The biggest discrepancies are between the vertical and flat positions, thanks to the effects of gravity on the balance and balance spring – Breguet’s thought was well, if I put the balance and its spring (as well as the escapement, including the escape wheel) inside a rotating cage that turns in the same plane as the dial, I’ll get a single, average rate for all the vertical positions. Daniels remarks, “The close vertical rates obtainable with a tourbillon will last longer than similar rates from a conventional watch.” He also points out, “The tourbillon watch is expected to keep better time than a conventional watch. If its potential to maintain a closer rate for a longer period of time is to be fully exploited, it should be fitted with an escapement that does not need oil at the impulse surfaces.” The reason for this is obvious – as lubricant deteriorates, the rate of the watch will vary.

Now, all this seems relatively straightforward. The notion that a tourbillon should be an aid to rate stability seems intuitive enough and yet it has become more or less a matter of conventional wisdom nowadays that the tourbillon is not a help – in fact, it’s often expressed that it’s more of a hindrance than anything else. Against the theoretical potential of the tourbillon, there are a number of practical arguments, which can be divided into arguments against the tourbillon in a wristwatch, and arguments against the tourbillon, period.

To take the second set of objections first, the tourbillon creates some issues while attempting to address another. You do get a single average rate for all the hanging positions, for sure, but you also need a more powerful mainspring because instead of the movement just moving the balance when the escapement unlocks, you have to move the mass of everything in the cage as well – balance, spring, lever, and escape wheel. On top of that you have to move the mass of the cage. Every single time the balance unlocks and the escape wheel advances a tooth, the whole kit and kaboodle has to move which is a major additional inertial load, as well as a lot of additional load on the pallets and escape wheel teeth. (If you want to get around this nowadays, and you’re making a tourbillon on the cheap, you can overcome lower manufacturing precision by throwing in a stronger mainspring; traditionally, however, making a tourbillon involved manufacturing to very high precision so as to avoid excess wear.)

In a wristwatch the objections become even more pronounced. The space issue is perhaps less urgent in a pocket watch, but in a wristwatch, the tourbillon really starts to hog available space; building a tourbillon into a 50mm or larger pocket watch is one thing, but fitting one into, say, a 30mm diameter watch movement is another. There were certainly tourbillon wristwatches built in the days before the post-quartz luxury watch boom, but they were very few in number – Patek Philippe made several wristwatch tourbillons for the observatory trials, as did Omega, but these are rarities, and the search for winning timekeeping solutions in the observatory trials was largely taken up with refining the lever escapement watch, and improving fine adjustment techniques. Audemars Piguet’s ultra-thin tourbillon watch, caliber 2780, from 1986, was intended as a demonstration of tour de force watchmaking prowess, not a serious contribution to precision timekeeping.

Finally, it’s doubtful whether, as the saying goes, the game is even worth the candle. We asked watchmaker Roger Smith, who’s known in the watch community for his advocacy of simplicity and robustness as traits of paramount desirability in a watch movement, for his views on the tourbillon in a wristwatch. He replied:

“I think George (Daniels) would have happily admitted that a tourbillon has no practical purpose within modern horology, and I would agree. The tourbillon was invented to cope with the vagaries of a split bimetallic balance, which due to its nature (two flexible arms) is impossible to poise, due to the varying temperatures, centrifugal forces and repeated shocks that a watch will receive. The tourbillon moderated these issues.”

“Today, the industry all use mono metallic balances which do not move once poised in the factory or workshop, and so for practical purposes relegates the tourbillon to the history books. Where a tourbillon does have a role today, is when a watchmaker is wanting to show off their skill, artistry and flare for miniaturized mechanics and to stand out from the crowd. I would still put one in a watch.”

It’s conventional wisdom that the natural home of a tourbillon is in a pocket watch but Smith makes an interesting and often forgotten point – the split balances used in pre-Glucydur balance watches, which were paired with steel balance springs, were extremely difficult to poise both statically and dynamically, and moreover would have any poise errors amplified by the expansion and contraction of the balance. In this context, a tourbillon could be effective, or at least it would seem so from observatory time trial records – a Girard-Perregaux observatory tourbillon from the 1890s, was recorded at the observatory at Neûchatel as having a difference in rate between the flat and hanging positions of just 0.19 seconds/day. But in a modern watch with a monometallic balance, the device seems superfluous at best and parasitic at worst, purely from a chronometry standpoint. Modern Rolex watches, to take just one obvious example, leave the factory adjusted to ±2 seconds per day, which a tourbillon wristwatch – certainly one of the same proportions as, say, a 39mm Oyster Perpetual –would be hard pressed to match.

Still, there are interesting outliers in the data on tourbillon wristwatches. The Concours International de Chronométrie was one of the few modern examples of a real accuracy competition, and in 2011, a Greubel Forsey Double Tourbillon Technique was the overall champion, with 915 points out of 1000 overall awarded – a Tissot, the winner in the Classique category, scored 764 points. The 2009 edition of the Concours also featured winning performances from tourbillons, with the JLC Master Tourbillon and Gyrotourbillon obtaining the highest overall scores. Such a small sample size, of course, provides little in terms of solid information on the benefits of the tourbillon and as with all time trials, may say as much about the adjusters as it does about the watch, but at least it shows that the device is not necessarily inimical to accuracy – which is not strong marquee, as they say, but it’s something.

A final point which is sometimes forgotten in the discussion of tourbillons in watches, is that the “single average rate for the vertical positions” is itself subject to instability. Having a single average rate is certainly a theoretical advantage, but that single rate is an average and itself subject to variation depending on the amount of power in the mainspring, the state of lubrication of the watch, and the amount of time it spends in varying positions where poise errors may be more or less exaggerated. The latter issue is the one that multi-axis wristwatch tourbillons are meant to address, and certainly, such tourbillons – those from Greubel-Forsey, for instance – remain fascinating as intellectual exercises. That may sound like damning with faint praise, but it isn’t – the intellectual interest of such watches isn’t for everyone (and certainly the price is not for every wallet) but they remain intriguing examples of the exploration of bleeding edge mechanical solutions to long term rate stability; experimental constant force escapements like Girard-Perregaux’s are another example of such experimentation, and I think the watch world is definitely richer for such things.

Asked if it’s true that a tourbillon could at least proved a theoretical benefit in a wristwatch – given its ability to produce a single average rate in the vertical positions – Smith replied, “A simple and well set up balance and spring will hold as good a rate in the verticals as any tourbillon. In the long term, a simple watch will far out perform a more complicated tourbillon, particularly as the lubricants begin to age.”

There is a certain finality to this assessment that brooks no dispute, but I still find tourbillons charming – given the chance, I would wear the Patek observatory tourbillon cased for Philippe Stern as my everyday wristwatch, efficiency be damned; I love the thing. It’s true that cheap, mass-produced tourbillons may seem to have deflated the tourbillon as a manifestation of craft, but I think they do so no more than cheap mass-produced watches deflate the value of a Dufour Simplicity – watchmaking nowadays is perhaps less about what you do, and much more about how you do it, and a beautifully made tourbillon is still as relatively rare and inherently interesting as ever.