Build a Winning Dream 11 Team Using Python and AI
Introduction
The exponential rise of IPL in India has correlatively resulted in the growth of fantasy cricket in the country. In recent years, we have seen more and more people playing on platforms like Dream 11, My Circle, etc. every day. Dream11 is a fantasy sports app where you can create your fantasy team for sports like cricket, football, kabaddi, etc. based on real-life matches. Once you build your team, you can join competitions and win daily cash prizes. In these games, for every single match, you have to create an 11-player sports team. The aim is to play with this team and earn maximum points to win maximum rewards. In this article, I will show you how to create a Dream 11 fantasy cricket team with which you can win up to Rs. 1 Crore in India, using Python and AI. So let’s hit it off!
Rules and Constraints for Selecting the Dream 11 Team
The team you will create on Dream 11 will have 11 players out of 22 players. Out of these, a maximum of 7 players can be from the same team in real life. Your Dream11 can have different combinations of players, but has to be within the 100-credit cap and must fulfill the following team selection criteria:
Building Your Dream 11 Team
Once you are done with your team, now it’s time to select your captain and vice-captain. In this article, as an example, I am going to create a team for a match between Chennai Super Kings and Punjab Kings.
We begin by running the following code:
!pip install openai==1.12.0
!pip install gradio==4.19.0
from bs4 import BeautifulSoup
import pandas as pd
from openai import OpenAI
import gradio as gr
from google.colab import userdata
client = OpenAI(api_key=userdata.get('OPENAI_API_KEY'))
def pull_match_data(m_link):
r = requests.get(m_link)
soup = BeautifulSoup(r.text, 'html.parser')
batting_columns = []
for i in soup.find_all(class_='ds-w-full ds-table ds-table-md ds-table-auto ci-scorecard-table')[:1]:
for j in i.find_all('th'):
batting_columns.append(j.text)
batting = []
for i in soup.find_all(class_='ds-w-full ds-table ds-table-md ds-table-auto ci-scorecard-table')[:1]:
rows = i.find_all('td', class_='ds-min-w-max')
for i in range(0, len(rows), 8):
batter = []
for j in range(i, i+8):
try:
batter.append(rows[j].text)
except IndexError:
break
batting.append(batter)
innings1_batting = pd.DataFrame(batting, columns=batting_columns)
batting_columns = []
for i in soup.find_all(class_='ds-w-full ds-table ds-table-md ds-table-auto ci-scorecard-table')[1:2]:
for j in i.find_all('th'):
batting_columns.append(j.text)
batting = []
for i in soup.find_all(class_='ds-w-full ds-table ds-table-md ds-table-auto ci-scorecard-table')[1:2]:
rows = i.find_all('td', class_='ds-min-w-max')
for i in range(0, len(rows), 8):
batter = []
for j in range(i, i+8):
try:
batter.append(rows[j].text)
except IndexError:
break
batting.append(batter)
innings2_batting = pd.DataFrame(batting, columns=batting_columns)
bowling_columns = []
for i in soup.find_all(class_='ds-w-full ds-table ds-table-md ds-table-auto')[:1]:
for j in i.find_all('th'):
bowling_columns.append(j.text)
bowling = []
for i in soup.find_all(class_='ds-w-full ds-table ds-table-md ds-table-auto')[:1]:
rows = i.find_all('td')
bowler = []
for i in range(0, len(rows)):
if len(bowler) < 11:
if len(rows[i].text) < 20:
try:
bowler.append(rows[i].text)
except IndexError:
break
else:
bowling.append(bowler)
bowler = []
if len(rows[i].text) < 20:
bowler.append(rows[i].text)
innings1_bowling = pd.DataFrame(bowling, columns=bowling_columns)
bowling_columns = []
for i in soup.find_all(class_='ds-w-full ds-table ds-table-md ds-table-auto')[1:2]:
for j in i.find_all('th'):
bowling_columns.append(j.text)
bowling = []
for i in soup.find_all(class_='ds-w-full ds-table ds-table-md ds-table-auto')[1:2]:
rows = i.find_all('td')
bowler = []
for i in range(0, len(rows)):
if len(bowler) < 11:
if len(rows[i].text) < 20:
try:
bowler.append(rows[i].text)
except IndexError:
break
else:
bowling.append(bowler)
bowler = []
if len(rows[i].text) < 20:
bowler.append(rows[i].text)
innings2_bowling = pd.DataFrame(bowling, columns=bowling_columns)
return innings1_batting, innings1_bowling, innings2_batting, innings2_bowling
def get_completion(m1_link, m1_batting_team, m1_bowling_team, m2_link, m2_batting_team, m2_bowling_team, m3_team1, m3_team2, model="gpt-4-turbo-2024-04-09"):
m1_innings1_batting, m1_innings1_bowling, m1_innings2_batting, m1_innings2_bowling = pull_match_data(m1_link)
m2_innings1_batting, m2_innings1_bowling, m2_innings2_batting, m2_innings2_bowling = pull_match_data(m2_link)
system_prompt = """
You are a huge cricket fan, and good with analysing match data.
"""
prompt = """
I will provide you batting & bowling data for two cricket matches.
Match 1 data provided between three backticks.
Match 2 data provided between separator ####.
Now, one team from each of these two matches are playing tonight.
Your task is to analyse the two matches data.
And create a Fantasy team for tonight's match between {m3_team1} & {m3_team2}.
This fantasy team should have 11 players.
```
Match 1: {m1_batting_team} vs {m1_bowling_team}
Innings 1 >> {m1_batting_team} Batting
{m1_innings1_batting}
Innings 1 >> {m1_bowling_team} Bowling
{m1_innings1_bowling}
Innings 2 >> {m1_bowling_team} Batting
{m1_innings2_batting}
Innings 2 >> {m1_batting_team} Bowling
{m1_innings2_bowling}
```
####
Match 1: {m2_batting_team} vs {m2_bowling_team}
Innings 1 >> {m2_batting_team} Batting
{m2_innings1_batting}
Innings 1 >> {m2_bowling_team} Bowling
{m1_innings1_bowling}
Innings 2 >> {m2_bowling_team} Batting
{m2_innings2_batting}
Innings 2 >> {m2_batting_team} Bowling
{m2_innings2_bowling}
####
"""
messages=[
{"role": "system", "content": system_prompt},
{"role": "user", "content": prompt}
]
response = client.chat.completions.create(
model="gpt-4-turbo-2024-04-09",
messages=messages,
temperature=0
)
return response.choices[0].message.content
iface = gr.Interface(
fn=get_completion,
inputs=[
gr.Textbox(label="Team 1 Last Match Scorecard", lines=1
After running the code, you will get this interface:
Here’s how to create your Dream11 fantasy cricket team using the above interface.
- Fill in ‘Team 1 Last Match Scorecard’
Input the value of your first team’s scorecard from the last match. For reference, you can copy-paste this website link which contains the match between Chennai Super Kings and Sunrisers Hyderabad.
- Select ‘Team Batting First’
Input the team that chose to bat first in that match.
- Select ‘Team Bowling’
Input the team that chose to bowl first in that match.
- Add ‘Team 2 Last Match Scorecard’
Input the value of your second team’s scorecard from the previous match. For reference, you can copy-paste this website link which contains the match between Kolkata Knight Riders and Punjab Kings.
- Select ‘Team Batting First’
Input the team that chose to bat first in that match.
- Select ‘Team Bowling’
Input the team that chose to bowl first in that match.
- Select ‘Teams Playing Today’
Choose the two teams who will be playing today’s fantasy match.
- Enter ‘GPT Model API’
Enter your GPT Model API or choose ‘gpt-3.5-turbo-1106’, if unsure.
- Click Submit
And with that final click, here is your dream team!
Conclusion
In this article, we explored the exciting possibility of using Python and AI to create a winning Dream11 team. By leveraging historical data and machine learning techniques, you can potentially gain an edge in selecting the best players and maximizing your points. Remember, while AI can be a powerful tool, factors like player form and pitch conditions also play a crucial role. So, for the ultimate Dream11 success, combine AI insights with your own cricketing knowledge. Now go ahead and take on the world of fantasy sports