Ko!
I was a KO-Full user for about a year, so if you have any other questions about the system, don't be afraid to ask...
I understand the problem with starting on a negative number for an IRC. Simply adjusting it to "0" makes everything much simpler, as it has been stated. One of my buddies who was using it at the time with me like to actually start at the given IRC in the book, but instead of saying "negative #(whatever number)" you could substitute the word "drop" for "negative". So, in this case for a 6D shoe, you could start the IRC at DROP 20. This makes thinking about the count much easier because you don't have to deal with taking the time to think about long words.
KO is proven to be a successful card counting system in the long run, but some advantage is given up for its simplicity. Because it is not as accurate as a level two or three system, you have a greater RoR, which therefore means you must be SUFFICIENTLY funded in order to be successful because you will experience bad variance and bad swings!
Trust the mathematics provided by the system and use the indices. Small sessions are meaningless whether you're up 50units or down 50units. You need to understand that, on a graph of your bankroll, the trend will never be a line which increases linearly (y=mx+b). Your BR will peak, fall, peak, and fall BUT the overall result will be an averaged line with a POSITIVE slope.
If you're interested in a stronger version of KO, there is TKO (True Counted KO). There's a thread or two on it and a book already written. It's called The Color of Blackjack by Daniel Dravot. And, as I've stated in these other threads, TKO is more accurate in estimating advantage and betting opportunities. There are points in 6D shoe games where you should actually be betting more, earlier than the KC. If you are prepared for the same system, but with deck estimation, then you should certainly move on to it. It's better.
Hope this helps!