Why are automated GUI tools so fragile?

After about a year and a half, I work with SilkTest, which is a GUI automation tool for both desktop and web applications. It simulates mouse and keyboard inputs that ultimately model end-user behavior. However, I find it to peel off a bit; Calling the Button.Click () or DialogBox.Close () method, which works very well 9 times in a row, seems unsuccessful on the 10th call, only to get back to work on the 11th. Normally, I would just do it using the SilkTest application (or the application under the test or the operating system or whatever you have), but then I see that there are similar problems with other GUI automation tools such as Selenium:

Selenium Click () error with anchor elements

Selenium Click () cannot click button object

I know that for desktop applications, each control / GUI dialog has a tag element associated with it (at least in Windows GUIs), and for web pages there is a hierarchy of elements of the page element domain model. I assume that these tools sometimes run into problems navigating through these hierarchies and finding unique elements and controls. But what is going on here? SilkTest is a relatively old commercial software package, while selenium is relatively new, open source and constantly evolving. The fact that both of them can have similar problems causes me a couple of flags.

Also, is this the case with other GUI testing tools? Or did I just have an unusual experience?