All Articles (10)
Psychology (1)
Behavior (1)
quant (9)

title: Vnpy tutorial 0x06 backtest description: Vnpy Quant Tutorial, Backtest keywords: Vnpy, quant tutorial, backtest date: 2024-09-05 category: "quant"

Warning: This page is translated by MACHINE, which may lead to POOR QUALITY or INCORRECT INFORMATION, please read with CAUTION!

After the data is connected according to [[vnpy_tutorial_0x02_futu]]]

Open the recovery system: Function -> CTABACKTESTER

  • Strategy Strategy: Hackstrategy
  • VTSYMBOL Local Code: Enter 00700.sehk
  • Start date Start time: 01/01/24
  • Fear rate: 0.000025
  • SIP PAGE transaction inventory: 0.2
  • SIZE contract multiplication 300
  • Price Tick price beating 0.2
  • CAPITAL Returga Back Testing funds: 1000000.0

Enter 0.000025 (0.25) in the handling rate editing box, 0.2 input of the transaction slippery point (that is, a slide cost of a unilateral transaction 1 jump), the contract multiplication is 300 (300 yuan each point), the price jump is also 0.2 (stock index futures futures futures (stock index futures futures The minimum price changes), we use 1 million for the recovery funds.

Click the "Start Test" button to pop up the parameter configuration dialog box:

  • Then click Start Backtesting

The Fast_window and Slow_window here are the parameter names we added to the parameters list before. Our back test engine will automatically start the entire process of the strategy back test:

Load data, data back, simulated matching, calculate daily profit and loss, statistical indicators,

finally draw a chart:

13:03:21	Initializing BacktesterEngine
13:03:21	Strategies imported
13:03:27	----------------------------------------
13:03:27	Start loading historical data
13:03:27	Loading progress: # [0%]
13:03:28	Loading progress: # [10%]
13:03:28	Loading progress: ## [20%]
13:03:29	Loading progress: ### [30%]
13:03:30	Loading progress: #### [40%]
13:03:31	Loading progress: ##### [50%]
13:03:32	Loading progress: ###### [59%]
13:03:33	Loading progress: ####### [69%]
13:03:34	Loading progress: ######## [79%]
13:03:35	Loading progress: ######### [89%]
13:03:36	Loading progress: ########## [99%]
13:03:36	Historical data loading completed, data count: 89852
13:03:36	Strategy initialization complete
13:03:36	Start replaying historical data
13:03:36	Backtesting progress: = [0%]
13:03:36	Backtesting progress: == [10%]
13:03:37	Backtesting progress: === [20%]
13:03:37	Backtesting progress: ==== [30%]
13:03:37	Backtesting progress: ===== [40%]
13:03:38	Backtesting progress: ====== [50%]
13:03:38	Backtesting progress: ======= [60%]
13:03:38	Backtesting progress: ======== [70%]
13:03:39	Backtesting progress: ========= [80%]
13:03:39	Backtesting progress: ========== [90%]
13:03:39	Backtesting progress: =========== [100%]
13:03:39	Historical backtest complete
13:03:39	Start calculating daily mark-to-market profit and loss
13:03:39	The daily mark-to-market profit and loss calculation is complete
13:03:39	Start calculating strategy statistical indicators
13:03:39	Strategy statistical indicators calculation completedcompleted```


 ![](https://xcy-1251434521.cos.ap-chengdu.myqcloud.com/picture/202409111309963.png?imageSlim) 


- 子图1：资金变化曲线，笔直向下说明稳定亏损
- 子图2：最大回撤曲线，越来越大说明策略亏损越来越多
- 子图3：每日盈亏统计，红绿分布平均，但绿色密度更大（亏损）
- 子图4：盈亏的概率分布图，尖峰在0轴左侧（中位数日期发生亏损）


中间的数据说明

| 功能                           | 数据                |
| ---------------------------- | ----------------- |
| 首个交易日                        | 2016-08-15        |
| 最后交易日                        | 2019-07-31        |
| 总交易日                         | 721               |
| 盈利交易日                        | 265               |
| 亏损交易日                        | 456               |
| 起始资金                         | 1,000,000.00      |
| 结束资金                         | -1,204,558.24     |
| 总收益率                         | -220.46%          |
| 年化收益                         | -73.38%           |
| Max Drawdown最大回撤             | -2,219,999.08     |
| Max Drawdown Duration百分比最大回撤 | -220.15%          |
| Toal Net P&L总盈亏              | -2,204,558.24     |
| Total Commission总手续费         | 519,338.24        |
| Total Slippage总滑点            | 1,154,700.00      |
| Total Turnover总成交额           | 20,773,529,520.00 |
| Total Trade Count 总成交笔数      | 19245             |
| 日均盈亏                         | -3,057.64         |
| 日均手续费                        | 720.30            |
| 日均滑点                         | 1,601.53          |

# 功能说明


* Backtesting Trades 回测交易
* Backtesting Orders 回测订单
* Daily Results 每日资金数据结果
* Candle Chart 蜡烛图
* Start Optimization 开始优化
* Optimization Results 优化结果
* Edit Strategy Code 编辑策略代码
* Reload Strategy Class 重新载入策略代码


# 策略自定义参数


在HackStrategies.py中，改数据

```python

# Definition parameter

Fast_window = 10
Slow_window = 20
+test_window = 30


# Definition variable
FAST_MA0 = 0.0
FAST_MA1 = 0.0
Slow_ma0 = 0.0
Slow_ma1 = 0.0
+test_ma0 = 0.0
+test_ma1 = 0.0

# Add parameter and variable name to the corresponding list
#parameters = [[["fast_window",,"slow_window"]
parameters = ["fast_window",,"slow_window",,"test_window"]
#variables = [[["fast_ma0",,"fast_ma1",,"slow_ma0",,"slow_ma1"]
variables = ["fast_ma0",,"fast_ma1",,"slow_ma0",,"slow_ma1",,"test_ma0",,"test_ma1"]

下面的Start Optimization 开始优化章节会提到。

Start Optimization 开始优化

点start Backtesting（开始回测）

我们想要看看

  • fast_window，从2到20（步进2），

  • slow_window，从20到100（步进10）

  • test_window，先不管。

  • step:增加步数。例：开始1，结束100，步数为2，那么1,3,5,7

  • Brutal force optimization暴力穷举优化：把所有可能都显示一次

  • Genetic algorithm optimization遗传算法优化

我们选择Brutal force optimization暴力穷举优化开始跑数据，等待结果出来

 

查看优化结果，点击Optimization Results 优化结果可以看见哪个参数的效果最好。

fast_window数据：30 slow_window数据：90

效果最好，亏得最少。只有-17%，本来亏损是64%，进步很大 2 本来

回测教程

Back to articles