The evaluative vs average metering option is a CF.n on the camera and it works with both versions.
But you need to grasp -- and this is not rocket science -- that regardless of how the flash is controlled exposure for any given output level will only be correct at one distance.
So average isn't a magic bullet, it simply evaluates the scene much differently. Most indoor flash scenes are mostly background. Pre-flash goes out from the flash but very little comes back. Camera evaluates the return and says to flash "Hey it looks really dark out there - pour on the power". What that extra power does is raise the exposure of any light areas in the foreground, which otherwise would be reproduced as gray, up to the point they are reproduced accurately (which you control via FEC). So its really the fact the background is darker than average which makes average metering appear to work better, in general. But it will usually blow highlights in some areas. If you fill the viewfinder with white - a bride against a white wall - average metering will underexpose it because the camera sees lots of light reflecting and says to the flash "Hey it looks really bright out there - cut back the power".
I used auto thyristor Vivitars from the mid 1970s until I switched to the Canons in 2005. They have an averaging sensor cell on the flash. You set f/stop on the camera per the dial, then as you change shooting distance the sensor regulates flash duration to keep the exposure the same. But changes in average scene reflectance would affect the average metering. With experience I learned to anticipate when the average was lighter and darker than middle gray and compensate by changing the aperture: opening it for brighter than average scenes which would cause the flash to output less power, and closing it for darker scenes because they would cause the flash to output more power which would blow the highlights.
The only difference with average, once you set FEC to obtain the results you desire, is that is will take a more significant change in scene reflectance to affect the exposure.
Manual isn't idiot-proof either. If you shoot a person in a white shirt and put their shoulder closer to the flash than the face you'll either correctly expose the face and blow out the shirt, or if you correctly expose the shirt wind up underexposing the face. The solution to that dilemma? Learn how to pose people so they look good and their faces are closer to the light.
An important part of making flash work well is learning how to compose for flash so what is most important is closest to the flash. When you start doing that you'll find any TTL flash mode will produce better results.
ETTL will help train you to compose more effectively for flash because if you put unimportant stuff in the foreground ETTL will expose it correctly and underexpose what you think is more important further back.
