You are on page 1of 10

2021-2023

Prediction IPL Player Price

Submitted By: -

NAME PRN
Akriti 21020841153
Rohan Parakh 21020841249
Siddhartha Mishra 21020841208
Raghav Jhawar 21020841032
Vaibhav Raju Sherekar 21020841192
Madhulika Srivastava 21020841197

Submitted to:

Dr. Swarnalakshmi Umamaheswaran

1
Study of Data and changes
The total number of rows and columns are less in number with 131 rows and 28 columns, data
cleansing is done on excel. The column names are not fully given and it’s hard to understand when
regression is done. Below are the changes made to the column names and there are two columns with
Total runs and Total wickets are added. With the total runs and total wickets columns it will be easy
to look at the player’s data to their sold price.
T-RUNS = TEST RUNS
T-WKTS = TEST WKTS
RUNS-S = IPL RUNS S
HS = HIGHEST SCORE
RUNS-C = IPL RUNS C
WKTS = IPL WKTS
TOTAL RUNS = TEST RUNS + IPL RUNS S + ODI-RUNS-S
TOTAL WICKETS = TEST WKTS + IPL WKTS + ODI-WKTS

Correlation
Based on the Pearsons R Heatmap, we can see there is a correlation between -

TEST RUNS – ODI RUNS S – A cricket player who is usually selected for the ODI team will have
better chances to play in a test match. All the players who play ODI matches have the high chances
of performing well in Test matches. So the runs scoring in Test match and ODI match will have a
correlation. Need to be that all the players who play well in ODI excel in Test Cricket

TEST WKTS – ODI WKTS – Same goes with the bowler. Bowler who bowls well in the ODI
matches will have a higher chance of getting more wickets in Test matches too. Though the bowler
has to bowl for many overs in a test match, the performance can have a correlation with the ODI
matches.

IPL RUNS S – HIGHEST SCORE – AVG – SIXERS – IPL is a cricket league with mixture of star
players and entertainment. The highest score will improve the IPL runs and average of the player,
need not be that highest score is achieved only based on sixers but a high probability that the highest
score can be achieved by hitting sixers. For a player who is coming at the end of the match will get

2
less balls to face and hitting a six will improve the average and strike rate but not the high score. For
few bowlers, the highest score and IPL runs will be low and sixers will be high, comparing with it
will not do justice for the player’s profile.

Pearsons R Heatmap

3
Linear Regression (Iteration- 1)

R2 value is 0.660. TOTAL RUNS column which is a sum of TEST RUNS, IPL RUNS S, ODI-RUNS-
S and TOTAL WICKETS column which is a sum of TEST WKTS, IPL WKTS, ODI-WKTS is
considered.
Country, team, player role is added in the factors. There is an observation made. There are teams
with ‘+’ sign. We are considering the data of the players who are currently in IPL team. No need to
know much on the players changing the teams or played in the past for different teams. So the ‘+’ can
be removed in the team’s column.

4
5
Linear Regression – (Iteration -2)
After removing the ‘+’ sign from the teams there is a change in R2 value. There is a drop from 0.660
to 0.633. In Jasp, the collinearity statistics (VIF) can’t be processed for the columns which are dragged
in factors. So, those three are removed for further analysis.

6
Linear regression (Iteration -3 without factors)

There is a significant drop in the R2 value from 0.633 to 0.531. P- value for Total runs – 0.038, Total
wickets – 0.048, Age – 0.011, Sixers - < 0.001 is less than 0.05. Total runs and total wickets can be
taken as significant to judge the players profile. These two are one of the main things for a batsmen
and bowler. As the player ages, the performance levels decreases. It happens in every sports.
Considering age is important and significant to bid the player. As mentioned earlier, IPL is a cricket
league with entertainment, sixers are the most important for any player’s profile. They decide him as
finisher or hitter.

The VIF for Average, Highest Score, Ave-BL, SR-BL is above 5. The average bowling rate and strike
rate of bowler is important but those totally doesn’t decide the bowlers ratings. Though the bowler
doesn’t take a wicket in the death overs and doesn’t give away runs. Then the performance is called
good. Sometimes the dot balls matters in cricket rather than wickets. So those two columns can be
removed.

7
Linear regression (Iteration 4 removing Ave-BL, SR-BL)

There is slight decrease in the R2 value from 0.531 to 0.524. VIF for highest score and average is
above. Average for a middle order batsman will be high because most of the times the chance comes
to bat during the end of the matches or during the middle of the matches. But if the whole match is
played by the openers and middle order batsman comes at the end, who plays with tension free and
scores runs, then both average and highest score won’t be justified. Highest score brings significance
only in one match. The same level of performance can’t be maintained throughout the IPL season. So
both highest score and average can be removed. ODI strike rate of batter and bowler varies according
to the position they play. A player hitting a six on the first ball and getting out on the second ball will
have a strike rate of 300. Strike rate will vary for an opener and bowlers. ODI SR B, ODI SR BL, SR
B can be removed.

8
Linear Regression – (Iteration 5) Removed HIGHEST SCORE, AVE, ODI SR B, ODI SR BL,
SR B

R2 value is 0.516. The final model has these 8 variables by dropping these HIGHEST SCORE, AVE, ODI
SR B, ODI SR BL, SR B coefficients as they are not much significant

To check the batsman performance, the


TOTAL RUNS more number of runs will lead to better
price
To check the bowlers performance, the
TOTAL WICKETS more number of wickets will lead to
better price
Increase in the age will lead to the
AGE
decrease in the bidding price
More number of sixers, higher the
SIXERS
bidding price
The less number of runs conceded will
IPL RUNS C help the bowler to get more bidding
price
A reasonable economy for a bowler,
ECON
higher the price
Every year the auction price seems to
AUCTION YEAR
increase for players
The starting point for any player. There
BASE PRICE can be chances of player getting sold for
base price.

9
10

You might also like