If all your pedals are true-bypass, then the length of cable your guitar is driving changes depending on which pedal is the first one engaged in your chain.
That's not such a big deal - most pedalboards are wired with nice short cables between pedals. However, when all the pedals are true-bypass and off, then your guitar is driving the cable to the board, all cables on the board, and the cable from the board to the amp. If you use the same length cable from guitar to board as you do from board to amp, then switching all of your all-true-bypass pedals off more than doubles the amount of cable your guitar has to drive, and that will have a noticeable effect on your tone.
The "solution" to this problem is to buffer the signal either at the guitar (e.g. use active pickups) or put a buffer amp as the first unit on your pedalboard. That way, the buffer amp drives the pedals and the cord from the board to the amp. Having a low output impedance, the buffer is (compared to the pickups) unaffected by changes in load.
On the downside, now you have a buffer in your signal path even when all of your pedals are off. Granted, that's better than the Boss approach of having an FET switch and buffer in every pedal, but wasn't the point of true-bypass to have nothing but wire between your guitar and amp when all the effects are off? And now we need an always-on buffer at the head of the chain to isolate the guitar from changes in cable load. Hmmm... One step forward, one back.
Now, what if all of your pedals are half- or hardwire-bypass? (Gawd, dontcha just love marketing terminology?) Well, the more pedals you put in your chain, the more they're going to load down your guitar when they're all turned off. That's because the guitar is driving the input of the effect circuit even when the effect is switched off. That's bad, huh? Sure it is, if you use a lot of effects chained together. But you can fix that by putting a buffer amp at the head of your effects chain. Just like when you use nothing but true-bypass pedals. Hmmm, again...
If you're only using one pedal (without a buffer), you might actually be better off with a half-bypass switching arrangement. Why? A true-bypass pedal will unload your guitar from the second half of the cable run when your turn on the effect. That means you're going to hit the pedal with a signal that's a bit brighter than what you send straight to the amp with the effect off. Maybe that brighter signal works well with the effect on, maybe it doesn't... On the other hand, the half-bypass pedal, when switched off, loads your guitar with not only the second cable run, but also with the input impedance of the pedal. Yes, that pedal input impedance lowers and flattens the resonant peak of your guitar. This is what most people mean when they talk about a pedal "sucking tone". That's the bad news. The good news is that the presence of that second cable run has a less audible effect on the tone - its incremental contribution is reduced by the always-present loading of the pedal itself. In simpler language, the pedal isn't going to get a suddenly brighter signal from the guitar when you switch on the effect.
Granted, that difference is small and applicable to people who use only one pedal in their chain. Everyone else ought to be using a buffer at the head of their pedal chain, and then it really doesn't matter whether the pedals are true-bypass or half-bypass. Or does it?
Did you ever notice that true-bypass pedals are more likely to click or pop when switched? There's a reason for that. The input and output circuits of a pedal are coupled using a capacitor to isolate the signal from the DC voltages present on the internal circuitry. These capacitors often "leak" a tiny -- almost insignificant -- amount of DC current.
When you disconnect one end of the capacitor, it tends to "float" toward the DC voltage found at the other end. When you kick the stomp button to turn on the effect, that floating end is suddenly connected to a zero voltage, and the capacitor has to suddenly replace the charge lost during the "float" time. This, of course, sounds like a pop or a click.
There's a really simple fix that eliminates most of the switching pop. All you have to do is to connect a resistor from the capacitor to ground to continually replenish the charge that would have been lost while the effect was off and the input (or output) capacitor was "floating".
But the true-bypass switching arrangement is at a disadvantage here. That extra resistor from the input capacitor to ground has to have a large value to keep from loading down the guitar. The large resistance can allow the capacitor to float a bit off ground, and you still get a small pop when switching the effect. This is especially noticeable when the effect has gain, e.g. distortion or compression. Because the output of an effect is at a low impedance, you can use a much smaller resistor to prevent the capacitor from floating off ground.
So, is a true-bypass switching arrangement ever beneficial? Maybe. If your pedal has a vintage transistor circuit with a low input impedance, a true-bypass switch will prevent that circuit from suppressing the upper-midrange peak of your guitar. (In other words, it'll prevent the tone-suck.) And because the input circuit is already at a low-impedance, the designer can effectively eliminate the switching pop due to the input capacitor "float". But you still have to account for the big difference in loading of the guitar between the effect off and effect on conditions. You still might be better off either using a buffer at the head of your effects chain. Or -- if you use just one effect -- stick with a half-bypass effect and adjust the rest of your rig to get a good tone with the pedal in the chain and turned off.
Honestly, I think we spend way too much time worrying about small differences like bypass switching. There are certainly engineering tradeoffs that can be made. And the marketing people have a field day relabelling something that's only "different" as something that's "better". In the end, these differences have way less effect on our sound than a bit of practice or even adjusting amp EQ. We want to hear true-bypass as "better" because we pay more for it. But it really is merely "different"... and, of course, more expensive.