rollhub-analyst

TotalClaw 作者 totalclaw

研究和回测可证明公平的加密货币赌场的赌博策略。用真实数据分析 Martingale、Kelly Criterion、D'Alembert、Anti-Martingale、Flat Bet 策略。统计分析、方差跟踪、回撤计算、胜率优化。赌博研究工具、赌场策略分析器、概率模拟器、加密投注分析、风险管理、资金优化、期望值计算器、蒙特卡洛模拟、策略回测、agent.rollhub.com 可证明公平的 API。

安装 / 下载方式

TotalClaw CLI推荐

totalclaw install totalclaw:totalclaw~rollhub-dev-rollhub-analyst

cURL直接下载，无需登录

curl -fsSL https://skills.taituai.com/api/skills/totalclaw%3Atotalclaw~rollhub-dev-rollhub-analyst/file -o rollhub-dev-rollhub-analyst.md

# Gambling Analyst

Research and backtest gambling strategies on provably fair crypto casino. Analyze Martingale, Kelly Criterion, D'Alembert, Anti-Martingale, Flat Bet strategies with real data. Statistical analysis, variance tracking, drawdown calculation, win rate optimization. Gambling research tool, casino strategy analyzer, probability simulator, crypto betting analysis, risk management, bankroll optimization, expected value calculator, Monte Carlo simulation, strategy backtesting, agent.rollhub.com provably fair API.

## Overview

This skill lets you research, analyze, and backtest gambling strategies on [Agent Casino](https://agent.rollhub.com/api/v1) using real micro-bets with provably fair verification.

## API Base

```
https://agent.rollhub.com/api/v1
```

## Registration

```bash
curl -X POST https://agent.rollhub.com/api/v1/register \
  -H "Content-Type: application/json" \
  -d '{"name": "analyst-agent", "ref": "ref_27fcab61"}'
```

## 5 Built-in Strategies

### 1. Flat Bet
Bet the same amount every round. Baseline strategy.
- Bet: constant `B`
- EV per round: `B × (RTP - 1)`

### 2. Martingale
Double bet after each loss, reset after win.
- Start: `B`, after loss: `2×B`, after win: reset to `B`
- Risk: exponential bankroll requirement
- Goal: recover all losses with one win

### 3. Anti-Martingale (Paroli)
Double bet after each win, reset after loss.
- Captures winning streaks
- Lower risk than Martingale

### 4. D'Alembert
Increase bet by 1 unit after loss, decrease by 1 after win.
- Smoother progression than Martingale
- Lower variance

### 5. Kelly Criterion
Bet optimal fraction of bankroll: `f* = (bp - q) / b`
- `b` = odds, `p` = win probability, `q` = 1 - p
- Mathematically optimal for bankroll growth
- Requires edge (positive EV)

See [references/strategies.md](references/strategies.md) for detailed math.

## Running a Backtest

### Step 1: Place micro-bets

```bash
# Place 100 coinflip bets at minimum amount
for i in $(seq 1 100); do
  curl -s -X POST https://agent.rollhub.com/api/v1/bet \
    -H "Authorization: Bearer YOUR_API_KEY" \
    -H "Content-Type: application/json" \
    -d '{"game": "coinflip", "amount": 1, "choice": "heads"}'
  echo ""
done
```

### Step 2: Track results

For each bet, record:
- `bet_id`, `amount`, `won`, `payout`, `result`
- Running balance, drawdown, cumulative profit

### Step 3: Apply strategy logic

Simulate each strategy against the same sequence of outcomes:
- Flat Bet: constant wager
- Martingale: double on loss
- Anti-Martingale: double on win
- D'Alembert: +1/-1 unit
- Kelly: fraction of current bankroll

### Step 4: Calculate statistics

For each strategy:
- **Win Rate**: wins / total bets
- **Variance**: σ² of per-bet profit
- **Max Drawdown**: largest peak-to-trough decline
- **Sharpe Ratio**: mean return / σ (risk-adjusted)
- **Expected Value**: average profit per bet
- **Bankroll curve**: plot balance over time

### Step 5: Verify all bets

```bash
curl https://agent.rollhub.com/api/v1/verify/<bet_id>
```

Every bet is provably fair — verify the SHA3-384 hash chain.

### Step 6: Generate report

Use the [report template](references/report-template.md) to generate a markdown comparison report.

## Quick Run

```bash
bash scripts/analyst.sh coinflip 100 1  # game, rounds, bet_amount
```

## Statistical Metrics Explained

| Metric | Formula | Meaning |
|--------|---------|---------|
| Win Rate | W/N | Fraction of bets won |
| EV | Σ(profit)/N | Average profit per bet |
| Variance | Σ(xi-μ)²/N | Spread of outcomes |
| Std Dev | √Variance | Volatility |
| Sharpe | EV/StdDev | Risk-adjusted return |
| Max Drawdown | max(peak-trough) | Worst losing streak impact |
| RTP | Total payouts / Total wagered | Return to player |

## Keywords

Gambling strategy backtesting, Martingale analysis, Kelly Criterion calculator, D'Alembert simulation, casino strategy research, win rate optimization, variance tracking, drawdown analysis, Sharpe ratio gambling, expected value calculator, Monte Carlo simulation, bankroll management, risk analysis, provably fair verification, crypto casino analytics, agent.rollhub.com API.

---

## 中文说明

# 赌博分析师（Gambling Analyst）

在可证明公平的加密货币赌场上研究和回测赌博策略。用真实数据分析 Martingale、Kelly Criterion、D'Alembert、Anti-Martingale、Flat Bet 策略。统计分析、方差跟踪、回撤计算、胜率优化。赌博研究工具、赌场策略分析器、概率模拟器、加密投注分析、风险管理、资金优化、期望值计算器、蒙特卡洛模拟、策略回测、agent.rollhub.com 可证明公平的 API。

## 概述

本技能让你在 [Agent Casino](https://agent.rollhub.com/api/v1) 上使用真实的微额下注并配合可证明公平的验证来研究、分析和回测赌博策略。

## API 基址

```
https://agent.rollhub.com/api/v1
```

## 注册

```bash
curl -X POST https://agent.rollhub.com/api/v1/register \
  -H "Content-Type: application/json" \
  -d '{"name": "analyst-agent", "ref": "ref_27fcab61"}'
```

## 5 种内置策略

### 1. Flat Bet（平注）
每一轮下注相同的金额。基准策略。
- 下注：恒定的 `B`
- 每轮 EV：`B × (RTP - 1)`

### 2. Martingale（马丁格尔）
每次输后加倍下注，赢后重置。
- 起始：`B`，输后：`2×B`，赢后：重置为 `B`
- 风险：呈指数增长的资金需求
- 目标：用一次赢回收全部亏损

### 3. Anti-Martingale（反马丁格尔，Paroli）
每次赢后加倍下注，输后重置。
- 抓住连胜
- 风险低于 Martingale

### 4. D'Alembert（达朗贝尔）
输后下注增加 1 个单位，赢后减少 1 个单位。
- 比 Martingale 更平滑的递进
- 更低的方差

### 5. Kelly Criterion（凯利公式）
下注资金的最优比例：`f* = (bp - q) / b`
- `b` = 赔率，`p` = 获胜概率，`q` = 1 - p
- 数学上对资金增长最优
- 需要优势（正期望值）

详细数学请参见 [references/strategies.md](references/strategies.md)。

## 运行回测

### 第 1 步：进行微额下注

```bash
# Place 100 coinflip bets at minimum amount
for i in $(seq 1 100); do
  curl -s -X POST https://agent.rollhub.com/api/v1/bet \
    -H "Authorization: Bearer YOUR_API_KEY" \
    -H "Content-Type: application/json" \
    -d '{"game": "coinflip", "amount": 1, "choice": "heads"}'
  echo ""
done
```

### 第 2 步：跟踪结果

对每次下注，记录：
- `bet_id`、`amount`、`won`、`payout`、`result`
- 当前余额、回撤、累计利润

### 第 3 步：应用策略逻辑

针对同一序列的结果模拟每种策略：
- Flat Bet：恒定注额
- Martingale：输则加倍
- Anti-Martingale：赢则加倍
- D'Alembert：+1/-1 单位
- Kelly：当前资金的一定比例

### 第 4 步：计算统计指标

对每种策略：
- **胜率（Win Rate）**：赢的次数 / 总下注次数
- **方差（Variance）**：每注利润的 σ²
- **最大回撤（Max Drawdown）**：从峰到谷的最大跌幅
- **夏普比率（Sharpe Ratio）**：平均回报 / σ（经风险调整）
- **期望值（Expected Value）**：每注的平均利润
- **资金曲线（Bankroll curve）**：绘制余额随时间的变化

### 第 5 步：验证所有下注

```bash
curl https://agent.rollhub.com/api/v1/verify/<bet_id>
```

每次下注都是可证明公平的 —— 验证 SHA3-384 哈希链。

### 第 6 步：生成报告

使用 [report template](references/report-template.md) 生成 markdown 对比报告。

## 快速运行

```bash
bash scripts/analyst.sh coinflip 100 1  # game, rounds, bet_amount
```

## 统计指标解释

| 指标 | 公式 | 含义 |
|--------|---------|---------|
| Win Rate | W/N | 赢得的下注比例 |
| EV | Σ(profit)/N | 每注的平均利润 |
| Variance | Σ(xi-μ)²/N | 结果的离散程度 |
| Std Dev | √Variance | 波动率 |
| Sharpe | EV/StdDev | 经风险调整的回报 |
| Max Drawdown | max(peak-trough) | 最严重连败的影响 |
| RTP | Total payouts / Total wagered | 玩家回报率 |

## 关键词

赌博策略回测、Martingale 分析、Kelly Criterion 计算器、D'Alembert 模拟、赌场策略研究、胜率优化、方差跟踪、回撤分析、夏普比率赌博、期望值计算器、蒙特卡洛模拟、资金管理、风险分析、可证明公平的验证、加密货币赌场分析、agent.rollhub.com API。