I think the theory is good.
If something has a higher price, it has a lower upside and a bigger downside in a fixed forward interval, so own less of it. And vice versa.

But...the thing that is more important is how sure you are of your outlook on each investment.
(then decreased by the "I was sure, but I was wrong" factor)

For two random picks, I love DG and I love BRK.
Based on my squinting at the tea leaves, DG is apparently a more attractive investment over the next 3 years.
But I am very much more sure of my opinions about the value of Berkshire than I am about DG, so I have a much bigger BRK position, and I'm not about to reverse them.

If you had a lot of investments that you were equally sure about, I think adjusting them based on relative forward returns would make a lot of sense.
Sometimes when I have two fairly similar investments I have a gander at the ratio of their prices.
Stockcharts.com has the stock price ratio feature, for example two dollar stores: stockcharts.com: SharpCharts | StockCharts.com
This makes it very clear that dollar Tree has done way better than DG recently.
When that sort of graph swings a lot to the high or low side, it's not a bad time to shift money from one to the other.
You don't have to pick a peak or trough--moving back and forth will increase your long run returns provided you alternate the direction of your movement.
But for me it's relatively rare to have two positions that I am equally sure about and fulfil similar functions within the portfolio.
And it's hard to see where cash fits into the formula.

The main reason that sitting on your hands is usually the right decision is that most investment decisions (including mine) are mostly based on forecasts of future outcomes that are much worse quality than we think they are.

Jim