Jon W. Hansen, FCIPS | Procurement Insights | Hansen Models™
A measure that looks successful during a crisis, but doesn’t hold, raises a hard question: was it ever a success?
On 9 October, hours before Hurricane Isaias made landfall on the Gulf Coast, the U.S. administration announced that Russia would make more than 300,000 metric tons of diesel immediately available to U.S. and global markets. Wholesale diesel fell more than 4 percent in a single session.
The surface calmed. The current did not.
Four weeks later, no cargo under the deal has reached a U.S. port. Of the three tankers reported as U.S.-bound, two have changed course, and the third has had no confirmed discharge port since 26 October. New Jersey’s governor has urged port operators to turn the diesel away. Meanwhile, the U.S. retail diesel price, which stood at $6.179 a gallon in the EIA survey of 12 October, has been above that level in every weekly survey since.
I call this the Riptide Effect.
What a riptide does
At the beach, a riptide is dangerous precisely because the water above it looks calm. People relax. They swim further out. The force pulling them away hasn’t gone. It has only been hidden.
A crisis measure can work the same way. It steadies the surface at the acute moment, when everyone is watching. In this case, it steadied the price of diesel on landfall day: wholesale diesel fell more than 4 percent in a single session. Within two weeks, retail diesel had risen 17 cents a gallon, and it remains above its October level. Whether the announcement only delayed that rise is what our lab is now testing. If it doesn’t change the condition underneath, the force reasserts itself later, after people have stopped paying attention, and sometimes more sharply than before.
That does not make every temporary measure a mistake. There is a difference between a bridge and a mask:
A bridge buys time to change the underlying condition: real supply, delivered, while inventories are rebuilt or new sources are secured.
A mask buys time and spends it: an expectation that steadies the market without changing what produces the price.
So three questions decide whether a crisis measure was a success:
Was it delivered, or only announced?
What was the time it bought used for?
What decisions were made on the calm it created, and would they have been made if people had seen the current?
It keeps happening
The diesel announcement is not an isolated case.
In January 2025, a ceasefire in Gaza took effect, and the Houthis said they would limit their attacks on shipping. The surface calmed. The ships did not come back. Suez Canal transits fell 21 percent from January to February, and by June, traffic through the Red Sea was still about 60 percent below its level before the attacks.
In April 2026, a ceasefire was announced on 8 April. The next day, the head of ADNOC said: “The Strait of Hormuz is not open. Every day the Strait remains restricted, the consequences compound.” That week, I was updating an exposure assessment for an Ottawa diagnostics company, and recorded exactly that: the ceasefire had not reopened the strait. Seven months later, the strait is still effectively closed.
And this week, an independent scorer confirmed that the waiver behind the diesel deal remains in force. On paper, Russian diesel is authorized to reach U.S. ports. In practice, none has arrived. The record stays “confirmed” while the outcome drains away. That is a riptide on paper.
How the strand moved from one lab to the next
This did not come from hindsight. It came from following a strand.
In October, our Isaias Strand Lab locked its predictions before the storm made landfall and tracked them daily. One prediction held that, with no slack in the diesel market, even a short interruption would move prices. On the official survey, diesel eased 2 cents after landfall. We scored that prediction not demonstrated, and we recorded something more important: the lab had missed the diesel announcement entirely. A policy action on the price strand, made on landfall day, never entered our daily record. We wrote that down as plainly as we wrote down our hits. We also recorded an analyst’s caution: announced shipments are not delivered supply.
That missed strand did not close with the lab. It became the starting point of the next one.
Our Isaias Reverberation Lab opened on 17 October to follow the policy strand forward, with seven dated expectations locked on 17 October, to be scored by an independent scorer, never by their author. This week’s check-in recorded the cargoes diverting, the refinery strikes resuming, and the price of crude falling sharply. And on 5 November, before the evidence was in, we recorded the Riptide Effect as a hypothesis, with a way to test it: if the announcement’s relief is unwinding, it should show first in wholesale diesel, before it reaches the pump.
It may hold. It may not. Either way, the record will say so, because the test was written down first.
Outcomes are a journey, not a destination
The Isaias lab judged its price prediction one week after landfall. The riptide only became visible at four.
That is the lesson I keep returning to. An outcome measured at the acute moment is a snapshot of the surface, not a verdict on what happened. A crisis measure can look like a success on day three and a failure on day thirty. A technology implementation can look like a success at go-live and a failure eighteen months later. If you stop measuring when the surface calms, you will mistake the calm for the outcome.
I have been skeptical of this kind of snapshot for a long time. In 2007, Ariba’s PR firm invited me to interview a senior executive about a new contract win. The release cited 108 percent growth in its on-demand client base, with nothing to compare it to. I wanted to know whether the new model would cannibalize the old one. Ariba couldn’t answer by the end of our third interview. Press releases and case studies share the same weakness: they measure at the moment the surface is calmest.
It is why we track outcomes over time, long after the event that prompted them. True outcomes are a journey, not a destination.
When you stop reading, the swan changes color
There are three more things a riptide teaches, and our labs have shown each of them this week.
First, not reported does not mean it did not happen. In our Cold Chain Strand Lab, we have so far documented nine food-sector cyberattacks from public sources. Every company reported what happened and when its systems came back. Not one record showed who had the authority to act, or whether they acted in time. Those decisions were made. They simply were not recorded anywhere the public can see. Aviation solved this problem decades ago with flight recorders: the forces on an aircraft are the same whether or not anyone is recording, but only a recording lets investigators see the sequence of decisions clearly. Most organizations operate without a recorder for their own decisions.
Second, the color of a swan depends on how often you read the strand. I have long defined the difference this way: the difference between a gray swan and a black swan is not the event itself. It is whether the signals that preceded it were in the scan. Timing adds a second test. A strand can be in the scan and still arrive as a surprise if you read it too rarely. Our lab reads its strands weekly. It recorded one tanker’s change of course ten days after it happened, and the resumption of strikes on Russian refineries six days after. Neither was a black swan, because both strands were being watched. But if the time it takes to notice, plus the time it takes to respond, is longer than the time a strand takes to act, the event arrives as a surprise however well the strand is understood. And because there is no time limit on when a strand can act, the riskiest decision of all is to stop reading it. That is when a gray swan becomes a black one.
Third, a strand does not have a single point of impact. It has an impact path. In the 1998 engagement, once we found what time of day the orders were arriving, we aligned delivery with the objective. But delivery was the flashpoint, not the end of the strand. A wrong part shipped, a faulty part, or a wrong part ordered belonged to the same strand, and only showed itself further along. Measured at the delivery point, the strand looked solved. The same is true of the diesel deal: at the point of authorization, it was confirmed. At the point of delivery, the cargoes were turning away. Where you take the reading decides what you see. In 1961, President Kennedy set the goal as “landing a man on the moon and returning him safely to the earth.” Success was defined at the end of the path, not at its most important, most dramatic point: the moon landing.
Three domains, one pattern
Step back, and the examples in this post come from three domains that rarely appear in the same conversation:
Energy policy: a diesel deal announced, authorized and confirmed on paper, and never delivered.
Maritime shipping: two ceasefires, in 2025 and 2026, that calmed the surface while the ships stayed away.
Food-sector cybersecurity: attacks reported in detail, while the authority to act went unrecorded.
The industries differ. The pattern does not. In each, what was recorded diverged from what was actually happening, and only following the strand over time showed where.
Three domains, one pattern. In energy policy, maritime shipping and food-sector cyberattacks, what was recorded diverged from what was actually happening. Illustrative.
That is the claim at the heart of Strand Commonality™: the same strands, crossing in the same way, across domains that have nothing else in common. It is also why we call the principles behind it Invariant Physics™. The technology, the industry and the headline change. The forces that decide the outcome do not.
Why the strands must be recognized and monitored
The diesel announcement was not an isolated event. It crossed the hurricane, the war in Iran, the Russia–Ukraine conflict, sanctions policy, the U.S. midterm elections, and the decisions of individual port operators and state governments. No single function, desk or model following one of those strands alone would have seen the riptide forming.
This is what Strand Commonality™ is for: recognizing which strands are moving together, and continuing to monitor them after the moment that made them visible. Recognition without monitoring gives you a good call at the acute moment, and no way of knowing whether it held.
And here is the most important point. A riptide is not a single event or a rare anomaly. It is a regular occurrence. Announcements outrun delivery. Press releases outrun results. Go-live outruns adoption. Every crisis produces measures that calm the surface before they change the current. That means riptides can be managed, not just survived. Once you have identified all of the strands that require your attention, you know where the current is likely to run, what to watch, and when the calm is real. What turns a riptide from a hazard into something you can manage is not predicting it. It is having already identified the strands it runs through.
That is what we do. We lock what we understand before the outcome is known. We follow the strands forward, not just to the headline. We record our misses as plainly as our hits. And we keep measuring until the outcome, not the surface, tells us whether we were right.
That is also why we run our labs in real time rather than relying on case references. A case reference is a snapshot too: it proves we were right once, at a moment of our choosing. A live lab has to keep proving it, on dates fixed in advance, in the open, including when the result goes against us. Like the principles we call Invariant Physics™, our reliability and the accuracy of our coverage are never proven once and for all. We have to keep proving them.
Across our labs, the output accuracy of ARA™ RAM 2025™ running on SLAP OS™, assessed against the evidence, is holding at an average of around 91 percent, and it fluctuates. That is a different measure from our prediction record, which includes misses like the Isaias diesel prediction. The point is not the number. It is that tracking both the hits and the misses gives us an accurate picture, and an accurate picture is what drives the Infinity Loopback: the continuous learning of our algorithms.
Truth Is Believing. Accuracy Is Knowing. Outcome Is Proof.™
The Riptide Effect: When a Crisis Measure Calms the Surface but Not the Current
Posted on November 6, 2026
0
Jon W. Hansen, FCIPS | Procurement Insights | Hansen Models™
A measure that looks successful during a crisis, but doesn’t hold, raises a hard question: was it ever a success?
On 9 October, hours before Hurricane Isaias made landfall on the Gulf Coast, the U.S. administration announced that Russia would make more than 300,000 metric tons of diesel immediately available to U.S. and global markets. Wholesale diesel fell more than 4 percent in a single session.
The surface calmed. The current did not.
Four weeks later, no cargo under the deal has reached a U.S. port. Of the three tankers reported as U.S.-bound, two have changed course, and the third has had no confirmed discharge port since 26 October. New Jersey’s governor has urged port operators to turn the diesel away. Meanwhile, the U.S. retail diesel price, which stood at $6.179 a gallon in the EIA survey of 12 October, has been above that level in every weekly survey since.
I call this the Riptide Effect.
What a riptide does
At the beach, a riptide is dangerous precisely because the water above it looks calm. People relax. They swim further out. The force pulling them away hasn’t gone. It has only been hidden.
A crisis measure can work the same way. It steadies the surface at the acute moment, when everyone is watching. In this case, it steadied the price of diesel on landfall day: wholesale diesel fell more than 4 percent in a single session. Within two weeks, retail diesel had risen 17 cents a gallon, and it remains above its October level. Whether the announcement only delayed that rise is what our lab is now testing. If it doesn’t change the condition underneath, the force reasserts itself later, after people have stopped paying attention, and sometimes more sharply than before.
That does not make every temporary measure a mistake. There is a difference between a bridge and a mask:
So three questions decide whether a crisis measure was a success:
It keeps happening
The diesel announcement is not an isolated case.
How the strand moved from one lab to the next
This did not come from hindsight. It came from following a strand.
In October, our Isaias Strand Lab locked its predictions before the storm made landfall and tracked them daily. One prediction held that, with no slack in the diesel market, even a short interruption would move prices. On the official survey, diesel eased 2 cents after landfall. We scored that prediction not demonstrated, and we recorded something more important: the lab had missed the diesel announcement entirely. A policy action on the price strand, made on landfall day, never entered our daily record. We wrote that down as plainly as we wrote down our hits. We also recorded an analyst’s caution: announced shipments are not delivered supply.
That missed strand did not close with the lab. It became the starting point of the next one.
Our Isaias Reverberation Lab opened on 17 October to follow the policy strand forward, with seven dated expectations locked on 17 October, to be scored by an independent scorer, never by their author. This week’s check-in recorded the cargoes diverting, the refinery strikes resuming, and the price of crude falling sharply. And on 5 November, before the evidence was in, we recorded the Riptide Effect as a hypothesis, with a way to test it: if the announcement’s relief is unwinding, it should show first in wholesale diesel, before it reaches the pump.
It may hold. It may not. Either way, the record will say so, because the test was written down first.
Outcomes are a journey, not a destination
The Isaias lab judged its price prediction one week after landfall. The riptide only became visible at four.
That is the lesson I keep returning to. An outcome measured at the acute moment is a snapshot of the surface, not a verdict on what happened. A crisis measure can look like a success on day three and a failure on day thirty. A technology implementation can look like a success at go-live and a failure eighteen months later. If you stop measuring when the surface calms, you will mistake the calm for the outcome.
I have been skeptical of this kind of snapshot for a long time. In 2007, Ariba’s PR firm invited me to interview a senior executive about a new contract win. The release cited 108 percent growth in its on-demand client base, with nothing to compare it to. I wanted to know whether the new model would cannibalize the old one. Ariba couldn’t answer by the end of our third interview. Press releases and case studies share the same weakness: they measure at the moment the surface is calmest.
It is why we track outcomes over time, long after the event that prompted them. True outcomes are a journey, not a destination.
When you stop reading, the swan changes color
There are three more things a riptide teaches, and our labs have shown each of them this week.
First, not reported does not mean it did not happen. In our Cold Chain Strand Lab, we have so far documented nine food-sector cyberattacks from public sources. Every company reported what happened and when its systems came back. Not one record showed who had the authority to act, or whether they acted in time. Those decisions were made. They simply were not recorded anywhere the public can see. Aviation solved this problem decades ago with flight recorders: the forces on an aircraft are the same whether or not anyone is recording, but only a recording lets investigators see the sequence of decisions clearly. Most organizations operate without a recorder for their own decisions.
Second, the color of a swan depends on how often you read the strand. I have long defined the difference this way: the difference between a gray swan and a black swan is not the event itself. It is whether the signals that preceded it were in the scan. Timing adds a second test. A strand can be in the scan and still arrive as a surprise if you read it too rarely. Our lab reads its strands weekly. It recorded one tanker’s change of course ten days after it happened, and the resumption of strikes on Russian refineries six days after. Neither was a black swan, because both strands were being watched. But if the time it takes to notice, plus the time it takes to respond, is longer than the time a strand takes to act, the event arrives as a surprise however well the strand is understood. And because there is no time limit on when a strand can act, the riskiest decision of all is to stop reading it. That is when a gray swan becomes a black one.
Third, a strand does not have a single point of impact. It has an impact path. In the 1998 engagement, once we found what time of day the orders were arriving, we aligned delivery with the objective. But delivery was the flashpoint, not the end of the strand. A wrong part shipped, a faulty part, or a wrong part ordered belonged to the same strand, and only showed itself further along. Measured at the delivery point, the strand looked solved. The same is true of the diesel deal: at the point of authorization, it was confirmed. At the point of delivery, the cargoes were turning away. Where you take the reading decides what you see. In 1961, President Kennedy set the goal as “landing a man on the moon and returning him safely to the earth.” Success was defined at the end of the path, not at its most important, most dramatic point: the moon landing.
Three domains, one pattern
Step back, and the examples in this post come from three domains that rarely appear in the same conversation:
The industries differ. The pattern does not. In each, what was recorded diverged from what was actually happening, and only following the strand over time showed where.
Three domains, one pattern. In energy policy, maritime shipping and food-sector cyberattacks, what was recorded diverged from what was actually happening. Illustrative.
That is the claim at the heart of Strand Commonality™: the same strands, crossing in the same way, across domains that have nothing else in common. It is also why we call the principles behind it Invariant Physics™. The technology, the industry and the headline change. The forces that decide the outcome do not.
Why the strands must be recognized and monitored
The diesel announcement was not an isolated event. It crossed the hurricane, the war in Iran, the Russia–Ukraine conflict, sanctions policy, the U.S. midterm elections, and the decisions of individual port operators and state governments. No single function, desk or model following one of those strands alone would have seen the riptide forming.
This is what Strand Commonality™ is for: recognizing which strands are moving together, and continuing to monitor them after the moment that made them visible. Recognition without monitoring gives you a good call at the acute moment, and no way of knowing whether it held.
And here is the most important point. A riptide is not a single event or a rare anomaly. It is a regular occurrence. Announcements outrun delivery. Press releases outrun results. Go-live outruns adoption. Every crisis produces measures that calm the surface before they change the current. That means riptides can be managed, not just survived. Once you have identified all of the strands that require your attention, you know where the current is likely to run, what to watch, and when the calm is real. What turns a riptide from a hazard into something you can manage is not predicting it. It is having already identified the strands it runs through.
That is what we do. We lock what we understand before the outcome is known. We follow the strands forward, not just to the headline. We record our misses as plainly as our hits. And we keep measuring until the outcome, not the surface, tells us whether we were right.
That is also why we run our labs in real time rather than relying on case references. A case reference is a snapshot too: it proves we were right once, at a moment of our choosing. A live lab has to keep proving it, on dates fixed in advance, in the open, including when the result goes against us. Like the principles we call Invariant Physics™, our reliability and the accuracy of our coverage are never proven once and for all. We have to keep proving them.
Across our labs, the output accuracy of ARA™ RAM 2025™ running on SLAP OS™, assessed against the evidence, is holding at an average of around 91 percent, and it fluctuates. That is a different measure from our prediction record, which includes misses like the Isaias diesel prediction. The point is not the number. It is that tracking both the hits and the misses gives us an accurate picture, and an accurate picture is what drives the Infinity Loopback: the continuous learning of our algorithms.
Truth Is Believing. Accuracy Is Knowing. Outcome Is Proof.™
-30-
Share this:
Like this:
Related