ZAU CTLR INVOLVED IN AN OPERROR; QUESTIONS USABILITY OF CONFLICT PROBE FEATURE OF CONFLICT ALERT.

Narrative

ACFT #1 WAS ISSUED A CLRNC TO DSND; AFTER TAKING INTO ACCOUNT ACFT #2. IT WAS BELIEVED THAT 5 MI LATERAL SEPARATION WOULD BE MAINTAINED AND THAT WITH AN EYE TOWARDS MONITORING THE SIT; IT WOULD NOT BE A PROB. THERE WAS; A SHORT TIME LATER; CONSIDERABLE COORD WITH THE SECTOR BORDERING TO THE S. RETURNING TO THE SIT PREVIOUSLY DESCRIBED; A TURN TO THE R WAS ISSUED TO ACFT #1 TO 'PAD' THE SEPARATION. A MEASUREMENT WAS TAKEN OF 5.6 MI WHEN THE 2 ACFT WERE NEARING A POINT UNTIL THEY WOULD PASS. CONFLICT ALERT DID NOT ACTIVATE; IT LOOKED GOOD; AN INSURANCE TURN WAS ISSUED. ACFT 2'S DATA BLOCK SEEMED TO HAVE 2/3 'JUMPS' TO THE L. AT THAT TIME; I ATTEMPTED TO STOP ACFT #1 1000 FT ABOVE. I BELIEVED THE 'JUMPS' WERE INVALID AS IT HAS BEEN SEEN SO MANY TIMES BEFORE. THE ATTEMPT TO REVERT TO ALT WAS A PRECAUTIONARY MOVE. HOWEVER; ACFT #1 RPTED ALREADY DSNDING THROUGH THE ALT. CONFLICT ALERT ACTIVATED AT THAT TIME. THE ACFT PASSED WITH 4.5 MI. IT IS MY UNDERSTANDING CONFLICT ALERT WENT OFF AFTER THE ACFT HAD PASSED. ALTHOUGH CONFLICT ALERT IS NOT RELIABLE AS A SEPARATION TOOL; IT SHOULD HAVE PROVIDED ADVANCED WARNING -- TIMELY ENOUGH TO MAKE A DIFFERENCE. CONFLICT ALERT IS USED AS A KIND OF ASSESSMENT TOOL. IF THE CLRNC LOOKS GOOD; BASED ON EXPERIENCE/TRAINING AND CONFLICT ALERT DOESN'T ACTIVATE -- THAT NON ACTIVATION ADDS A SENSE OF CONFIRMATION THAT THE CLRNC IS GOOD.