Okay, you asked for a specific in your sources material. Here is what you source fails to include:
One problem that is common with all radiometric dating techniques, namely that the system remains closed following the volcanic event. In other words, there has been no loss of argon following crystallisation. Loss of argon could occur if the mineral sample has been weathered or if there has been a further episode of reheating; in both cases an underestimate of age will be obtained. However, petrographic examination may reveal evidence of weathering, while the presence of secondary minerals, such as xenocrysts, will tend to indicate that recrystallisation has occurred. A further way possible loss of argon can be detected is to compare age estimates derived from the whole rock and mineral fractions to determine whether or not they are concordant. This approach can also be used to establish whether older mineral fractions have been incorporated into the rocks. Secondly, the measurement of 36 argon (the Ar/Ar method) accounts for any excess atmospheric 40Ar they may be present in extrusive rocks.
Source: Quaternary Dating Methods, Mike Walker 2006. Chapter 3.2.3, page(s) 58-66.
Also note that the K/Ar method was developed back in the 1960's. And yes, back then some problems with excess or diffused Ar did occur then. That is rarely the case with today's methods and technology. And BTW, as mentioned above, one way of determining whether the system is closed or not is through the observation of whether the sample has been weathered or not. Good grief! A 5th grader can tell whether a rock sample has been weathered or not. It is more than obvious just from casual visual observation.