When they calculate the statistical back test performance, by default, do they base the trade gain/loss on how many shares/contracts you could have traded per fraction of capital allocated (rounded to nearest integer shares), or simply the gain/loss between the securities price change during the trade? I'm assuming the former, but just checking. TIA.
to make it a little more clear. Suppose a stock went from 10 to 12.5, the relative gain would be 25%/trade here. However, if you had, say 1,000 and bet a fraction of 2.5%/trade, you would bet the maximum fractional shares near $25, or 2 shares. Thus your gain would be 2shares*2.5 or 5 dollars for a $20 investment (fractional trade). However, relative to the 1,000 principle, the gain was not 25%/trade, rather it was .5%/trade. Also, assuming zero commission here. So when reading trade station results from various books, etc.. I am just curious if they calculate all the gain/trade based metrics on the former approach, or the latter. Just asking, since the first approach could be misleading, and many books show trade station results in their summaries. And while your at it, same question applies for any platform default you are familiar with.