To Design Simple-to-Build, High Efficiency Antennas with a Simple Feed System
And what could be more simple than a direct feeding of your antenna with no matching arrangement? In fact, why should we accept that a matching arrangement is required on a Yagi at all?
Often I hear hams talk about 'loading up anything with a decent ATU'. Yes, we can make anything presentable to the transmitter in order that it will deliver RF but how much actually gets to the antenna and more importantly, how much of that RF will the 'antenna' radiate? In most cases we are using 50 Ohm coax to feed our antenna system, if the load (antenna) is not presenting a 50 Ohm impedance, no matter what our ATU does for the transmitter, the load is still way off on the coax side and in these instances, our coax becomes apart of our antenna. This does not matter so much when we are in a 'load anything' scenario as perhaps we just want to radiate some RF regardless of the fact that much less RF is being radiated than is leaving the transmitter. However, when using a Yagi and looking towards directional properties, if the coax is radiating, less of our RF is firing in the direction we want it to go. So do we need an ATU or matching arrangement at the antenna? Not if you want to maximise the amount of RF radiated in the desired way. In these instances we need to choose the right antenna.
You may or may not know that as elements are added to a yagi, the impedance drops. Hence, a badly designed (or unfinished project, depending on how you look at it) Yagi will require some form of match at the antenna in order for the feed line to see a 50 Ohm impedance. With the match occurs loss and inefficiency. If we are able to maintain the input impedance to around 50 Ohms, no matching system would be required and one area that losses occur would be removed.
Yagi Gain and Impedance
Within the ideal world, we would have maximum forward gain, maximum bandwidth and maximum front to back ratio and all whilst throwing in minium reactance and SWR at the same time. All this whilst maintaining a 50 Ohm Impedance. However, this Holy Grail Yagi does not exist. If you want maximum forward gain, we have to sacrifice one or more of the other areas. One of the easiest ways to increase performance generally (forward gain and front to back ratio) is to drop the input impedance.
Along with the neccessity of having to have a matching arrangement in order to use this antenna, bandwidth is also compromised along with the antenna stability. By stability I refer to the effects of rain, snow and ice which will all effect the tuning of the antenna. Additionally, close by objects can effect the tuning. This could be other antennas or even houses, trees or other buildings if they are too close. Add narrow bandwdith and instability to the mix and you can quickly end up with an antenna that is not usable in certain weather conditions or locations!
The challenge is to create an antenna which has excellent forward gain, good front to back ratio, nice flat SWR curves across a wide bandwidth, that has good stability whilst maintaining a 50 Ohm feed impedance. Yes, this maybe more difficult than dropping feed impedance and take time and some skill in order to acheive. Whilst the ultimate result may not 
have the on paper performance of a lower impedance antenna, the overal radiating efficiency and performance of the OWA ( Optimised Wide-band Array) Yagi is the one to beat. Do we really want 300Khz of bandwidth with radiating efficiency dropping drastically 50Khz each side of the centre frequency?
Think long and hard before making your decision as to what type of antenna to build. If you are a contest station setting up on one frequency in perfect (weather) conditions once per year, with hundreds of watts to throw at the antenna (where losses of 10-20% don't matter too much), that all out performance maybe for you. If you are an everyday Ham looking for a good everyday Yagi with DX performance which is reliable and easy to build and requires no maintainance,
ALWAYS think 'OWA is the only way'!