Apparently there is a belief, in some audiophile circles, that increasing the power rating of your amplifier is always a good thing. The premise is that a high power amplifier is unlikely to clip, and so a very high power amplifier is very unlikely to clip. The problem is that this reasoning does not consider the power rating of the speakers in the system, the benefits or lack of benefits in the acoustic domain, and the potential waste of money from either unused power capabilities or blown speakers.
I started looking at a more sensible approach to amplifier selection than a blind "more is better" philosophy. I began calculations from the loudspeaker's specifications and the acoustic space I wanted to fill with sound. Crown Amplifiers has an excellent calculator on their website. You need to know the loudspeaker's sensitivity rating, which is normally expressed as a db SPL (sound pressure level) measured at 1 meter from the speaker when the power dissipated is 1 watt.
Crown's formula is: dBW = [required SPL at listener] - [speaker sensitivity] + 20Log([speaker to listener distance]/1 meter) + [headroom], and [final amplifier rating]W = 10^(dBW/10).
Crown's calculator asks for the desired "amplifier headroom." This is the idea that can lead to the original premise that "more watts is always better." The minimum calculation you can come up with using the sensitivity rating does not take into account musical peaks. The sensitivity rating is assumed to have been measured with a sine wave, which allows for voltage rms measurements, and is comparable to the dBu metering that will be used at the preamp (hopefully). Peak signals from irregular waveforms have instantaneous voltages that can be much higher than the equivalent rms/dBu reading. Here it is very important to note that voltage differences expressed in dB and power differences expressed in dB are on two different scales. A doubling of voltage is +6dB, whereas a doubling of power is +3dB.
Adding +3dB, or doubling, the minimum calculated power needed is supposed to account for the unknown peak levels that may exist in your source material. This is not a very definitive answer, but serves as a "rule of thumb" for quick calculations. The open nature of the recommended minimum feeds the notion that "more watts is better."
Certain factors can be considered that put a limit on the unknown peak levels. The maximum output of the preamp is one such factor. This could be determined by test, reviewing the manual and/or specification sheet, or by reviewing a schematic of the unit. Most preamps are going to be running op amp ICs off of +/-15V power supplies, which will cap the instantaneous clean signal voltage output at somewhere near, but below, the power supply rails. Some designs may allow clean output that is near 15V, and others may only allow 13V. Testing with an oscilloscope will reveal the actual specs of your system. Some designs may run off of +/-24V or higher, especially those that are marketed as using "discrete" electronics.
The next bottleneck for the peak capabilities of your system is the power amp input section. This is usually another preamp section, and may again be powered by +/-15V rails in many commercial designs. The ability for this section to clip signal voltage will have nothing to do with the power rating of the output section, so that needs to be carefully considered.
Finally, we reach the power amplifier output, which will be rated at some power level into some load, expressed in ohms. The power level may be expressed in "continuous," sometimes labelled rms, usually the smaller number if multiple levels are given, and "peak," "instantaneous," "burst," or some other label for the larger number. If you have access to the amplifier's schematic, you usually roughly calculate these levels from the output section's power supply. Just as with the preamp, an oscilloscope, test signal, and dummy load with allow you to verify the amp's specifications on your test bench. Just like the preamp, the amplifier's power supply puts a hard limit on peak voltage capability.
After considering all of these factors, selecting an amplifier based on its power capabilities becomes clearer. The calculations, including your reserved dBs of headroom, should reveal to you the number of watts needed to fulfill your application. Purchasing an amp than can deliver more power than you expect to ever use is questionable. You may end up with an amplifier that costs more, weighs more, and could pose a significant threat to your loudspeakers if you aren't careful.
Finally we get the title of the article. "Overpowering" loudspeakers is the notion that it is OK to hook up a pair of near field monitors with power amplifiers of excessively large power capabilities. A white paper entitled "Speaker Power Requirements" from JBL asserts that connecting loudspeakers with a n watt rating can be connected to power amplifiers with a 2n power rating, so long as the system is carefully monitored. The alleged benefits are "cleanereration of the whole system," and "less listener fatigue." This is very misleading advice. The premise that you can hook up a 1W speaker to a 2W amp, a 10W amp, or even a 1000W amp is valid. This is what JBL refers to as "careful monitoring." Why you would risk blowing up your speakers due to a moment of, let's say, "careless monitoring," is not clear from reading the JBL paper. JBL also fails to describe what "cleaner operation" or "listener fatigue" actually is. An internet search of "listener fatigue" was not promising. It does not appear to be a scientific term.
The JBL paper does describe a more sensible approach of either matching power ratings (100W speaker to a 100W amp) or, perhaps the best advice, make the speaker's power rating double the output power rating of the amplifier. JBL recommends these sensible approaches only for live sound and musical instrument amplification respectively.
Connecting a heavy duty speaker to a low power amp is sometimes called "underpowering" and there is a myth that this is a bad idea. The myth comes from confusion over loudspeaker and amplifier ratings. Common sense dictates that driving a 1000W rated speaker with a 1W amp is not going to blow the speaker. If an amp has a "continuous" power rating that is derived at by driving the amp with a sine wave, that power rating will be 1/2 of what the output will be if that sine is fully clipped into a square wave. So if you have a 100W continuous amp, and you wind up clipping it in the attempt for maximum volume, your speaker will be dissipating 200W. This is probably not a problem if you were smart in the first place and bought a 200W or higher rated speaker to go with your 100W amp. The problem was with the customers who bought 100W amps, and then were frustrated that their "100W" amps blew up their 100W speakers. Instead of blaming the customer for clipping their amp, the customer was told to buy a bigger amp so they wouldn't clip it. Such was born the myth that you can "underpower" speakers, and also the myth that "overpowering" is a good thing.
At my work we had a driver blow out on a very expensive PMC monitor. This monitor was bridged across the two power amps of a very large Bryston dual unit. The Bryston was spec'd for 500W in 4Ω, and 900W in 8Ω bridged. This monitor was 4Ω, which means bridging it was either beyond the amp's capabilities, or the potential power going into the PMC was over 1000W. Worse yet, the send was a +4dBu balanced line, but the controls were set for maximum gain. The operator at the source had to send mic level signal to get normal (85dB SPL within 10 meters) acoustic levels.
After lowering the gain to minimum, I calculated that there was no way, given the very good sensitivity of the PMCs, that anything near 500W was needed to drive the monitors. I reported this to my managers, but they were doubtful. Why? Because the PMC salesman had told them to use the Brystons in bridge mode. They reluctantly allowed me to rewire the speakers to normal mode (and thus an entire Bryston amp was freed up), but I could tell they still had suspicions that somehow the system would sound worse off because of it. Once that seed of doubt is sown, it is quite difficult to root it out.
Moral of the story is: don't let a salesman convince you to buy a huge amplifier if you are only trying to listen to a stereo set of speakers within a few yards/meters. If you want a reliable system, get speakers with good power ratings. Even there, do some math and buy what you need. Don't rely on "careful monitoring." Make the system as fool proof as you can.