在blotter中编程策略时，我遇到了交易后的End.Eq与手动计算的预期结果不匹配的问题。所以我写了一些简单的R代码来更好地理解吸墨纸是如何工作的。这里的结果也与我预期的不同。
也许有一个更有洞察力的人可以帮我解释一下。
# R version 3.2.1 (2015-06-18)
# Platform: x86_64-pc-linux-gnu (64-bit)
# blotter: 0.9.1666
require(blotter)
Sys.setenv(TZ="UTC")
rm(list =ls(envir=.blotter), envir=.blotter)
initDate <- '2008-01-01'
initEq <- 1000
currency("USD")
stock("MDY", currency = "USD", multiplier = 1)
getSymbols("MDY", from='2008-01-01', to='2008-06-30', index.class="POSIXct", adjust=T)
b.strategy <- "b.test"
initPortf(name = b.strategy, symbols = "MDY", initDate = initDate)
initAcct(name = b.strategy, portfolios = b.strategy, initDate = initDate, initEq = initEq)
addTxn(b.strategy, Symbol = "MDY", TxnDate = "2008-04-30",
TxnPrice = as.numeric(Cl(MDY["2008-04-30"])), TxnQty = 1, TxnFees = 0)
updatePortf(b.strategy, Dates = "2008-04-30")
updateAcct(b.strategy, Dates = "2008-04-30")
updateEndEq(b.strategy, Dates = "2008-04-30")
addTxn(b.strategy, Symbol = "MDY", TxnDate = "2008-05-30",
TxnPrice = as.numeric(Cl(MDY["2008-05-30"])), TxnQty = -1, TxnFees = 0)
updatePortf(b.strategy, Dates = "2008-05-30")
updateAcct(b.strategy, Dates = "2008-05-30")
updateEndEq(b.strategy, Dates = "2008-05-30")
perTradeStats(b.strategy, "MDY")
tradeStats(b.strategy)
getAccount(b.strategy)$summary
上面的代码用1000美元的现金初始化一个投资组合。我以"2008-04-30"的收盘价买了1个"MDY"。然后我在一个月后以"2008-05-30"的收盘价平仓。
[1] "2008-04-30 00:00:00 MDY 1 @ 151.920844224256"
[1] "2008-05-30 00:00:00 MDY -1 @ 160.178061444191"
由于投资组合不再包含未平仓头寸，我预计End.Eq的收益为1000+8.257217=1008.257217，但由于Unrealized.PL = -6.72，我看到的结果要低得多。
但这是从哪里来的？所有仓位都已平仓，所以我预计Unrealized.PL = 0。尽管交易似乎仍然给了我Net.Trading.PL = 8.257217的预期结果。
> perTradeStats(b.strategy, "MDY")
Start End Init.Pos Max.Pos Num.Txns Max.Notional.Cost Net.Trading.PL MAE MFE
1 2008-04-30 2008-05-30 1 1 2 151.9208 8.257217 0 8.257217
Pct.Net.Trading.PL Pct.MAE Pct.MFE tick.Net.Trading.PL tick.MAE tick.MFE
1 0.0543521 0 0.0543521 825.7217 0 825.7217
查看投资组合账户，我发现Unrealized.PL = -6.72，我无法解释，但这显然也导致了意外的End.Eq结果。
> getAccount(b.strategy)$summary
Additions Withdrawals Realized.PL Unrealized.PL Interest Gross.Trading.PL Txn.Fees Net.Trading.PL
2008-01-01 0 0 0.000000 0.00 0 0.000000 0 0.000000
2008-04-30 0 0 0.000000 0.00 0 0.000000 0 0.000000
2008-05-30 0 0 8.257217 -6.72 0 1.535756 0 1.535756
Advisory.Fees Net.Performance End.Eq
2008-01-01 0 0.000000 1000.000
2008-04-30 0 0.000000 1000.000
2008-05-30 0 1.535756 1001.536
总之，我的两个问题是：
- 寻求解释Unrealized.PL是如何产生的
- 如果可能的话，我该如何避免
首先，您应该只在需要标记书本时调用update*函数。如果你实际上不需要使用它们计算的一些值，那么在每次交易后调用它们是低效的。
其次，您使用单个日期调用update*函数，这(通常(没有意义。这导致这本书只标记了一个日期，这并没有特别的帮助。使用"2008-04-30/2008-05-30"这样的日期范围会更有意义。
非零的Unrealized.PL是因为您的平仓交易发生在2008-05-30的端，这意味着您有未实现的p&L代表那一天。
下面代码的修改版本应该会给出您期望的结果。
require(blotter)
Sys.setenv(TZ="UTC")
rm(list =ls(envir=.blotter), envir=.blotter)
initDate <- '2008-01-01'
initEq <- 1000
currency("USD")
stock("MDY", currency = "USD", multiplier = 1)
getSymbols("MDY", from='2008-01-01', to='2008-06-30', index.class="POSIXct", adjust=T)
b.strategy <- "b.test"
initPortf(name = b.strategy, symbols = "MDY", initDate = initDate)
initAcct(name = b.strategy, portfolios = b.strategy, initDate = initDate, initEq = initEq)
addTxn(b.strategy, Symbol = "MDY", TxnDate = "2008-04-30",
TxnPrice = as.numeric(Cl(MDY["2008-04-30"])), TxnQty = 1, TxnFees = 0)
addTxn(b.strategy, Symbol = "MDY", TxnDate = "2008-05-30",
TxnPrice = as.numeric(Cl(MDY["2008-05-30"])), TxnQty = -1, TxnFees = 0)
updatePortf(b.strategy, Dates = "2008-04-30/2008-05-30")
updateAcct(b.strategy, Dates = "2008-04-30/2008-05-30")
updateEndEq(b.strategy, Dates = "2008-04-30/2008-05-30")
perTradeStats(b.strategy, "MDY")
tradeStats(b.strategy)
getAccount(b.strategy)$summary
getPortfolio(b.strategy)$summary