Exactly. One of my systems has 4 losers/1 winner but is extremely profitable as the winners are, on average, 11 times larger than the losers. That's on full auto. If I deploy it using discretionary measure it's 2/1.
I have attached the most accessible pdf covering these formulas. I have no idea what the ET limit on pdf attachments is, so PM me if you want more papers on the subject. FWIW, I don't think any single ratio can fully characterize risk vs. reward for a given model/system or trader.