Path Dependent Perils

We don’t speak enough about path dependency.

The order in which you learn things can, sometimes literally, make or break you.

“What is the highest launch elevation you’ve done?” I asked a North Carolinian housemate this past week. “Probably seventy-five hundred feet” came his reply. “Your intuition about ground speed on launch will lie to you” I cautioned.

We were, you see, attending Utah’s annual Red Rocks fly-in where the Monroe Peak paragliding launch resides in the rarefied air of 11k’ MSL. Your wing, we should note, cares not how fast you are zipping over the ground when generating lift, but rather “indicated airspeed“, a quantity that hinges not just on the rapidity with which you are traversing an air mass but also its density. Your brain, meanwhile, loves to conflate cause and correlation, often in ways you won’t notice until it is too late.

If you have trained predominantly in the viscous air of lower altitudes, you may imagine during a takeoff run that a given ground speed will soon yield the lift you need to safely launch. This may have worked as an approximation at mild altitudes and yet in the high mountain desert reality of Utah this could cause you to run off a cliff and into a dangerous dive as your wing seeks trim speed in the surprisingly thin air. The inputs that won’t lie to you, and on which you ought to rely, include instead how pressurized the wing feels through the risers attached to your hips and the control toggles in your hands. You have to feel launch-readiness, not see it, if you want a universal mechanism.

The next day, after expediting a landing when the overhead cloud began to rain, I found myself in the LZ chatting with an Alaskan native visiting Utah for the first time. “How much precipitation should I be concerned by?” he asked. “Well, out of respect for Utah’s climate, I landed as soon as I got sprinkled on” I replied. I mentioned risks around airflow separation and the accumulation of water in a paraglider’s cells.

I then elaborated on how the humidity of one’s home environment tampers with one’s intuition of how concerned one ought be by a cloud of a given size, explaining that the more arid an environment the more powerful a weather phenomenon you are witnessing as evidenced by a given cloud size. If you learn in Utah, an Alaskan cloud might scare you out of a flight that would have been lovely. If you learn in Alaska, a Utahn cloud might seem inviting until it tries to kill you.

In both cases your safety can prove dangerously dependent on the path of your learning. In one direction you find yourself in the harsher environment first and thus well protected later, the worst outcome being an overly conservative approach that you can dial up later. In the other direction you find yourself in a deceptively easy environment at first and an unforgiving one later, possibly with inadequate margins and catastrophic consequences.

I may have lost count of the number of times that traversing a learning gradient in the dangerous direction of a particular facet nearly punched my paragliding ticket in a permanent way. But, for funsies, let’s take a trip down memory lane to explore a few, some where I was progressing in the safer direction and some not.

The majority of my earliest training took place on the “southside” of Point Of The Mountain. There the risks around venturi, the tendency of a fluid or gas to accelerate as it transits a constriction, generally prove relatively benign, owing variously to the local wind flows usually making it a morning site as conditions power up as well as the site topography rendering it simpler and more protected. Contrast this to the “northside”, the companion site on the Salt Lake valley side, generally flown later in one’s progression and in the evening when catabatic flows out of the Wasatch Front drain into the valley and accelerate at the pinch point where POTM exists, and in hindsight I see how inadequate appreciation of a paraglider’s vulnerability to venturi, a phenomenon that often precedes wind shear and mechanical turbulence, nearly killed me.

I was on the very first circuit of my very first ridge soaring flight on the “northside” when I waited too long to initiate my 180 at the west end of the lower bench and found myself parked in the air. Suddenly my wing began to wobble and I sank precipitously into what I gather was the lee of a wind blowing across the site, ending up on a small ledge where after landing I got plucked, slammed, and dragged wildly as I attempted to subdue and disconnect my wing. That was probably the most task-focused I have ever been in my life. By the time I unclipped one of my carabiners the wing was wrapped over a tall cliff and had a brick-sized rock in it. Having prevailed, I celebrated by purchasing a hook knife, and a few days later received my initial certification. Survive a few days like this and a guy might start developing theories about Quantum Immortality.

My next most memorable misadventure happened just a couple months later and bookends thematically with an experience from this past week. In many situations, the only correct levels of commitment are 0% and 100%, something impressed upon me as a baby bird by botching a “northside” launch in light and variable winds wherein I caught my foot in a bush and then tumbled wildly down a steep slope. As horrifying as the moment was subjectively, I mostly got a pass by virtue of my pummeling being at the hands of soft dirt and fragile bushes, not unforgiving rocks and electrified cables.

By the time I was launching “Cove” in Monroe a couple days ago I blessedly had inculcated instincts to launch aggressively and be ready for anything, allowing me to smash a bush, dodge a boulder, and sail over power lines with a wide margin despite a reconnection with the ground on launch when the wind suddenly went cross. I wish the pilot who handled a similar situation three days earlier with a less excellent outcome could have flubbed a “northside” launch like I had and thereby attained a less expensive lesson. Some launches are less forgiving than others. Better to play above soft ground before hard.

One of my worst periods came in the spring of 2021, also just months after my initial certification, perhaps owing in large measure to my being on the dangerous version of the path along seasonal transitions. The late Utah fall in which I learned is relatively mild, and the Utah winter milder still, just the recipe for over-confidence going into the spring where a surge of moisture coupled with greater insolation often results in shockingly punchy conditions. Why, I wonder, don’t we do a better job warning pilots about psychological traps attendant the time of year that they began their training?

That March I compressed two SIV clinics worth of “maneuvers” into just a few seconds with no warning — while at the “southside”, heading east at ~100′ over the east end gash that separates the hang glider and SuperFly areas, I took a massive asymmetric collapse on the right side, looked up at it like the clueless idiot I was, induced a 225-degree spiral dive, avoided smashing into the hill with an amount of left brake that I gather fomented a spin followed by a deep stall, brake-checked the surge so late that I could see above my wing the berm of the gash that precipitated the mayhem, and then just barely avoided hitting said berm during the swoop of my wing re-energizing. I don’t have to imagine what failing to check that surge would have done since six months later during an actual SIV I botched a stall recovery and Superman’d through a free fall, just barely slapping the wing over my head to avoid being gift-wrapped.

Then just three weeks later, laboring under the weight of a deep fear injury, I completed what I now call the “deflation-to-stall story arc”, wherein terror at the thought of another deflation led to very heavy hands on the brakes which when coupled with thermic conditions while flying a low-performance glider highly prone to steep pitch-back on thermal entry culminated in a stall about 100′ off the ground (yes, a FULL stall). The last thing I remember prior to picking myself up off the ground is my wing deforming in a terrifyingly novel way and thinking “ahhh this is it”. The concussion from impact wiped the moments immediately preceding it but in piecing the story together from a pilot on the scene and the all-sides scuff marks on my helmet I gather I managed to ride either a back-fly configuration or a parachutal stall all the way to the ground and impact on a tumble-favoring slope in something vaguely resembling a Parachute Landing Fall. Score one in each column for path direction luck: a dangerous entry into the annual cycle of seasons but saved again by the soft ground of the baby bird nest. Not exactly a cheap lesson, per se, but also survivably inexpensive…

The second-hit fear injury would cut deep enough that I couldn’t reliably enjoy flying again until reaching cloudbase in Colombia nearly nine months later. The three SIV clinics I attended as penance in 2021 may have made me an objectively safer pilot but my brain didn’t believe it. Often the only way to beat recency bias and negativity bias is to slog through a long valley of despair as you pile up positive experiences. You can’t reason your way out of PTSD and you certainly can’t predict how long recovery will take.

So-called Beginner’s Luck is real but not for the superstitious reason most people ascribe to it. Really it’s a function of survivorship bias — most people who get walloped early quit, they aren’t around to tell you their story, and that’s a lot of people in the sport of paragliding. Once you have finished scrolling all the fabulous paragliding videos on Instagram, do as Nassim Taleb counsels, and “now take a look at the cemetery”. Perversely, in a sport so complex, consequential, and random as paragliding, Beginner’s Luck may actually not be so lucky after all, putting one on a dangerous luck gradient — those suitably chastened early may come to hew more conservative earlier in the learning process than the relatively “lucky” pilots who, emboldened by their implausibly fortuitous dice rolls, take ever greater risks until a bill comes due for an unaffordable tab.

Every time I see a pilot mortified, maimed, or killed I seek to avoid the “wouldn’t happen to me” trap and instead think “there but for the grace of God go I”. I remember a couple months into paragliding wondering if I had finally experienced all the ways you could be surprised in this sport. Some time later I accepted that unless you stop progressing (or quit) you will never stop finding ways to be surprised.

How about that time I went to Alaska in 2022? There I discovered the path dependency regarding terrain stickiness. I learned in Utah where rocks, weeds, cacti, and sagebrush are forever ensnaring you in infuriating ways. You have but to land in a mild crosswind to put your wing down in a mess that could take fifteen minutes to disentangle. But the slick hillside tundra plants of Alaska? Holy smokes can you end up feeling like a giraffe on ice skates attached to a wing that wants to get under your feet for a very different kind of “sled ride”.

My very first flight at the Lake Hill launch of Hatcher Pass saw me falling repeatedly while attempting a takeoff from a steep, plant-strewn, high-wind face that ultimately involved getting plucked with a cravat in my wing. That might have resulted in tragedy with, say, a spin back into the hill or another low-altitude stall but instead was largely a non-event as I had already practiced cravat clearing during SIV — every chance you have to familiarize yourself with a scenario in SIV is a chance not to maim or kill yourself in the wild when you encounter a new-to-you wing fault over hard ground. That’s a nicer path. I only wish I had done my first stall on purpose and in SIV instead of 100′ off the ground at the “southside”. Can’t win’em all.

How about terrain complexity? When I found myself in Woodrat in April of 2023, most of my previous XC flying had been in places where there is essentially a single-direction extended ridge perhaps punctuated by the occasional valley crossing — Utah’s Wasatch Front, Colombia’s Piedechinche and Roldanillo, Italy’s Bassano, and most of what I had effectively flown in Alaska. Woodrat is different. The instant you launch you are playing in the complex terrain of many tightly connected small mountains and small valleys which conspire to generate thermals and channel wind in complex ways.

Across the valley below the main launch runs another perpendicular valley known as “China Gulch” which has a reputation for eating pilots. Despite reading the site guide and noting that “having to land in China Gulch has resulted in many tree landings and injuries over the years” I still got thus eaten though thankfully without harm to the limbs of either human or tree. I found a field in which to land just outside the red “Do Not Land” square in the FlySkyHy app. But let’s be honest — I didn’t give a rat’s ass in that moment about red squares on my flight deck, just staying out of trees and avoiding rotor so as not to take a big collapse next to the ground. I landed as downwind as I could in the field yet still had a scary wobble on short final. Yikes. Cheap lesson learned. Barely.

Thinking of a meta issue, perhaps big breakthroughs after long struggles represent particularly dangerous moments. I have noted previously that “if paragliding were an adventure role-playing game on a computer then you might get notifications that looked like ‘skill unlocked — fifty additional quests available; ten new ways to die'”. Such breakthroughs in one area, preceded by progression stagnation in others, can lead to weird learning gradients that present surprising risks.

For me that manifested notably by having an early crash, sustaining a deep fear injury, and dialing my subsequent ad hoc flying way back, coupled meanwhile with multiple SIVs greatly dialing up my wing handling competence, and yet self-limitation (especially in Utah) keeping me from developing an intuition about how thermals generate not just turbulence (now less of a concern) but also highly unpredictable glides. This got me into serious trouble at least twice before I realized what was happening to me.

On one occasion, at an SIV in Oregon, I found myself dealing with an extremely tight LZ on a very thermic day. I was not prepared for a tailwind generating thermal to pop as I was pointed toward trees during an S-turn to bleed off altitude before touching down. I just barely managed to avoid playing a life-altering game of tetherball with my body as ball, the margins so slim that the left tip of my wing grazed a tree while I was maybe 50′ off the ground, after which I had to “dance” through brush on short final in a manner where I just barely managed to avoid impaling myself on a pike-like dead tree. Ah the irony of training intended to make you safer very nearly wrecking you.

On another occasion, toward the end of a short flight at The V in Centerville Utah, I failed to appreciate the combination of marginal altitude plus a strong northwest wind when I made a committing crossing of a dense residential area, finding myself continually reassessing which tree was my best “LZ” until finally and implausibly I made it to the main LZ at the school. I note with some irony that my flight at the same site on the previous day was not just far better but also much safer as the result of the school LZ being shut on account of school being in session — not only did this preclude the committing glide but also made me more willing to head back to the hill’s lift band and fight to stay up, which I did instead of faux-conservatively folding the hand and playing FAFO over a bunch of houses, trees, and power lines.

I could go on digging up examples but perhaps let’s close the paragliding narratives with one last one. At the aforementioned fly-in, another of the North Carolinian crew with whom I was sharing a house had a fun flight right up until the landing that had me loaning him a handful of Advil. “I didn’t appreciate that the wind could be in such a different direction with so little height change so close to the ground”, he remarked of his shoulder-punishing yard-sale of a surprise tailwind landing.

That struggle is real but probably you’re on a less surprising gradient if you learn of that possibility toward the beginning of your paragliding life in a high mountain desert area like Utah as opposed to climates and terrain less likely to generate dangerous wind shear. But does less surprising mean less dangerous in this case? Maybe… though I’m less sure on this one than others.

I can think of a September 2023 flight, a less eventful V flight that immediately preceded the aforementioned one, whereupon landing on the fire road I seemingly caught a ripper of a thermal at the last moment but “luckily” protected my shoulders by landing on my head. Yes, actually, that often is the preferable thing to do, at least if you’re wearing a helmet and think to twist your head to avoid a full-on faceplant as well as manipulate your body in a way where your head contacts last after several other body parts. Give yourself a hug, say a prayer, and do your best to distribute the deceleration over both time and tissues.

When similar happened to me just a week later in Monroe I was more ready for the surprise tailwind and “raised the landing gear” for a butt-landing that I’m sure my harness didn’t love but left me unscathed, if choking on the dust my tumult had kicked up. There is probably another terrain complexity story here as well — the more varied the structures around your would-be landing spot, the more likely a hyper-local thermal will tag you at an inopportune moment, and the confluence of desert, crops, roads, and power lines made this one a dice roll I perhaps ought have avoided.

Switching gears, what about, as is the case with many in the paragliding community, if you fly multiple kinds of wings? I myself have had a funny progression with this — a few years before learning to paraglide I made an abortive attempt to get my Private Pilot License for airplanes, an experience disrupted in 2017 by my instructor getting killed while out with another student, and then I got back to airplanes nearly five years into my paragliding epoch. Consequently I have found myself paradoxically learning airplanes at a basic level both before and after getting somewhat proficient at paragliding.

How much easier it is to do a power-on approach after hundreds of landings in an aircraft without an engine! It’s surreal to recall how much I struggled with this in my earlier airplane epoch. Last month the instructor who did my pre-solo checkride remarked that I had “very good ground awareness”. As it turns out, a bunch of splatter landings in an “aircraft” where your landing gear is your legs is great motivation to develop that awareness!

I have, meanwhile, observed pilots deep in airplane backgrounds and relatively new to paragliding do things on paragliders that make me wonder about the risks attendant one’s order of exposure to free flight and powered flight.

In an airplane, apart from an engine-out situation, you nearly always want to follow a very long stabilized approach to landing, applying continuous small adjustments to pitch and power to maintain your glide as you encounter the vagaries of dynamic air, and pushing in the throttle for a go-around if you float too long.

On an unpowered glider, however, there is no “power” to apply, so on an approach to a tight field in dynamic conditions the correct play involves bleeding off altitude either upwind or over your LZ, often flying just a touch above obstacle height at field’s edge shortly before touchdown to simultaneously manage the risks of both undershoot and overshoot alike.

Fly the aircraft you’re in with respect for its unique properties! I myself have at times on a paraglider forgotten which aircraft I was flying but thankfully been chastened cheaply as I figured things out. Never run out of altitude, airspeed, and ideas, and definitely don’t get into the habit of landing a paraglider like an airplane. You might get away with it much of the time but eventually it will yield a trip to the repair shops for both people and paragliders. Beware so-called “negative transfer” — experience in one aircraft type may not merely be non-applicable to another aircraft type but actually make things more dangerous.

Lastly, we would be remiss in not mentioning that the perils of path dependency are limited not just to the direction of experiential learning but also the steepness of the gradient and furthermore how many parameters you are modifying simultaneously. Want to reduce the risk of surprise expensive education? Dial up just a single parameter by itself and by a small increment. Better still, dial up such a parameter on purpose instead of by accident. Want to protect yourself from yourself even more? Purposefully dial up a single parameter while deliberately dialing down other parameters to increase margins for error while reducing cognitive load.

Do this with some degree of reliability and you just might survive an aviation odyssey of your own. Apply this methodology more broadly to your life and you may even have a happier story arc in your career and other aspects of your life.

I left federal service in 2016 after eleven years at the DoD as a civilian in the cyber-security space, then spent about three years at a hedge fund with a strong security focus, then spent about a year as an executive at a seed stage startup, and when the COVID-19 pandemic struck I set out on my own to be a solopreneur consultant for an assortment of startups. I’d love to tell you that I consciously, deliberately, and brilliantly navigated those transitions by shrewdly dialing a carefully limited number of variables by a carefully calibrated amount at each step, from company size, sector, and maturity, to job role and location, to social graph overlap, and so on, but that would be an exercise in retrofitting self-serving narrative onto happenstance. At best I maybe had some good intuitions. In fairness I may have been just barely skiing from one crazy situation to the next with only the scarcest of a plan in my head.

What I can tell you is that I have watched many people over the past decade blow themselves up professionally because they didn’t twiddle all the knobs in a manageable way to attain “optimal distance” from one circumstance to the next — the sweet spot of uncomfortable enough that you know you’re learning but not so far outside your zone of competence that you are destined to crash and burn.

I saw muckety mucks from Big Government and Big Tech alike go to the aforementioned hedge fund and struggle because they lacked the troops to command and also had lost the ability to pick up a rifle themselves. I saw previously technical people whose skills had long since atrophied working as Project Managers or Team Leads in established companies think they could dust off those skills by going straight to a seed stage startup with a five-person engineering team only to crash the car. I’ve lost count of the times I’ve seen someone choke on complexity when their most involved previous project was a web app running on a single server in a corporate data center and they went straight to a microservices architecture on a public cloud without considering whether they even needed that, much less were up to the task.

And I, meanwhile, have survived my aviation journey thus far without any seriously life-altering consequences yet while relying on chance to a degree that I frankly find uncomfortable to consider. I hope that I’m past the steep part on the left side of the bathtub graph, but I’m not sure, and beyond there anyway lies the dangers of complacency. What I really hope as I finish the swing on this piece is that others might learn a few things that make them less reliant than I have been on the most fickle of friends — luck.