Sample a team to play Dream11

The code can create a random team(s) given playing 11 of both the teams.

• Name of the teams in variable t1 and t2 respectively.

  Sample input:

  '''

  t1 = 'srh'

  t2 = 'kkr'

  '''

• Following details of all the players to be stored in the variable p1, p2, ..., p22, which are all dictionaries with the below key-value pairs.

  name : Name of the player (To be input as strings; should be unique)

  type: 'wk', 'bat', 'bowl', 'ar' for wicket-keeper, batsman, bowler, and all-rounder respectively. (To be input as one of the variables wk, bat, ar and bowl depending on the type of the player on main.py file)

  points: Cost of the player from Dream11 App (To be input as integer or floating point numbers)

  team: Name of the team player belongs to (To be input as variable t1/t2 on main.py file)

  star player identifier: 1 or 0; all players with value 1 for this key will be selected in all the teams and selection will only happen on players with the value 0 for this key

• Number of teams to be sampled given by variable 'numOfTeams'

A random combination is chosen from the below list of possible combinations:

[1,3,2,5], [1,3,3,4], [1,4,1,5], [1,4,2,4], [1,4,3,3], [1,5,1,4], [1,5,2,3]

• Step1: Each of the tuple in the above list refers to a team combination where the four numbers denote number of wicket-keepers, batsman, all-rounder and bowler respectively in a hypothetical playing 11. A random combination is selected with equal probability for each of the combination. (Will be configurable in the next version)

• Step 2: Repeat below for each of the four categories i.e. wicket keeper, batsman, all-rounder and bowler.

  Players are divided between stars and non-stars players set.

  All star players are included in the team by default. In case number of star players in a category exceeds required number of players, a random sample out of the star players is taken (Using softmax probabilities derived from the star players set)

  After star players are included and their are still a few spots to fill from the non-star players, a random comibination is selected from the non-star players set (Using softmax probabilities derived from the non-star players set)

• Step 3: The sampled team from step 2 has to satisfy a number of criterias as follows:

  • A minimum of 4 and a maximum of 7 players are allowed from each team.

  • Maximum number of points to create team is limited to 100.

  • The team should not be very similar to already selected teams. To check this, Dis-similarity index is defined as follows:

    To calculate the dis-similarity index, all teams (currently sampled team and already selected teams) are converted to a 22 dimensional vector with each dimension corresponding to 22 players playing in the match. The value of the entry will be equal to the points for the player if the player is part of sampled team, otherwise 0.

    The L2 norm is computed for the sampled team vector with the team vector of previously selected teams. This number has to be greater than a threshold to pass the dis-similarity criteria.

  If the team satisfies the above criteria, then team is added to the list of already selected teams.

Sample teams with a few additional details put together

Sample Output for a hypothetical SRH vs KKR match:

------------ Team 1 ------------

S.No. Player C/VC Team Type Points

1 J BAIRSTOW SRH wk 10.0

2 D WARNER SRH bat 11.0

3 S GILL vc KKR bat 8.0

4 D HOODA SRH bat 8.0

5 A RUSSELL KKR ar 10.5

6 S AL HASAN c SRH ar 8.5

7 RASHID KHAN SRH bowl 9.0

8 S KAUL SRH bowl 8.5

9 L FERGUSON KKR bowl 8.5

10 BHUVI SRH bowl 8.5

11 P KRISHNA KKR bowl 8.0

Total points invested: 98.5

Wicket-keepers: 1 , Batsman: 3 , All-rounders: 2 Bowlers: 5

SRH : 7 KKR : 4

Similarity indices: []

------------ Team 2 ------------

S.No. Player C/VC Team Type Points

1 J BAIRSTOW SRH wk 10.0

2 D WARNER SRH bat 11.0

3 C LYNN KKR bat 9.5

4 N RANA KKR bat 9.0

5 D HOODA c SRH bat 8.0

6 S GILL KKR bat 8.0

7 A RUSSELL KKR ar 10.5

8 V SHANKAR vc SRH ar 8.5

9 BHUVI SRH bowl 8.5

10 S SHARMA SRH bowl 8.5

11 P KRISHNA KKR bowl 8.0

Total points invested: 99.5

Wicket-keepers: 1 , Batsman: 5 , All-rounders: 2 Bowlers: 3

SRH : 6 KKR : 5

Similarity indices: [24.76893215300167]

------------ Team 3 ------------

S.No. Player C/VC Team Type Points

1 J BAIRSTOW c SRH wk 10.0

2 D WARNER vc SRH bat 11.0

3 N RANA KKR bat 9.0

4 M PANDEY SRH bat 8.5

5 A RUSSELL KKR ar 10.5

6 S NARINE KKR ar 9.0

7 V SHANKAR SRH ar 8.5

8 S SHARMA SRH bowl 8.5

9 L FERGUSON KKR bowl 8.5

10 P CHAWLA KKR bowl 8.5

11 P KRISHNA KKR bowl 8.0

Total points invested: 100.0

Wicket-keepers: 1 , Batsman: 3 , All-rounders: 3 Bowlers: 4

SRH : 5 KKR : 6

Similarity indices: [29.609964538985857, 24.253865671269807]

------------ Team 4 ------------

S.No. Player C/VC Team Type Points

1 J BAIRSTOW c SRH wk 10.0

2 D WARNER SRH bat 11.0

3 R UTHAPPA KKR bat 9.0

4 D HOODA SRH bat 8.0

5 S GILL KKR bat 8.0

6 A RUSSELL vc KKR ar 10.5

7 S NARINE KKR ar 9.0

8 RASHID KHAN SRH bowl 9.0

9 P CHAWLA KKR bowl 8.5

10 K YADAV KKR bowl 8.5

11 S SHARMA SRH bowl 8.5

Total points invested: 100.0

Wicket-keepers: 1 , Batsman: 4 , All-rounders: 2 Bowlers: 4

SRH : 5 KKR : 6

Similarity indices: [27.050877989447958, 27.69927796892908, 26.90724809414742]