Such values have no logic to qualify as trade guide/predict direction.
Anyways most of the methods we use have no logic so I tested this long back and the results are absolutely inconsistent.
Moreover the idea itself has a flaw. Because there are multiple strikes which are very liquid in current week and monthly expirys.
In PE's itself one strike can show "open=high and another strike can be open=low" which are opposing signals. So which one do we base on ?
Similarly in calls side different strikes may show different signals. Hence the method has a flaw which invalidates testing.
( we may not see these on one or 2 days data. But when we take it up for a test and observe numbers on number of days/ weeks of data we will see all these imperfections).