Did you know with gigabit and higher, it's against spec to make patch cables the exact length you need? The issue is with short cables.
I found this oh HP's site, but it is noted in other places too:
It is possible that short cables will not meet requirements due to the impact of Return Loss.
For example, if a cable is 110ohms in impedance and it is plugged into a port with impedance from 100ohms at 1MHz to 85 ohms at 100MHz, reflections may occur because the cable has almost no attenuation.
Multiple reflections may result in this scenario, possibly resulting in standing waves that build in strength until the signal at the receiver is overwhelmed.
Therefore, it is necessary for a cable assembly (.5m or 2.5m) to meet Return Loss, Near End Crosstalk, and the other specifications set forth in the IEEE 802.3ab 1000BASE-T standard.
Cables under 2.5m have strict requirements to length. They basically have to be of multiples at to not promote standing waves. You have to know the wavelength of the signal, based on the propagation rate of electrons in that particular cable, at a particular frequency. You don't want the cable to be of a length that results in waveform peaks standing at the ends of the cables, else they'll be reflected back by the receiver (because it doesn't attenuate the signal enough) and cancel out other arriving signals.
Under 0.5m it's possible for the reflections to move all the way up the cable and reflect again from the sending side. This causes true standing waves, and will make communication at speed impossible.
Once a cable hits about 2.5m it'll offer enough attenuation of it's own, that combined with the receiver there won't be enough energy to cause interference reflecting back up the line.
