Introduction

Balatro is a popular card game that fuses poker and ability cards ( jokers, planets, tarots, and spectrals ) among other elements in order to masterfully create combinations that ensure you obliterate your round score.

Playing the game allowed me to notice that there were certain Joker cards that implicit synergy between them. An example of this is the “Four Fingers” joker card, which says that all straights and flushes can be made with 4 cards. the “Square Joker” states that this Joker gains +4 chips if the hand being played has exactly 4 cards. I would claim that these two cards have high synergy between them in the game, though there are other cards that obviously don’t necessary have that.

That got me thinking about other potential synergies among other jokers within the game, and wanted to create my own synergy tag system between the joker cards.

1. Can we find synergies among Jokers, and how many cards are universally standalone Joker cards which don’t really have synergies ?

2. Which Joker pairs have the strongest synergies based on the system created ?

Web Scraping - Obtaining joker data from wiki

library(tidyverse)
library(rvest)
library(stringr)

url <- "https://balatrogame.fandom.com/wiki/Jokers"
page <- read_html(url)

# grabs the third table that appears on the site
joker_table = page %>%
  html_elements("table") %>% .[-3]

# Get all table rows (skipping the header)
joker_rows <- joker_table %>% html_nodes("tr") %>% .[-1]


joker_df <- joker_rows %>% 
  map_df(~{
    cells <- html_nodes(.x, "td") %>% html_text2() %>% str_squish()
    tibble(
      name = cells[2],
      effect = cells[3],
      cost = cells[4],
      rarity = cells[5],
      unlock_requirement = cells[6],
      type = cells[7],
      activation = cells[8]
    )
    
  })

head(joker_df)

## # A tibble: 6 × 7
##   name                 effect   cost  rarity unlock_requirement type  activation
##   <chr>                <chr>    <chr> <chr>  <chr>              <chr> <chr>     
## 1 ... Retrigger Jokers +$ Econ… ""    <NA>   <NA>               <NA>  <NA>      
## 2 <NA>                 <NA>     <NA>  <NA>   <NA>               <NA>  <NA>      
## 3 Joker                +4 Mult  "$2"  .mw-p… Available from st… +m    Indep.    
## 4 Greedy Joker         Played … "$5"  Common Available from st… +m    On Scored 
## 5 Lusty Joker          Played … "$5"  Common Available from st… +m    On Scored 
## 6 Wrathful Joker       Played … "$5"  Common Available from st… +m    On Scored

After evaluation of table, I thought the data had been extract pretty well except for a few things:

Removing the first two rows that aren’t necessary
Having the original Joker’s columns overwritten with correct info
Translating the type column in order to properly assess cards

joker_df <- joker_df[-c(1:2),] # removing rows 1 & 2

# Changing row 1
joker_df$effect[1] <- "+4 Mult"
joker_df$rarity[1] <- "Common"

# translating type column 
joker_df <- joker_df %>%
  mutate(
    type = case_when(
      type == "+c" ~ "Chips",
      type == "+m" ~ "Additive Mult",
      type == "Xm" ~ "Multiplicative Mult",
      type == "++" ~ "Chips and Additive Mult",
      type == "!!" ~ "Effect",
      type == "..." ~ "Retrigger",
      type == "+$" ~ "Economy",
      TRUE ~ type
    )
  )

Creating our “synergy” patterns in order to evaluate Joker cards among themselves.

tag_patterns <- list(
  "4-card hand" = regex("four cards|4 cards", ignore_case = TRUE),
  "flush synergy" = regex("flush", ignore_case = TRUE),
  "straight synergy" = regex("straight", ignore_case = TRUE),
  "face card synergy" = regex("face card", ignore_case = TRUE),
  "king card synergy" = regex("king", ignore_case = TRUE),
  # ensure that the value is referring to a played card
  "even card synergy" = regex("(each played|each card|played cards|cards of value|card with value).*\\b(2|4|6|8|10)\\b", ignore_case = TRUE),
  "odd card synergy" = regex("(each played|each card|played cards|cards of value|card with value).*\\b(1|3|5|7|9)\\b", ignore_case = TRUE),
  "ace card synergy" = regex("ace", ignore_case = TRUE),
  "diamonds card synergy" = regex("diamond", ignore_case = TRUE),
  "hearts card synergy" = regex("heart", ignore_case = TRUE),
  "spades card synergy" = regex("spade", ignore_case = TRUE),
  "clubs card synergy" = regex("club", ignore_case = TRUE),
  "glass card synergy" = regex("glass card", ignore_case = TRUE),
  "stone card synergy" = regex("stone card", ignore_case = TRUE),
  "discard synergy" = regex("discard", ignore_case = TRUE),
  "joker synergy" = regex("each Joker card |leftmost joker|joker to the right|Jokers each", ignore_case = TRUE),
  "probability synergy" = regex("probabilit|chance", ignore_case = TRUE),
  "four of a kind synergy" = regex("four of a kind", ignore_case = TRUE),
  "three of a kind synergy" = regex("three of a kind", ignore_case = TRUE),
  "two pair synergy" = regex("two pair", ignore_case = TRUE),
  "high card synergy" = regex("high card", ignore_case = TRUE),
  "straight flush synergy" = regex("straight flush", ignore_case = TRUE),
  "poker hand synergy" = regex("poker hand", ignore_case = TRUE),
  "cards held synergy" = regex("held in hand", ignore_case = TRUE),
  "planet card synergy" = regex("planet", ignore_case = TRUE),
  "uncommon synergy" = regex("uncommon", ignore_case = TRUE)
)

# creating the tags column that will be used to identify potential synergies
joker_df <- joker_df %>%
  rowwise() %>%
  mutate(
    tags = {
      matched_tags <- names(tag_patterns)[
        vapply(tag_patterns, function(p) str_detect(effect, p), logical(1))
      ]
      list(if (length(matched_tags) > 0) matched_tags else "universal")
    }
  ) %>%
  ungroup()

# Unnest the tags list column to long format
tag_counts <- joker_df %>%
  select(name, tags) %>%
  unnest(tags) %>%
  count(tags, sort = TRUE)

tag_counts

# creating new df w/o cards with no 'synergies'
library(purrr)

tagged_jokers <- joker_df %>%
  filter(map_lgl(tags, ~ !"universal" %in% .x))

head(tagged_jokers)

Some takeaways from the table above:

Even vs. Odd: 5 more Jokers reference odd numbers than even numbers, suggesting odd-number-related effects might be found more frequently in game play.
Suit synergy: Clubs are the preferable choice when looking for joker synergies, Hearts and Diamonds are slightly less common.
Favorable hand Among Jokers effects: Straight and flush hands (respectively) are the most favorable hands, cumulatively being ≈62% of the all hands synergy.
Out of the 150 jokers in game, 59 (≈40%) don’t have any identified synergies based on the tags system created here. This maybe indicate that many Jokers offer standalone values and have abilities outside the typical mechanics described.

Looking to find strong synergies between Joker pairs

library(tidyverse)

# Create synergy pairs dataframe
create_synergy_pairs <- function(joker_df) {
  
  # Get all joker names
  joker_names <- joker_df$name
  n_jokers <- length(joker_names)
  
  # Create empty dataframe
  synergy_pairs <- data.frame(
    name.x = character(),
    name.y = character(),
    n = integer(),
    stringsAsFactors = FALSE
  )
  
  # Compare each pair of jokers
  for (i in 1:(n_jokers - 1)) {
    for (j in (i + 1):n_jokers) {
      
      joker1_name <- joker_names[i]
      joker2_name <- joker_names[j]
      
      # Get tags for each joker
      joker1_tags <- joker_df$tags[joker_df$name == joker1_name][[1]]
      joker2_tags <- joker_df$tags[joker_df$name == joker2_name][[1]]
      
      # Count shared tags
      shared_count <- length(intersect(joker1_tags, joker2_tags))
      
      # Add to dataframe
      synergy_pairs <- rbind(synergy_pairs, data.frame(
        name.x = joker1_name,
        name.y = joker2_name,
        n = shared_count,
        stringsAsFactors = FALSE
      ))
    }
  }
  
  # Sort by shared tags (highest first)
  synergy_pairs <- synergy_pairs %>%
    arrange(desc(n))
  
  return(synergy_pairs)
}

joker_synergy_pairs <- create_synergy_pairs(joker_df)

library(ggplot2)
library(dplyr)

# Get exactly the top 5 pairs
top_5_pairs <- joker_synergy_pairs %>%
  slice_head(n = 5) %>%
  mutate(pair = paste(name.x, "×", name.y))

# Create bar chart
ggplot(top_5_pairs, aes(x = reorder(pair, n), y = n)) +
  geom_col(fill = "#45876a") +
  coord_flip() +
  theme_minimal() +
  labs(
    title = "Highest Tag Count Among Jokers",
    subtitle = "based on 'synergy' pattern system",
    x = "Joker Pairs",
    y = "Shared Tags"
  ) +
  theme(
    text = element_text(size = 12, family = "sans"),
    panel.grid.major.y = element_blank(),
    panel.grid.minor = element_blank(),
    plot.title = element_text(color = "#ff9800", 
                              size = 18, 
                              hjust = 0, 
                              face = "bold"),
    plot.subtitle = element_text(hjust = 0)
    
  )

Some takeaways from the visualization above:

Tag overlap doesn’t guarantee synergy strength: While Flower Pot and Smeared Joker share 4 tags—the highest count—this reflects mechanical compatibility rather than power level. Their pairing enhances Flower Pot’s X3 multiplier flexibility, but jokers with fewer shared tags may actually create stronger combinations. This suggests shared tags indicate synergy potential rather than synergy quality.
Hub jokers enable multiple synergy paths: Three jokers (Flower Pot, Smeared Joker, and Blackboard) appear multiple times in the bar chart, indicating they have a ‘link’ within the ‘synergy’ system. This suggests focusing on these versatile jokers could enable flexible deck-building strategies, as players can pivot between different synergy combinations depending on what cards become available.
Accessibility of High-Synergy Pairs: Notably, all jokers in the top 5 pairings are common or uncommon rarity, making these synergies highly accessible to players. This accessibility factor significantly increases the practical value of these findings, as players can reliably build strategies around these combinations rather than relying on rare drops. This suggests the synergy pattern system identifies not just theoretical optimal pairings, but actionable strategic options for consistent game play.