Musicnoise is correct in the sense that McIntosh's Power Guard circuit will essentially keep the amp from clipping to any significant degree . . . and amplifier clipping causes the overwhelming majority of (power-related) speaker failures. And I agree with others that a simple wattmeter of any sort is of very limited use for preventing speaker damage . . . your ears and basic common sense are better things to go by.
But for your specific meter accuracy question, it varies from model to model. For the amps without autoformers (including IIRC the 6450), it's essentially a voltmeter, calibrated to a specific impedance load. So when your meter reads "100 watts", what it really means is "20 volts" . . . which is the same thing IF the amp is driving a 4 ohm resistive load . . . which your loudspeaker is obviously not. But in your case, "100 watts" on the meter will still be pretty close to the maximum power output when driving your speakers, so concept of "meter full scale" is the same, but the wattage calibration will be inaccurate.
For the later McIntosh amps that have autoformers, the meters are much more accurate, because it's the drive to the autoformer that's measured, not the voltage at the speaker terminals. They also have a current sensor that gives a certain amount of correction factor for the typical variations in a loudspeaker load. So if the load on, say, the 4-ohm autoformer tap is maybe within +/- 20% of 4 ohms, the meter reading will be very accurate. But if you were to put a 2-ohm load on the 8-ohm tap, this would be outside of the correction range of the current sensor, and the meter wouldn't be accurate at all.
But in the real world, the 6450 isn't a hugely powerful amp . . . and with clipping removed from the picture by the Power Guard, you really shouldn't need to worry a whole lot with an average domestic loudspeaker. The meters are handy for checking i.e. channel balance on a mono record, or verifying that your cables are hooked up correctly, or as an educational tool to understand the rough relationships between power and SPL, but they're not sufficient to really tell when you're about to damage your speakers.
But for your specific meter accuracy question, it varies from model to model. For the amps without autoformers (including IIRC the 6450), it's essentially a voltmeter, calibrated to a specific impedance load. So when your meter reads "100 watts", what it really means is "20 volts" . . . which is the same thing IF the amp is driving a 4 ohm resistive load . . . which your loudspeaker is obviously not. But in your case, "100 watts" on the meter will still be pretty close to the maximum power output when driving your speakers, so concept of "meter full scale" is the same, but the wattage calibration will be inaccurate.
For the later McIntosh amps that have autoformers, the meters are much more accurate, because it's the drive to the autoformer that's measured, not the voltage at the speaker terminals. They also have a current sensor that gives a certain amount of correction factor for the typical variations in a loudspeaker load. So if the load on, say, the 4-ohm autoformer tap is maybe within +/- 20% of 4 ohms, the meter reading will be very accurate. But if you were to put a 2-ohm load on the 8-ohm tap, this would be outside of the correction range of the current sensor, and the meter wouldn't be accurate at all.
But in the real world, the 6450 isn't a hugely powerful amp . . . and with clipping removed from the picture by the Power Guard, you really shouldn't need to worry a whole lot with an average domestic loudspeaker. The meters are handy for checking i.e. channel balance on a mono record, or verifying that your cables are hooked up correctly, or as an educational tool to understand the rough relationships between power and SPL, but they're not sufficient to really tell when you're about to damage your speakers.