This column was originally published on RealMoney on July 28 at 9:59 a.m. EDT. It's being republished as a bonus for TheStreet.com readers.On Thursday night, Richard Suttmeier and I had a chat in Columnist Conversation about the difficulty of valuing the homebuilders. One of the themes we touched on is how experience increases one's ability to look a bit beyond the numbers and find patterns; specifically, to see the valuation patterns evolving. Doesn't it make sense that the best analysts, irrespective of discipline, apply a healthy dose of experience to look past the raw numbers? If consistently good trading and investment decisions were possible using only raw data with a rigid interpretation, we could all just go home and let the machines do it. There's got to be some aspect of learning that goes into the trading process. What about program trading, you ask? Well, even program traders employ system learning tools like neural networks and walk-forward optimization of their systems to enable them to adapt to changing market conditions. In essence, their programs are geared to do the very things we do as traders: adapt. Over the past couple of weeks, there has been a bit of chatter on RealMoney about head-and-shoulder patterns. I've seen a few lately, and they're all different. By examining them more closely, I think you'll learn that not all patterns are created equally. Some offer more profit than others. Let's take a look. An inverse head-and-shoulders pattern is close to completion on the weekly chart of Pfizer ( PFE). But there's one problem: This relatively low-beta stock has run about 15% in the last couple of weeks. That's an advance that just begs for profit-taking. So a breakout above the neckline would have more profit potential if more work was done at current levels to allow for some profit-taking. Notice the slope of the neckline: downward.