Analysis Fields
Algorithm Science
Trading Strategy Discovery
The Analysis tab copies the algorithm spotilighted in the Control Panel from the Strategy tab for more detailed analysis and graphing. We do the analysis for both the Long and Short sides of the algorithm.
Algorithm Long/Algorithm Short
In these columns we show the return and drawdown for the long and short side of the algorithm. The results are displayed in the performance table and on the Growth Graph on the Results tab.
Days in mkt
If a position was held for a partial day, it is counted as one full day, unless both a long and short position was held on a particular day, in which case 0.5 days is assigned to each. This can lead to different percentages for "Days in mkt" if you switch between SAR and Long/Short styles. The figure is an approximation and tends to total higher than the actual time positions were held.
Time in mkt.
Time in mkt is a estimate of the actual time that Long and Short positions are held for. It will total less than the "Days in mkt" figure above because partial days are only assigned a fraction of a day, not a whole day. If a buy or sell occurs, it counts as Long for a half day. If a short sell or cover occurs it counts as Short for half a day. If a C/B and S/Sh cycle occurs, since we don't know if a buy/sell cycle or a cover/short cycle was executed (it depends on signal arrival order), we assign one quarter day to each Long and Short.
Optionally, you can display Time in mkt in the performance table and use it in the FOM and efficiency calculations, but the buy/sell point scans will always display Days in mkt.
Sliding Returns
The compounded return for the last N trading days, where N is the number at the top of the column and can be set by the user. This data is used in the longievity graph.
Profit and Loss
We generally use compounded returns in all the analyses, where profits/losses are re-invested. If you are interested in simple returns, they are shown here. The simple return would be the profit/loss realized if you always invested the exact same amount of cash for every transaction. Typically simple returns produce less profit and more loss than compounded returns (in fact simple losses can exceed 100%, compounded losses can't). As with compounded returns, transaction fees are not included. There are three columns because occasionally you will run into the situation where a profit or loss cannot be attributed to Long or Short position. Specifically, if reset is active and the reset state is Out, then a C/B and S/Sh cycle can occur. However whether the algorithm went long or short will depend on the order of arrival of the signals. On these days you will see "C/B" in the "Buy/Cov Action" column and "S/Sh" in the "Sell/Sht Action" column.
At the bottom of the column you will also see how many of the trades were good (i.e profitable) and how many were bad. Also, we give the average P/L per trade--the simple return divided by the number of postions held.
Algorithm Long/Algorithm Short
In these columns we show the return and drawdown for the long and short side of the algorithm. The results are displayed in the performance table and on the Growth Graph on the Results tab.
Days in mkt
If a position was held for a partial day, it is counted as one full day, unless both a long and short position was held on a particular day, in which case 0.5 days is assigned to each. This can lead to different percentages for "Days in mkt" if you switch between SAR and Long/Short styles. The figure is an approximation and tends to total higher than the actual time positions were held.
Time in mkt.
Time in mkt is a estimate of the actual time that Long and Short positions are held for. It will total less than the "Days in mkt" figure above because partial days are only assigned a fraction of a day, not a whole day. If a buy or sell occurs, it counts as Long for a half day. If a short sell or cover occurs it counts as Short for half a day. If a C/B and S/Sh cycle occurs, since we don't know if a buy/sell cycle or a cover/short cycle was executed (it depends on signal arrival order), we assign one quarter day to each Long and Short.
Optionally, you can display Time in mkt in the performance table and use it in the FOM and efficiency calculations, but the buy/sell point scans will always display Days in mkt.
Sliding Returns
The compounded return for the last N trading days, where N is the number at the top of the column and can be set by the user. This data is used in the longievity graph.
Profit and Loss
We generally use compounded returns in all the analyses, where profits/losses are re-invested. If you are interested in simple returns, they are shown here. The simple return would be the profit/loss realized if you always invested the exact same amount of cash for every transaction. Typically simple returns produce less profit and more loss than compounded returns (in fact simple losses can exceed 100%, compounded losses can't). As with compounded returns, transaction fees are not included. There are three columns because occasionally you will run into the situation where a profit or loss cannot be attributed to Long or Short position. Specifically, if reset is active and the reset state is Out, then a C/B and S/Sh cycle can occur. However whether the algorithm went long or short will depend on the order of arrival of the signals. On these days you will see "C/B" in the "Buy/Cov Action" column and "S/Sh" in the "Sell/Sht Action" column.
At the bottom of the column you will also see how many of the trades were good (i.e profitable) and how many were bad. Also, we give the average P/L per trade--the simple return divided by the number of postions held.
