NiMH's don't like to be discharged much below one volt. Definitely not good to pull them down to zero. If it's a good quality cell, doing it once or twice probably won't kill it, but it'll shorten it's cycle life. The size of resistor you use will depend on how fast you want to discharge it. Discharge rate is usually stated as a ratio of the mAH capacity of the cell. For example, a 1/5C discharge rate for a 2000mAH cell is 400mA's for 5 hours (2000mA / 5 = 0.4A). Use Ohm's law, R = V / I, to figure the ohm value of the resistor. The nominal voltage for MiMH's is 1.2 volts. So, for the 400mA 1/5C example, R = 1.2V / 0.4A or 3.0 ohms. Power (P) = V * A. Using our example, 1.2V * 0.4A = 0.48 Watts, so a half-watt (or bigger,) three-ohm resistor would work fine. It's best to use a resistor that will discharge the cell at roughly the same rate as the device you use it in, but if it's a LED light you might not want to wait that long. If you can find the spec sheet for the cells you're using, lookup the maximum recomended discharge rate and stay well below that. Use a voltmeter to monitor voltage during discharge and stop discharging when the voltage matches the other cell of the pair (which was measured using the same load resistor you use for discharging). Then charge the pair. It's probably best to forgo all of this and get a charger with independent (single cell) charging capability. If the device you're using the cells in automatically powers down when the voltage is too low, then you could use a multiple of three cells, say six, charge them up, run both groups of three cells through the device, and then charge the three pairs. This will work because the device will shutdown at a predetermined point, and all your cells would stay roughly matched. I use paint to color code my matched sets of NiMH's. But again, your best option is a charger that doesn't require cells be charged in pairs. I love my MAHA 401.