Usually shorting phases is fine.
If you short the phases while the turbine is over-speeding, then that pulse of current can be very high. But, then, you already have a problem if they're overspeeding, anyway.
If the generator is connected in star and if the star point is hidden somewhere in the winding, you can only make short-circuit in star. The peak of the Q-n curve for short-circuit in star is lying about a factor 0.8 lower than for short-circuit in delta. The Q-n curve for short-circuit in delta is lying higher than for short-circuit in star because higher harmonic currents can circulate in the winding for short circuit in delta.
You can make short-circuit before or after the rectifier and close to the generator or at the position of the batteries. Make short-circuit after the rectifier is only allowed if the rectifier diodes can have the large short-circuit current. You get a voltage drop of about 1.4 V over the rectifier and also a voltage drop if there are long cables in between the generator and the short-circuit switch. This voltage drop results in shifting to the right of the Q-n curve and the P-n curve of the generator. This shifting may make that the P-n curve of the generator for short-circuit is no longer lying to the left side of P-n curve of the rotor for the maximum wind speed.
In figure 4 of my public report KD 484, I give the P-n curves of the rotor for different wind speeds, taken into account the effect of the safety system for higher wind speeds, together with the measured P-n curves of the generator for short-circuit in delta and in star. It can be seen that the P-n curve for short-circuit in delta is lying most to the left of the P-n curve of the rotor for V = 9.5 m/s and so braking in delta will result in a larger power and torque difference and so the rotor will slow down faster. The P-n curve of the rotor for V = 9.5 m/s is the highest rotor curve as the safety system turns the rotor out of the wind such that this curve is also valid for higher wind speeds.
Making short-circuit over a resistor creates a voltage drop over that resistor and this voltage drop also results in shifting to the right of the P-n curve of the generator. So you have the advantage that not all the heat is dissipated in the winding but the disadvantage that now there is a larger chance that the P-n curve of the rotor is lying no longer to the left side of the P-n curve of the rotor for the maximum wind speed. My advise would be to make short-circuit in delta before the rectifier and at a distance as close as possible to the generator. If you won't climb the tower, this position will be at the tower foot.