Thank you for reminding me that this is P&N
You can retrieve the final report by the US-Canada Task Force through the following link:
http://www.ferc.gov/industries/electric/indus-act/reliability/blackout.asp
The report identifies deficiencies in four areas that contributed to the initiation of the blackout. If you read the report, you will see that plant tripping isn't among them.
Many fossil-fueled steam driven generators run at or very close to their maximum outputs because it is the point at which they are most efficient (i.e. fewest BTUs required to generate a MWhr). There are also significant time delays (i.e. minutes) required to increase fuel delivery to the boilers to increase steam to finally increase power output. These two things limit their immediate (i.e. within a few seconds) response to automatic governor action (based on frequency deviation). In fact, some combustion turbine generators actually exacerbate the frequency deviation by responding in the wrong direction due to plant control systems keyed on exhaust temperature limits. This is enough of a concern to the industry that NERC is nearing completion of new reliability standards that will specify minimum levels of "frequency responsive reserves".
The kinds of output change requests you get via telephone or by telemetry from a central control center don't require action within seconds. The electrical system is stable (or has been restabilized by automatic actions like governor response) and we're just adjusting which generation needs to be run how hard.
I gather that your experience may be more with smaller, perhaps isolated, power systems. The scale of the eastern interconnection may be orders of magnitude greater than what you're used to. A 1000 MW generator loss is likely to result in less than a 0.05 HZ drop in frequency.