I recently had the change to speak with a current co-pilot of the CAF B-29 "Fifi". He stated that the pilot or co-pilot would typically advance the throttles for takeoff with his hand on the levers in a "twisted" position so that any given lever would be further forward than all its neighbors to the right, i.e. the tops of the levers would form a diagonal line from forward left to aft right. So differential thrust is used to compensate for a left-turning tendency. He stated that you wouldn't use full throttle at this point, because then you'd have nothing remaining to work with-- this implies that a correction to the left might sometimes be needed, since a further correction to the right could always be accomplished by reducing the right-hand throttles. He stated that once the rudder became effective, the pilot flying the airplane would then call for the flight engineer to "set takeoff power", at which point all the throttles would be advanced to the same position.
This pilot was somewhat skeptical of the idea that it would be common technique for a pilot in any group flying the B-29 to use the brakes to keep the airplane going straight during the take-off roll. He said it would be considered bad technique today in the CAF, due to the risk of skidding a tire and leaving a flat spot.
The pilot's general description of the technique he used, and the specific point that some "room" was left to allow corrections to the left as well as to the right, suggests that the left-turning tendency in the B-29 on takeoff wasn't too extreme. However, since the plane didn't have a steerable nosewheel, some other form of compensation was required.
Moving on from the specific comments by this pilot, to some general observations about left-turning tendencies on take-off in tricycle-gear aircraft that sit in a level attitude until the aircraft is rotated for lift-off--
In general, engine torque creates a rolling (banking) tendency to the left, but the aircraft can't roll (bank) very much when the wheels are on the ground. However the rolling tendency does increase the weight on the left tire, and thus the rolling friction of the left wheel.
In a tricycle-gear aircraft that sits in a level attitude until it is rotated for lift-off, P-factor does not create a turning tendency during the portion of the takeoff run that precedes rotation. In general, P-factor creates a yawing tendency, but P-factor only exists when the aircraft is at a positive angle-of-attack, i.e. when the aircraft is in a nose-high pitch attitude in relation to the direction of the flight path. This is not the case when the aircraft is in a level attitude, with all the wheels on the ground.
Gyroscopic precession does not create any turning tendency during the portion of the takeoff run that precedes rotation. In general, gyroscopic precession creates a yawing tendency whenever the aircraft is pitching, i.e. whenever the pitch rotation rate is not zero. When the propellers rotate clockwise, gyroscopic precession will create a left yaw torque when the nose is pitching up, and will create a right yaw torque when the nose is pitching down. So gyroscopic precession will make an aircraft with clockwise-rotating propellers yaw to the left as it is being rotated for takeoff, but not before or after that point.
During the portion of the takeoff run that precedes rotation, the spiralling (helical) slipstream (propwash) does exert a force on the vertical fin. When the propellers rotate clockwise, the upper part of this spiral is moving toward the right, which exerts a rightwards force on the vertical fin, yawing the nose toward the left. This force will be greater at lower airspeed than at high airspeed, since when the airspeed is low, the propwash or slipstream will tend to dominate over the free relative wind created by the aircraft's motion through the airmass. Even though the vertical fin of the B-29 was not directly centered behind any one of the engines, the airflow over the vertical fin surely had some degree of left-to-right component which undoubtedly contributed to the aircraft's left-turning tendency on take-off.
Harvey S Plourde's excellent book "The Compleat Taildragger Pilot" (1991) contains a pair of graphs showing the individual and combined effects of all the different factors that contribute a left-yawing tendency during the takeoff roll in a propeller-powered aircraft with clockwise-turning propellers, for both the tailwheel case and the tricycle gear case, from the start of the takeoff roll all the way to the moment of liftoff. Different factors are dominating at different points in the takeoff roll.
Now to address one other aspect of the original question:
In the vast majority of twin-engine or multi-engine propeller-powered aircraft, all the propellers turn the same way. As another answer has noted, this simplifies production, and the supply of replacement parts. Of the well-known American-made twin- or multi-engine aircraft built in large numbers during the second world war, only the P-38 Lightning had propellers rotating in opposing directions.
Note-- the links embedded in this answer go to sections in John S Denker's excellent "See How It Flies" website.
Note-- all references to "clockwise" in this answer are from the perspective of someone standing behind the aircraft looking forward, as is the accepted convention in aviation, presumably because it also relates to what the pilot sees when looking forward through the windscreen at the propeller of a single-engine airplane of the "tractor" configuration.
Related ASE answers--
What is the influence of yaw on P-factor?
