TidyTuesday: Risk Valuation Matrices

TidyTuesday

Data Viz

Finance

Risk

Six risk matrices - R-multiple vs winrate, drawdown vs recovery, expectancy, losing streaks, exponential growth, consecutive losses - with ggplot2

Author

chokotto

Published

February 11, 2026

Overview

This TidyTuesday builds the same six risk valuation matrices as the companion MakeoverMonday, using ggplot2: R-multiple vs required winrate, drawdown vs recovery rate, expectancy, losing streaks and capital loss, exponential growth, and consecutive-loss probability. Metrics are read from data/risk_metrics.parquet (generated by prepare_data.py).

Data

Show code

library(tidyverse)
library(scales)
library(arrow)
library(patchwork)
library(glue)

Show code

data_dir <- file.path(getwd(), "data")
risk_path <- file.path(data_dir, "risk_metrics.parquet")

if (file.exists(risk_path)) {
  risk <- read_parquet(risk_path) %>% slice(1)
  win_rate   <- as.numeric(risk$win_rate_pct)
  r_multiple <- as.numeric(risk$r_multiple)
  max_dd     <- as.numeric(risk$max_drawdown_pct)
  max_streak <- as.integer(risk$max_consecutive_losing_days)
  account_jpy <- as.numeric(risk$account_size_jpy)
  pos_pct    <- as.numeric(risk$position_size_pct)
  cat("Loaded risk_metrics: WR =", win_rate, "%, R =", r_multiple, ", max DD =", max_dd, "%, max streak =", max_streak, "\n")
} else {
  win_rate <- 55; r_multiple <- 1.5; max_dd <- 10; max_streak <- 3
  account_jpy <- 1e7; pos_pct <- 2
  cat("Using default risk metrics (run prepare_data.py to use your data)\n")
}

Loaded risk_metrics: WR = 55 %, R = 1.5 , max DD = 0 %, max streak = 0

1. R-Multiple and Required Winrate

Show code

r_vals <- seq(0.5, 3.5, by = 0.5)
req_wr <- 100 / (1 + r_vals)
tbl1 <- tibble(r_multiple = r_vals, required_winrate_pct = req_wr)

p1 <- ggplot(tbl1, aes(x = factor(r_multiple), y = required_winrate_pct)) +
  geom_col(fill = "#3b82f6") +
  geom_vline(xintercept = which.min(abs(r_vals - r_multiple)), linetype = "dashed", color = "red", linewidth = 0.8) +
  annotate("text", x = Inf, y = Inf, label = glue("Formula: Required WR = 1 / (1 + R)\nYour R ≈ {round(r_multiple, 2)}"), hjust = 1.1, vjust = 1.2, size = 3) +
  labs(title = "R-Multiple vs Required Winrate", x = "R-Multiple", y = "Required Winrate (%)") +
  theme_minimal(base_size = 12)
p1

2. Drawdown and Recovery Rate

Show code

dd_pcts <- c(5, 10, 20, 30, 40, 50, 60, 70, 80, 90)
recovery <- 100 * dd_pcts / (100 - dd_pcts)
tbl2 <- tibble(drawdown_pct = dd_pcts, recovery_pct = recovery)

p2 <- ggplot(tbl2, aes(x = factor(drawdown_pct), y = recovery_pct)) +
  geom_col(fill = "#10b981") +
  geom_vline(xintercept = which.min(abs(dd_pcts - max_dd)), linetype = "dashed", color = "red", linewidth = 0.8) +
  annotate("text", x = Inf, y = Inf, label = glue("Recovery % = 100 × DD / (1 − DD)\nYour max DD ≈ {round(max_dd, 0)}%"), hjust = 1.1, vjust = 1.2, size = 3) +
  labs(title = "Drawdown % vs Recovery Rate %", x = "Drawdown (%)", y = "Recovery needed (%)") +
  theme_minimal(base_size = 12)
p2

3. Expectancy

Show code

wr_dec <- win_rate / 100
exp_pct <- (wr_dec * pos_pct * r_multiple) - ((1 - wr_dec) * pos_pct)

p3 <- ggplot() +
  annotate("text", x = 0.5, y = 0.5, label = glue(
    "Expectancy = (Winrate × Pos.Size × R) − ((1 − Winrate) × Pos.Size)\n",
    "With your data: Winrate = {win_rate}%, Pos = {pos_pct}%, R = {round(r_multiple, 2)}  →  Expectancy = {round(exp_pct, 2)}% per trade"
  ), size = 4, hjust = 0.5, vjust = 0.5) +
  scale_x_continuous(limits = c(0, 1)) +
  scale_y_continuous(limits = c(0, 1)) +
  labs(title = "Trade Expectancy (your parameters)") +
  theme_void()
p3

4. Losing Streaks and Loss of Capital

Show code

n_losses <- 1:10
risks <- c(0.01, 0.03, 0.05)
tbl4 <- expand_grid(n = n_losses, risk_pct = risks) %>%
  mutate(loss_pct = 100 * (1 - (1 - risk_pct)^n), risk_lab = paste0(risk_pct * 100, "% risk"))

p4 <- ggplot(tbl4, aes(x = n, y = loss_pct, color = risk_lab)) +
  geom_line(linewidth = 1) +
  geom_point(size = 2) +
  geom_vline(xintercept = max_streak, linetype = "dotted", color = "gray40", linewidth = 0.8) +
  annotate("text", x = max_streak, y = max(tbl4$loss_pct) * 0.95, label = glue("Your max streak: {max_streak}"), hjust = -0.1, size = 3) +
  labs(title = "Losing Streaks: Capital Loss % after n Consecutive Losses", x = "Consecutive losses (n)", y = "Loss %", color = NULL) +
  theme_minimal(base_size = 12) +
  theme(legend.position = "top")
p4

5. Exponential Growth (Gain per Trade)

Show code

E <- exp_pct / 100
account_0 <- if (account_jpy > 0) account_jpy else 1e7
account_m <- round(account_0 / 1e6, 1)
n_trades <- c(1, 10, 50, 100, 200, 500)
gain_per <- account_0 * ((1 + E)^n_trades - 1) / n_trades
tbl5 <- tibble(n_trades = factor(n_trades), gain_million_jpy = gain_per / 1e6)

p5 <- ggplot(tbl5, aes(x = n_trades, y = gain_million_jpy)) +
  geom_col(fill = "#8b5cf6") +
  labs(
    title = glue("Gain per Trade (million JPY) — Account ¥{account_m}M, Expectancy {round(exp_pct, 2)}%"),
    x = "Number of trades", y = "Gain per trade (million JPY)"
  ) +
  theme_minimal(base_size = 12)
p5

6. Consecutive Losses (Probability Heatmap)

Show code

winrates <- seq(30, 70, by = 5)
k_streaks <- 1:5
tbl6 <- expand_grid(winrate_pct = winrates, k = k_streaks) %>%
  mutate(prob_pct = 100 * (1 - winrate_pct/100)^k)

p6 <- ggplot(tbl6, aes(x = factor(k), y = factor(winrate_pct), fill = prob_pct)) +
  geom_tile(color = "white", linewidth = 0.3) +
  geom_text(aes(label = sprintf("%.1f%%", prob_pct)), size = 2.8) +
  scale_fill_gradient(low = "white", high = "darkred", name = "P(k losses) %") +
  geom_hline(yintercept = which.min(abs(winrates - win_rate)), linetype = "dashed", color = "blue", linewidth = 0.8) +
  labs(
    title = "P(k consecutive losses) by Winrate — your winrate highlighted",
    x = "Consecutive losses (k)", y = "Winrate (%)"
  ) +
  theme_minimal(base_size = 11) +
  theme(panel.grid = element_blank())
p6

Key Findings

R and winrate: Higher R lowers the winrate needed to break even.
Drawdown: Recovery needed grows non-linearly (e.g. 50% DD → 100% gain to recover).
Expectancy: Combines winrate, position size, and R into one per-trade edge.
Streaks: A few consecutive losses at 3–5% risk can erase a large share of capital.
Growth: Positive expectancy compounds; gain per trade rises with trade count.
Consecutive losses: Even at 60% winrate, 3+ consecutive losses are non-negligible.

This post is part of the TidyTuesday weekly data visualization project.

Disclaimer

This analysis is for educational and practice purposes only. Risk metrics and interpretations are based on the provided dataset and do not constitute investment advice.

--- title: "TidyTuesday: Risk Valuation Matrices" description: "Six risk matrices - R-multiple vs winrate, drawdown vs recovery, expectancy, losing streaks, exponential growth, consecutive losses - with ggplot2" date: "2026-02-11" x-posted: false author: "chokotto" categories: ["TidyTuesday", "Data Viz", "R", "Finance", "Risk"] image: "thumbnail.svg" engine: knitr code-fold: true code-tools: true code-summary: "Show code" twitter-card: card-type: summary_large_image image: "thumbnail.png" title: "TidyTuesday: Risk Valuation Matrices" description: "R-multiple, drawdown, expectancy, streaks, growth, consecutive losses" --- ## Overview This TidyTuesday builds the same **six risk valuation matrices** as the companion MakeoverMonday, using **ggplot2**: R-multiple vs required winrate, drawdown vs recovery rate, expectancy, losing streaks and capital loss, exponential growth, and consecutive-loss probability. Metrics are read from `data/risk_metrics.parquet` (generated by `prepare_data.py`). ## Data ```{r} #| label: load-packages #| message: false #| warning: false library(tidyverse) library(scales) library(arrow) library(patchwork) library(glue) ``` ```{r} #| label: load-risk-metrics #| message: false data_dir <- file.path(getwd(), "data") risk_path <- file.path(data_dir, "risk_metrics.parquet") if (file.exists(risk_path)) { risk <- read_parquet(risk_path) %>% slice(1) win_rate <- as.numeric(risk$win_rate_pct) r_multiple <- as.numeric(risk$r_multiple) max_dd <- as.numeric(risk$max_drawdown_pct) max_streak <- as.integer(risk$max_consecutive_losing_days) account_jpy <- as.numeric(risk$account_size_jpy) pos_pct <- as.numeric(risk$position_size_pct) cat("Loaded risk_metrics: WR =", win_rate, "%, R =", r_multiple, ", max DD =", max_dd, "%, max streak =", max_streak, "\n") } else { win_rate <- 55; r_multiple <- 1.5; max_dd <- 10; max_streak <- 3 account_jpy <- 1e7; pos_pct <- 2 cat("Using default risk metrics (run prepare_data.py to use your data)\n") } ``` ## 1. R-Multiple and Required Winrate ```{r} #| label: viz-r-winrate #| fig-width: 8 #| fig-height: 4 r_vals <- seq(0.5, 3.5, by = 0.5) req_wr <- 100 / (1 + r_vals) tbl1 <- tibble(r_multiple = r_vals, required_winrate_pct = req_wr) p1 <- ggplot(tbl1, aes(x = factor(r_multiple), y = required_winrate_pct)) + geom_col(fill = "#3b82f6") + geom_vline(xintercept = which.min(abs(r_vals - r_multiple)), linetype = "dashed", color = "red", linewidth = 0.8) + annotate("text", x = Inf, y = Inf, label = glue("Formula: Required WR = 1 / (1 + R)\nYour R ≈ {round(r_multiple, 2)}"), hjust = 1.1, vjust = 1.2, size = 3) + labs(title = "R-Multiple vs Required Winrate", x = "R-Multiple", y = "Required Winrate (%)") + theme_minimal(base_size = 12) p1 ``` ## 2. Drawdown and Recovery Rate ```{r} #| label: viz-drawdown-recovery #| fig-width: 8 #| fig-height: 4 dd_pcts <- c(5, 10, 20, 30, 40, 50, 60, 70, 80, 90) recovery <- 100 * dd_pcts / (100 - dd_pcts) tbl2 <- tibble(drawdown_pct = dd_pcts, recovery_pct = recovery) p2 <- ggplot(tbl2, aes(x = factor(drawdown_pct), y = recovery_pct)) + geom_col(fill = "#10b981") + geom_vline(xintercept = which.min(abs(dd_pcts - max_dd)), linetype = "dashed", color = "red", linewidth = 0.8) + annotate("text", x = Inf, y = Inf, label = glue("Recovery % = 100 × DD / (1 − DD)\nYour max DD ≈ {round(max_dd, 0)}%"), hjust = 1.1, vjust = 1.2, size = 3) + labs(title = "Drawdown % vs Recovery Rate %", x = "Drawdown (%)", y = "Recovery needed (%)") + theme_minimal(base_size = 12) p2 ``` ## 3. Expectancy ```{r} #| label: viz-expectancy #| fig-width: 8 #| fig-height: 2.5 wr_dec <- win_rate / 100 exp_pct <- (wr_dec * pos_pct * r_multiple) - ((1 - wr_dec) * pos_pct) p3 <- ggplot() + annotate("text", x = 0.5, y = 0.5, label = glue( "Expectancy = (Winrate × Pos.Size × R) − ((1 − Winrate) × Pos.Size)\n", "With your data: Winrate = {win_rate}%, Pos = {pos_pct}%, R = {round(r_multiple, 2)} → Expectancy = {round(exp_pct, 2)}% per trade" ), size = 4, hjust = 0.5, vjust = 0.5) + scale_x_continuous(limits = c(0, 1)) + scale_y_continuous(limits = c(0, 1)) + labs(title = "Trade Expectancy (your parameters)") + theme_void() p3 ``` ## 4. Losing Streaks and Loss of Capital ```{r} #| label: viz-streaks #| fig-width: 8 #| fig-height: 4 n_losses <- 1:10 risks <- c(0.01, 0.03, 0.05) tbl4 <- expand_grid(n = n_losses, risk_pct = risks) %>% mutate(loss_pct = 100 * (1 - (1 - risk_pct)^n), risk_lab = paste0(risk_pct * 100, "% risk")) p4 <- ggplot(tbl4, aes(x = n, y = loss_pct, color = risk_lab)) + geom_line(linewidth = 1) + geom_point(size = 2) + geom_vline(xintercept = max_streak, linetype = "dotted", color = "gray40", linewidth = 0.8) + annotate("text", x = max_streak, y = max(tbl4$loss_pct) * 0.95, label = glue("Your max streak: {max_streak}"), hjust = -0.1, size = 3) + labs(title = "Losing Streaks: Capital Loss % after n Consecutive Losses", x = "Consecutive losses (n)", y = "Loss %", color = NULL) + theme_minimal(base_size = 12) + theme(legend.position = "top") p4 ``` ## 5. Exponential Growth (Gain per Trade) ```{r} #| label: viz-growth #| fig-width: 8 #| fig-height: 4 E <- exp_pct / 100 account_0 <- if (account_jpy > 0) account_jpy else 1e7 account_m <- round(account_0 / 1e6, 1) n_trades <- c(1, 10, 50, 100, 200, 500) gain_per <- account_0 * ((1 + E)^n_trades - 1) / n_trades tbl5 <- tibble(n_trades = factor(n_trades), gain_million_jpy = gain_per / 1e6) p5 <- ggplot(tbl5, aes(x = n_trades, y = gain_million_jpy)) + geom_col(fill = "#8b5cf6") + labs( title = glue("Gain per Trade (million JPY) — Account ¥{account_m}M, Expectancy {round(exp_pct, 2)}%"), x = "Number of trades", y = "Gain per trade (million JPY)" ) + theme_minimal(base_size = 12) p5 ``` ## 6. Consecutive Losses (Probability Heatmap) ```{r} #| label: viz-consecutive #| fig-width: 7 #| fig-height: 5 winrates <- seq(30, 70, by = 5) k_streaks <- 1:5 tbl6 <- expand_grid(winrate_pct = winrates, k = k_streaks) %>% mutate(prob_pct = 100 * (1 - winrate_pct/100)^k) p6 <- ggplot(tbl6, aes(x = factor(k), y = factor(winrate_pct), fill = prob_pct)) + geom_tile(color = "white", linewidth = 0.3) + geom_text(aes(label = sprintf("%.1f%%", prob_pct)), size = 2.8) + scale_fill_gradient(low = "white", high = "darkred", name = "P(k losses) %") + geom_hline(yintercept = which.min(abs(winrates - win_rate)), linetype = "dashed", color = "blue", linewidth = 0.8) + labs( title = "P(k consecutive losses) by Winrate — your winrate highlighted", x = "Consecutive losses (k)", y = "Winrate (%)" ) + theme_minimal(base_size = 11) + theme(panel.grid = element_blank()) p6 ``` ## Key Findings 1. **R and winrate**: Higher R lowers the winrate needed to break even. 2. **Drawdown**: Recovery needed grows non-linearly (e.g. 50% DD → 100% gain to recover). 3. **Expectancy**: Combines winrate, position size, and R into one per-trade edge. 4. **Streaks**: A few consecutive losses at 3–5% risk can erase a large share of capital. 5. **Growth**: Positive expectancy compounds; gain per trade rises with trade count. 6. **Consecutive losses**: Even at 60% winrate, 3+ consecutive losses are non-negligible. *** _This post is part of the [TidyTuesday](https://github.com/rfordatascience/tidytuesday) weekly data visualization project._ :::{.callout-caution collapse="false" appearance="minimal" icon="false"} ## Disclaimer ::: {style="font-size: 0.85em; color: #64748b; line-height: 1.6;"} This analysis is for educational and practice purposes only. Risk metrics and interpretations are based on the provided dataset and do not constitute investment advice. ::: :::