1-9	R 1-10
1.000	0.627	0.663	0.678	0.701	0.666	0.640	0.709	0.693	0.756
R 1-1	R 1-12	R 1-13	R 1-14	R 1-15	R 1-16	R 1-17	R 1-18	R 1-19	R 1-20
0.717	0.704	0.647	0.702	0.679	0.660	0.721	0.587	0.691	0.730

The above grey correlation data are based on the relevant data of the first game. The grey correlation research has a good analysis effect in the establishment of the gain and loss score model of mixed doubles in table tennis. Because of the space relationship, this paper will not give examples the remaining data and calculation process.

6. Discussion on the Contribution Rate of Men and Women in Mixed Doubles

Based on the grey correlation degree analysis, the overall score analysis of different serve rounds is obtained. First of all, the 20 world's top table tennis mixed doubles matches, men and women athletes in different service rounds were analyzed. In mixed doubles related to table tennis, it can be divided into eight rounds according to the different serve and receiver. They are recorded as male serve male, male serve female, male receive male, male receive female, female serve male, female serve female, female receive male, female receive female. The meaning of the name is: the first male/female is the winning player, the hair/receiver refers to the winner this round is serve/receiver, and the second male/female is the losing player. For example, the male player who fails to serve the ball for the first time refers to the male player who fails to serve the ball for the first time. Female receiving male finger: this round is for the male player of the losing side to serve, and the female player of the winning side receives the ball for the first time, and then the hitting order shall be carried out according to the mixed doubles rules. Gross score G represents the score of each round. Net score N represents the score minus the lost score under each round, as shown in Table 7. The net score contribution of female athletes was significantly improved in the four conditions of female hair and female receiver, except that the net score of female hair was lower than that of female athletes. In other cases, the net scoring contribution of female athletes plays a key role. This shows that the net score (or actual winning ability) of male and female athletes varies with the combination of hair extensions. Correlation differences can be evaluated, for example, whether the cooperation between male and female athletes is poor, or whether the opponent's scoring ability is high.

Table 7

Gross/net average scores in different service rounds.

	Male s male	Male r male	Male s female	Male r female	Female s male	Female s female	Female r male	Female r female
G	6.4	4.95	5.65	5.9	6.65	7.5	5.05	5.55
N	1.9	0.7	1.8	−0.25	2	2.9	0.35	−0.7

From Figure 7, it can be seen that female hair, female hair, and male hair have the highest scores in three rounds, and N is also the highest in the three cases, indicating that the winner has obvious advantages and strong scoring ability in these three rounds.

Figure 7

Average scores in different service rounds.

Through the analysis of the male and female athletes' sending and receiving of the victory side, it can be concluded that in the four cases of male hair, male receiving, female hair and female receiving, the G and N of female hair rounds are the highest, indicating that in the female hair rounds, the victory side has the strongest offensive ability, the least mistakes, and the strongest comprehensive scoring ability. Further, the four situations of male, female, hair, and reception are classified and analyzed, respectively. From the results, the following can be seen:

Whether male or female athletes receive the service for the first time, there is little difference in the net average score, and female athletes are slightly higher than male athletes.

The net average score of the serving round (7.5) is much higher than the net average score of the receiving round (1.2), indicating that the serving round has certain advantages for the server, and the winner is often better able to grasp the opportunity of the serving round and get higher scores.

Strategy:

When our players serve, female players should be given priority to serve and serve to female players of the opposing team. G and N is higher.

When the opposing player serves and our player receives the ball, the male player shall be given priority to receive the ball. The G and N phases of this serving round shall be considered for higher.

The analysis method of graph theory model [1, 15–23] and grey correlation degree can comprehensively evaluate the mixed doubles of table tennis. This method has certain intuition and effectiveness and can provide reference basis for the training and preparation of coaches and athletes. The graph theory established in this paper can also be used in the analysis of other sports events, such as badminton, volleyball, basketball, football, tennis, billiards, hockey, and baseball. Therefore, the mathematical model established in this paper has good popularization and application value, and the universality of the model is strong. It has more important significance and value for multiperson (more than two people) team cooperation sports competition.

This paper summarizes the technical and tactical analysis of mixed doubles based on Atlas analysis and makes an in-depth analysis and comparison of the technical and tactical characteristics of the two pairs of mixed doubles players on the basis of grey correlation analysis. In addition, it also makes a specific research and analysis on the respective contribution rates of male and female athletes and draws a lot of referential opinions. These conclusions can help mixed doubles coaches and athletes prepare and train effectively. But the deficiency of this paper is that it cannot calculate the contribution rate of men and women to what extent can win, which needs to be excavated in future research.