I'm running MI1 biwires from the 4 Ohm tap on a Music Reference RM-200 to VMPS RM-30s.
Here's my understanding of Alpha-Core's MI offerings:
MI3s offer twice the material (i.e., raw copper by weight) and half the resistance of MI2's, which in turn offer twice material and half the resistance of MI1's. Specifically, the equivalent gauge of MI3s, MI2s, and MI1 is approximately 7 AWG, 10 AWG, and 13 AWG respectively. This corresponds to a resistance of 0.0011, 0.0022, and 0.0044 Ohms/ft respectively (
http://www.alphacore.com/mispeaker.html).
I thought I recalled a time when the cost of bare, unterminated wire scaled exactly with material (i.e., cost of unterminated MI3's = 2 * MI2's = 4* MI1s), but according to
http://www.alphacore.com/shop/Speaker_CablesznCopper.html this is not true today, and indeed you do get a bit of a "bulk" discount as you move up the scale.
The resistance measures are relatively unimportant. Any system will overcome the difference between 0.0011 Ohms/ft and 0.0044 Ohms/ft with only the slightest adjustment to the volume. But impedance is another story.
Resistance is strictly an inhibition to electron flow for DC: i.e., a zero frequency, infinite wavelength signal. Once we introduce alternating current--which is what amps use to drive speakers--then capacitive and inductive reactance interact to also impede current flow. The (complex) sum of resistance, inductance, and capacitance is called impedance. It is the net inhibition to AC flow. Inductance can be thought of as an inhibition to a *change* in current (not simply an amount, or amplitude of current) caused by the generation of a magnetic field. Because music as encoded in oscillating electrical signals is highly dynamic (lots of change), excessive inductance per se--without some other reason for its presence--cannot be considered a good thing. Inductance acts as a low-pass filter, so the closer you get to DC, the better the current flow at the frequency, while the higher the frequency, the greater the inhibition.
Alpha-core's claim to fame is that by sandwiching the (+) and (-) wires so tightly together, they greatly reduce inductance (at the cost of increasing capacitance), to the net effect of low, final impedance numbers. Specifically, they claim a characteristic impedance of ~1.7 Ohms for MI3, 2.5 Ohms MI2, and 4 Ohms for MI1 (
http://www.alphacore.com/mispeaker.html). Additionally, they do this in a very smart way: by sandwiching the strands so closely together, the mutual inductance that is generated in one wire as a direct result of oscillating current within it, is canceled almost exactly by the opposite effect in the other wire. So they get low impedance numbers by lowering net inductance--a smart quantity to lower for audio applications. Some folks claim that wires that tend to maximize inductance (the opposite of Alpha-Core's offerings) roll the highs (which may be good or not so good, depending on what's coming out of your amp).
Lower inductance comes at the cost of increasing capacitance. Because the (+) and (-) strands are so closely sandwiched together, it is easier for current to cross the insulation between the strands. Capacitance acts as a high-pass filter, so it is high frequencies that are "shorting" across the wire. But the capacitance numbers are in the pico-Farads/ft, so only spurious frequencies way off the audio scale (and beyond the capacity of most amps to generate) would be affected. So you are kinda' getting something (low inductance) for nothing (high capacitance): a good trade. Some amps may pathologically oscillate when presented with a high capacitance load, so the simple solution is to short the wires with a capacitor and a resistor (called a Zobel network), so the HF energy is simply dissipated as heat in the resistor, leaving the amp unaffected.
What does all this mean? Imagine, if you will, a complicated Rube-Goldberg contraption of pipes of differing sizes, all connected in funny ways and filled with water. This contraption is our analogy to your system as it runs from the wall to your source, pre-amp, amp, and finally speakers. Now start an oscillation at one end by tapping (e.g., at 60 Hz), and the goal is that via this convolution of pressures and relays and valves and so forth, a complicated oscillation of waves--an amplified mimic of the musical source--will come out the other end. With this analogy you can see that when you join a wide pipe to a narrow pipe, or vice versa, as the wave hits the junction it will create reflective waves back down the stream, which in turn will interact and (slightly, or not so slightly) corrupt the signal. This is an example of an impedance mis-match, and with electricity there is the same, analogous phenomenon. When you change impedance, waves are generated on the line. This is one reason why digital cables from a transport to a DAC can make a difference. Even though virtually all transports and all cables will send S/PDIF signals bit-perfect, minor anomalies caused by impedance mismatches generate jitter on the line, and affect the signal (specifically, it's timing). So, impedance matching has some relevancy to audio, and Alpha-Core's claim is that they do a better job than most in minimizing impedance mismatches between the amp the speakers, while at the same time lowering inductance.
How is this standardized? If you take a piece of speaker wire (w/ (+) and (-) strands), and don't hook it up to either the amp or the speaker, but simply measure the impedance between the strands it will measure essentially infinite. That is, the insulation between the strands creates a very high inhibition--which is good, because anything measurable would be a short. But in reality the inhibition is not infinite. There is some material separating the (+) and (-) strands, even if it is only air or an almost perfect vacuum. And indeed, there is a slight crossing of electrons between strands--that is, there is both a voltage and a current. An indication of the insulation material's propensity to carry--or inhibit--current is called the dielectric. The longer the cable, the more insulation, or opportunity, there is for electrons to cross. If we took an infinite length of cable, and then asked "Given this infinite length, what is the net inhibition to election flow across the insulation?" the answer would be the characteristic impedance. For RCA interconnects, this is historically 75 Ohms. For speaker cable there is no standard, so Alpha-Core says, "Well, let's make it close to the speaker's average band-wide impedance in a effort to minimize impedance mismatch." (Quotes are mine, not theirs).
At the end of the day, I think most systems would be challenged to uncover a difference between 1.7 Ohms characteristic impedance, 2.5 Ohms, and 4 Ohms--but, no doubt, some will show a difference with different wires. But do this part on listening—I don’t think we can rationalize/intuit that 2, or 4, or 5 Ohms will be better just because one number is closer to a speaker’s average impedance rating: it is better to compare Alpha-Core's offerings as a whole to cables that are in the 100+ Ohms category. I’m sure MI1’s would sound great on some 8 Ohm speakers. Alpha-Core gives a bit of a guide to choosing cables based on characteristic impedance here:
http://www.alphacore.com/mifaq.html#select.
Hope this helps