The only way to avoid battery problems is to look after them carefully, always make sure they are charged before the contest and test them regularly. For this you need a good quality charger, one that can cycle the battery and measure its capacity. Check the the capacity at least once a year and keep a record of the results of every test. Merely looking at the results for a battery can tell you a lot. I discard a battery once it has fallen below 80% of its original capacity.
If your d/t release, like the Koster timers, pulses a solenoid or heats a wire to burn through a rubber band, its also worth checking the shape and size of the voltage drop when the d/t releases. The Koster may only use 80-300 uA to run the timer, but it pulls 500 mA for 15 mS on d/t and this pulls a fresh NiCd down 0.5 V in a nice, flat bottomed bathtub shape. As the battery ages, the voltage drop increases and the curve becomes a saw tooth. I discard a battery once the pulse becomes a sawtooth or is more than 1 volt deep.
This is easy for me to do since I have an oscilloscope, but there are other ways of checking.
The following test is probably also valid if your timer uses servo(s) to control the model. To do this all you need is a digital voltmeter and a switcheable load that you can connect and disconnect while the voltmeter is monitoring the battery. Adjust the size of the load so it discharges the battery at 0.1C, i.e. 5 mA for a 50 mAh battery. When the load is connected the voltage should drop no more than 100mv and stabilise at the new voltage. Leave the load on for 4-5 seconds to check that the voltage is stable. Now disconnect the load. You should see a return to within 2 mV of the original voltage within 2-3 seconds. If the voltage under load doesn't stabilise but continues to drop or if it doesn't return to the original voltage when the load is removed then I'll consider the battery is starting to fail and discard it.
If you can measure the current drawn by timer and servo(s) with the model fully hooked up, it would be sensible to revise this test to draw the largest current you see instead of the rather arbitrary 0.1C I suggested.
You should also consider how much of the battery's capacity is used during a normal contest or on a long trimming day and adjust your charging regime to suit. The Kosters are one extreme: Thomas once calculated that if you made 10 test flights an hour for 18 hours you'd only use 33% of the standard 50 mAh battery's capacity. In other words, recharging every night with a 1 mA trickle charger is enough to allow you to fly all day and to go on doing the same throughout a World Champs and the associated International. At the other end of the scale an M&K timer with a standard 50 - 100 mAh battery won't quite manage a full contest on a charge and is unlikely to handle a busy trimming day. My solution here was to buy a battery operated peak charger with a 100 mA output and run it on the field from a 1000 mAh battery pack. I always have one spare flight battery on the field. I fly with one battery while the spare is on charge and swap them before every flight: the charger has recharged its battery before I get back from making a flight and want to use it. This way I know that the model always has a fresh battery in it and, because the swap is part of my standard preflight routine, it won't get forgotten. I keep a bigger charger back at the ranch and recharge the peak charger's battery overnight.