Because I've been singing the praises of the Profit vs Risk (PvR) metric I thought it would be useful to post a particular type of example.

This illustrates a point in the main chart that everyone developing algorithms should understand in order to be able to make sound decisions regarding your iterative changes, as to which ones are positive and which ones are negative.

The first column is a cost basis multiplier parameter.
When that parameter is increased by 1/1000th, the Quantopian return value goes up by only 9/100ths of 1%, while ...
PvR increases by a whopping 14 whole percentage points (from 88% to 102%).
That increase in output could easily have been missed.
An additional value in PvR is that one no longer needs to worry about the amount of cash used or leverage until ready.

param  
1.228  QRet 75.42          PvR 88.62               CshLw 2664 MxLv 0.79 RskHi 8510  
1.229  QRet 75.51 up .09   PvR 102.95 up 14 points CshLw 2664 MxLv 0.79 RskHi 7335  
1.23   QRet 75.51          PvR 102.95              CshLw 2664 MxLv 0.79 RskHi 7335  
1.24   QRet 75.51          PvR 102.95              CshLw 2664 MxLv 0.79 RskHi 7335  
1.25   QRet 75.51          PvR 102.95              CshLw 2664 MxLv 0.79 RskHi 7335  
1.26   QRet 68.63 -7       PvR 89.83  -13          CshLw 2359 MxLv 0.83 RskHi 7640  
1.27   QRet 58.00 -10      PvR 71.95  -18          CshLw 1938 MxLv 0.87 RskHi 8061  
1.28   QRet 65.64 +7       PvR 85.91  +14          CshLw 2359 MxLv 0.84 RskHi 7640  
1.29   QRet 65.64          PvR 85.91               CshLw 2359 MxLv 0.84 RskHi 7640  

There are also times when QRet and PvR move in opposite directions.
This backtest beats the benchmark by over 50 percentage points in terms of profit per dollar activated, it's just that it didn't use all of the starting capital.

My philosophy: The better everyone else does the better Quantopian does and the better I'm likely to be able to do then. So do well, try PvR.