进入退出信号的盯市变化

Question

我正在寻找一种方法来计算股票市场头寸在进入和退出日期之间的每日变化。例如，如果我以951.84 的价格进入02/06/08，并以967.42 的价格在02/19/08 退出，那么基于Daily market price 的每日价格变化是多少？ 我想merge 两个数据框merge(df1,df2,all=T) 但同时考虑买卖日期并计算两者之间的变化。

>DailyMarketPrice

date        price
01/21/08    917.75
01/22/08    955.93
01/23/08    921.28
01/24/08    958.55
01/25/08    961.69
01/28/08    949.08
01/29/08    956.97
01/30/08    951.14
01/31/08    949.04
02/01/08    976.43
02/04/08    976.46
02/05/08    945.63
02/06/08    951.84
02/07/08    925.87
02/08/08    920.76
02/11/08    911.39
02/12/08    945.41
02/13/08    949.05
02/14/08    950.84
02/15/08    938.79
02/18/08    962.13
02/19/08    967.42


>signals

date    Cond
01/21/08    Buy
01/22/08    Sell
02/06/08    Buy
02/19/08    Sell

Answer 1

不幸的是，我不熟悉足够的软件包来为您提供可以处理此类事情的东西。但是，由于这里只有一种资产在交易，因此您自己编写代码似乎很简单。这是您正在寻找的那种输出吗？

    date  price inv.change inventory pnl.change   pnl
01/21/08 917.75          1         0       0.00  0.00
01/22/08 955.93         -1         1      38.18 38.18
01/23/08 921.28          0         0       0.00 38.18
01/24/08 958.55          0         0       0.00 38.18
01/25/08 961.69          0         0       0.00 38.18
01/28/08 949.08          0         0       0.00 38.18
01/29/08 956.97          0         0       0.00 38.18
01/30/08 951.14          0         0       0.00 38.18
01/31/08 949.04          0         0       0.00 38.18
02/01/08 976.43          0         0       0.00 38.18
02/04/08 976.46          0         0       0.00 38.18
02/05/08 945.63          0         0       0.00 38.18
02/06/08 951.84          1         0       0.00 38.18
02/07/08 925.87          0         1     -25.97 12.21
02/08/08 920.76          0         1      -5.11  7.10
02/11/08 911.39          0         1      -9.37 -2.27
02/12/08 945.41          0         1      34.02 31.75
02/13/08 949.05          0         1       3.64 35.39
02/14/08 950.84          0         1       1.79 37.18
02/15/08 938.79          0         1     -12.05 25.13
02/18/08 962.13          0         1      23.34 48.47
02/19/08 967.42         -1         1       5.29 53.76

如果是这样，这是我用来合并 DailyMarketPrice 和 signals 的代码：

DailyMarketPrice <- do.call(rbind, list(
  data.frame(date = "01/21/08", price = 917.75),
  data.frame(date = "01/22/08", price = 955.93),
  data.frame(date = "01/23/08", price = 921.28),
  data.frame(date = "01/24/08", price = 958.55),
  data.frame(date = "01/25/08", price = 961.69),
  data.frame(date = "01/28/08", price = 949.08),
  data.frame(date = "01/29/08", price = 956.97),
  data.frame(date = "01/30/08", price = 951.14),
  data.frame(date = "01/31/08", price = 949.04),
  data.frame(date = "02/01/08", price = 976.43),
  data.frame(date = "02/04/08", price = 976.46),
  data.frame(date = "02/05/08", price = 945.63),
  data.frame(date = "02/06/08", price = 951.84),
  data.frame(date = "02/07/08", price = 925.87),
  data.frame(date = "02/08/08", price = 920.76),
  data.frame(date = "02/11/08", price = 911.39),
  data.frame(date = "02/12/08", price = 945.41),
  data.frame(date = "02/13/08", price = 949.05),
  data.frame(date = "02/14/08", price = 950.84),
  data.frame(date = "02/15/08", price = 938.79),
  data.frame(date = "02/18/08", price = 962.13),
  data.frame(date = "02/19/08", price = 967.42)
))

signals <- setNames(as.data.frame(matrix(c(
  "01/21/08", "Buy",
  "01/22/08", "Sell",
  "02/06/08", "Buy",
  "02/19/08", "Sell"
), ncol = 2, byrow = TRUE)), c("date", "Cond"))

# Set `inv.change` to `+1` on "Buy" days, `-1` on "Sell" days and `0` on every
#  day without a transaction.
merged <- cbind(DailyMarketPrice, inv.change = c(+1, -1, 0)[match(signals$Cond[match(DailyMarketPrice$date, signals$date)], c("Buy", "Sell", NA))])
# Set `inventory` to 0 on "Buy" days, 1 on "Sell" days, and 1 on every day
#  between "Buy" and "Sell" days.
merged$inventory <- head(diffinv(merged$inv.change), -1)
# Mark-to-market the inventory positions by calculating change in prices during
#  holding periods.
merged$pnl.change <- merged$inventory * c(0, diff(merged$price))
# Record the cumulative P&L.
merged$pnl <- tail(diffinv(merged$pnl.change), -1)
print(merged)