The following was developed in response to a question from Richard Johnson who asked: What is your thought process when you are thinking about tacking on a header? How do you make the decision to go or not to go? To which I replied: It's a hard thing to totally nail down, and deciding when to tack often seems to be more of an art form than scientific method. But I have given it some serious thought, so here goes:

Going up the first beat, you get a nice, juicy knock (a.k.a. header). "Tack!" I hear you cry. And tacking would most often be the right move. But alas, not always. I can think of several situations in which you should not tack on that juicy header, tempting as it may be. A well executed tack in a dinghy takes only seconds, yet that little tack can have a profound effect on your finish position - and not always for the better. A tack without thinking is fraught with risk to your finish position (not to mention your boat!). So, before you make this potentially crucial move, there is serious thinking to be done. "We don't have time for all that thinking, we need to tack now!!" Well, the good news is: you can decide now, in fact you should decide - and most likely tack - now, because most of your thinking was, of course, done ahead of time. Do I hear you say "What thinking?" If so, the rest of this article is for you.

Strategic considerations: Position on the course must be a major factor in the decision-making process. Here is what Marc and I consider aboard SHADES:
1. In principle, we want to avoid reaching the layline any earlier than necessary. So, the first strategic question is: Are we nearer to the rhumbline (the middle of the course) or getting rather close to the layline? Is our tack-o-meter in the red zone? It is if we are nearing a layline but still quite a ways from the mark. The closer we get to the layline, the nearer we are to running out of useful options. And being out of options is bad news in any game. Being near or at the layline represents the danger zone as far as tacking is concerned. The closer we are getting to the layline, the more we should look for any excuse to tack and dig back in towards the middle. Conversely, in such a red-zone situation, we should be more and more reluctant to commit to a tack that will take us away from our present wise strategy of digging back in towards the rhumbline, and instead move us even closer to the dreaded layline. Thus, when we meet a shift near the layline, we already know whether we are eager to tack to dig back in towards the middle of the course, or will not want to tack unless we meet a truly exceptional shift that virtually demands a tack. (That same knowledge by the way, ideally governs our reaction - tack or bear away - if we meet a starboard boat!)
Of course, as we near the windward mark, we do have to get to the starboard layline. In general, we try to put this off as long as possible. If there are no other boats near us, we can wait until we are two or three lengths from the mark and our main concern will be not to have to tack twice in rapid succession and then immediately bear off onto a run, which would cause major speed loss. If we expect a bigger the crowd to join us at the mark rounding, then we have to join the layline "parade" earlier, of course, to avoid the major risk of getting cut off at the mark and losing lots of distance while we wait, or, God forbid, have to do a two-turn penalty.
If the shift hits when we are near the middle of the course and our options remain abundant, then the merits of the shift itself become more important. But wait, even here, there are still other, important questions to answer before we tack on our juicy shift.
2. Did our pre-race game plan call for playing one side of the beat in preference to the other? Before race 1 of the 2005 Midwinters for instance, Marc and I stood in the boat several times and came to the conclusion that the right side seemed to be getting more wind, and that we would try to play that side of the beat. Of course, such observations are not always reliable, and we are not so vain as to think we know it all. So we do not just bang the corner of what we hope will turn out to be the favoured side. No, sirree, we consider a further aspect of our position on the course as follows: 
3. Where are we in relation to the main body of the fleet or vis-à-vis the boats we especially want not to lose to in this race? In race 1 of the Midwinters, for instance, our preliminary estimate was that we needed to keep a specially close eye on Peter, Gale, Nick and Joe. Thus we would happily tack on a shift if doing so would help us to stay with our main competitors and/or between them and what we considered to be the favoured side of the course. And if we do tack, Marc and I try not to sit on anyone's wind if we can avoid it at no cost to our own progress. We Wayfarers have a wonderful "live and let live" approach to racing that I believe, adds greatly to the fun we have out on the water. Of course, late in the race, prudence dictates protecting whatever position we may be defending.
In any case, covering other boats in oscillating shifts, the kind we usually see in small-lakes racing, can be an unrewarding business and we usually find it preferable to "sail our own race" in such conditions while obeying with our strategic imperatives: avoiding the laylines, and protecting the favoured side of the course against the main body of the fleet.

By making all of the above considerations a fairly consistent topic of conversation between Marc and me as we go up the beat, we fulfill the first requirement: before any shift arrives, we know whether we are eager to tack, reluctant to tack or merely open to the idea of a tack. And this gives us a much better chance of correctly deciding whether we should tack or not.

 
Wind-related considerations:  Purely wind-related questions that that merit contemplation before we make our tacking decision include:
1. Will the shift last long enough to make a tack worthwhile? This can often be seen from watching what boats ahead are doing, and sometimes from watching wind patterns on the water ahead.
2. Is the header a "velocity header", i.e. have we sailed into a lull where our continuing faster boat speed from the previous stronger wind is combining with a weaker true wind and is moving the apparent wind vector further forward until the boat slows down? This is rarely a good time to tack and lose even more speed, when all we really want to do is ghost through the dead spot as quickly as possible - major gains and losses possible here! A subsidiary question then becomes:
Will a tack take us into an area that appears to be getting less wind or perhaps more wind? Ripples on the water, and again, other boats, may well provide valuable input here. It always pays to keep in mind that even the best of shifts cannot make up for a significant lack of wind. It may even be worthwhile, on rare occasions, to go against all the tried and true strategic odds, if by doing so we seem much more likely to keep the wind. And it doesn't have to be much - just more than the other guys are getting!
3. How big is the header? If we are in reluctant-to-tack mode, the header will need to be substantial to let greed overcome our fear of tacking, but if, on the other hand, we are eager to change tacks, then even the slightest and most fleeting of headers may be reason enough to go.
4. Is this shift a persistent shift, e.g. a new wind such as a thermal-lift-related on-shore breeze coming in, or a bend in the wind around a point of land? In that case - rather rare in our Wayfarer racing here in North America - it may pay to sail well into the header if we expect the wind to swing around even further. Then, a later tack will get us an even better lift on the other tack than the guys got who tacked immediately.

If none of the above factors help us to make up our mind, we remember that, everything else being equal, it's always a good safety play to get closer to the middle of the course. In other words, if there's a small header, we would be more likely to tack if a tack took us back towards the middle from our current position off to one side of the beat.

Other considerations: Since we North American Wayfarers sail many of our races in shifty winds, we need to be prepared to tack often and well. It behooves us to hone our tacking skills by practicing them until they are second nature and a joy to behold. Not the least of the benefits that accrue from top-notch tacking skills is the fact that we can now readily admit that our first tack was a mistake and tack back to our original tack with minimum pain. Of course, we all have days when every tack seems to have been a mistake, and then we tend to tack too often. But that is still better in the long run, than tacking and then sitting back and saying: "Well, that takes care of things."
I still vividly recall our many lovely long-distance races on North Bay's Trout Lake, where we would beat 5 miles into the prevailing SW breezes that were angled such that you could sail a long port tack nearly parallel to the north shore. Eventually you would need to tack to starboard to avoid hitting the shore, and this is where we regularly "made our money". Most of our competitors would, having tacked to avoid the shore, proceed a long way on starboard tack, some of them more than a mile to the south shore, for no good reason. We, on the other hand, would sail out 50-100 metres, and then tack back to port, and repeat the procedure as necessary. All this time we would patiently await the inevitable port-tack header that invariably occurred a few times during the long beat to the west end of Trout Lake. Every time that header hit, we would be ready and waiting to sail the starboard lift (20-30° most times) for as long as the shift lasted. And every time, we gained a fair bit. The moral being: If you are forced to tack away from the favoured tack, or you realize that your tack was a mistake (because the shift only lasted a second or two for instance), get back to the favoured tack as soon as is reasonably possible.

Related reading: The Fine Art of Pinching (see http://www.wayfarerinternational.com/Pointing.htm)