Avoiding knight forks

Though I try to vigiliantly watch for places where my pieces could forked by an opponent's knights, and make adjustments accordingly, I still find myself losing material to this tactic again and again (I have also won my share of material by employing it). Though I am usually able to avoid knight forks in the more common scenarios, such as when two of my pieces are on the same rank either one or three squares apart and my opponent has a knight withing striking distance, it happens in other situations where it would seem very difficult to forsee. These include when two pieces are on the same rank -- again, separated by either one or three squares -- or even more cryptically, when arranged in the pattern of two parallel, overlaping "L's" lying back to back (e.g. on a1 and e3) or two perpendicular, intersecting "L's" (e.g. on a1 and d4). There are probably still other complex patterns in which knight forks can occur that haven't even conceived of.

Anyway, as I said before, I am constantly losing material and having to resign games to this tactic and I want to know, is this, to some extent, inevitable? Do even high-level chess masters occasionally get "forked" by an opponent's knight, or does it only happen to amateurs? Are only computers able to predict with complete accuaracy how and when a knight fork might occur if the targeted pieces are arranged in a complex pattern?

I am at the point where I am ready to just start targeting my opponents' knights for "immediate termination" (i.e. trading them for a minor piece) at the beginning of the game. 

I know this isn't quite what you're talking about, but here's an example I saw in the forums of Karpov getting forked by a queen and resigning: http://www.chess.com/forum/view/fun-with-chess/karpov-loses-in-12-moves

I know what you mean about the frustration of a knight coming seemingly out of the blue and undermining all my beautiful plans for the destruction of my opponent.  I don't know of a foolproof way of avoiding forks except to make scanning both knight's available options, maybe two moves out, just part of what I do each move.  I do that sometimes, but not regularly enough to avoid getting forked regularly.  I think that in order to regularly avoid being forked over by an opponent scanning the knights' moves and becoming comfortable with their attacking style just has to become part of the chess program in my head.  Until then getting forked will just be a regular and unhappy surprise for me.

A.

Knights are a tricky piece to get the hang of. At one point this was what I did to help. If you put a knight in the middle of an empty board, it can go to 8 places. By looking at one place the knight can go, and comparing it to all the other places you see all the possible knight forks that can happen ... does this make sense? What I mean is one possible knight fork is when 2 squares are next to each other on a diagonal. Knight fork number 2 if you start looking around the knight moves could be two squares with just one square inbetween on the horazontal or vertical. Let me see if I can make a diagram.

edited for clairity

Knights always jump from one color square to the other.  If you're really bad at spotting knight forks, you can ensure your safety from them by keeping your major pieces on opposite color squares.  Obviously you can't do this all the time or else it hurts your game, but if you think a knight fork is coming and you can't envision a way out, it will work as a safety principle.

Also, a knight cannot fork two pieces separated by two squares diagonally.

I think we are wording things badly. What I meant was 2 squares right next to each other on a diagonal.

The same color thing is a good point. I hope the diagram helps.

Thanks for the help. The diagram was particularly useful. How do you make diagrams and post games, by the way?

I've had my fair share of blunders due to an overlooked knight. I think it helps too to look for patterns within your prefered openings. You have to learn not only your openings' strengths, but also their weaknesses. As an example, I often play the bird. However, when starting out I frequently would overlook the weakness of my d pawn after developing. I now have grown accustomed to defending the previously weak g4 square:

since knight can go max two square for each direction(up or down), keeping 4 squares between the pieces(just in one direction) would be enough

Checking all possible knight moves two turns away seems like a very mentally expensive solution. A more economical solution is to work backwards instead of forwards:

1. Identify tactical targets for your opponent
2. If there are more than 2 on the same color, then check if a combination of targets on the same color can be connected with a knight move

I have been victim to various forks when it comes to knights, theres seems to be a few rules to take into consideration when dealing with knights:

1. The next move a knight will employ will always have it move to a square of the opposite color, meaning if its on a light color square, it will always jump to a dark color square and vice-versa

2. If you see a knight will fork you in the next move and a major piece (rook or queen) is under attack, it is best you create a counter attack that will make the oponnent think before moving the knight, equalizing the threat on the board is the best way to have knights back off, as opposed to attacking the square the knight will land for a new fork or defending the pieces that are already under attack

If your pieces are near each other and on same-color squares, with some exceptions, they can be forked. (Exceptions include the pieces being diagonally adjacent, such as on c3 and e5.) Make sure that potential forking squares are protected if you'd like to keep your pieces where they are.