tetlock-style-economists - Predictions for:

I estimate a 20% probability that the event, under its relatively strict resolution criteria, occurs before December 1, 2026.

The base rate for a deliberate Russian kinetic attack on NATO territory is low. Since the start of the Ukraine war, Russia has repeatedly conducted dangerous operations near NATO borders, but it has generally avoided an unmistakable, attributable attack on Alliance territory. A major invasion or sustained incursion remains very unlikely, probably around 2–4%.

The probability is nevertheless materially higher than the historical baseline because roughly two months remain and several developments point toward a plausible limited-strike pathway: an explosive-carrying drone entered Lithuania from the Belarusian direction; Russian attacks have occurred close to Poland; Polish leadership has publicly warned of possible Russian strikes disguised as accidents; and military activity in the Baltic and along NATO’s eastern flank is dense. These factors make a small drone or missile incident operationally feasible.

My factor assessment is approximately:

• Opportunity for a limited kinetic incident: 35–45%. Russian operations near NATO borders, dense air activity, and Belarusian access create meaningful opportunity.
• Russian or Belarusian willingness to conduct a deliberate probe: 25–35%. Moscow has incentives to test allied cohesion, but the escalation risks remain substantial.
• Successful execution or penetration of NATO territory: 35–50%. A drone or missile could be intercepted, but even an intercepted weapon might qualify if deliberate targeting is subsequently established.
• Clear attribution and evidence of deliberate intent to test NATO: 45–60% conditional on an incident. This is the largest constraint. The Lithuanian case demonstrates that even an armed incursion may remain ambiguous, accidental, Ukrainian, criminal, or otherwise difficult to attribute.

Combining these factors gives a lower probability than the research document’s central estimate of 22%, but still within its stated range. I reduce the estimate modestly because the document may double-count related indicators: the Lithuanian incident, Polish warning, nearby strikes, and NATO activity all reflect the same underlying deterioration. I also assign substantial weight to the strict requirement that the event be more than accidental spillover and be plausibly intended, at least in part, to test Article 5.

The most likely qualifying scenario is a small, deniable drone or missile attack against a border-area logistics, military, communications, or transport target in Poland or the Baltic states, potentially involving Belarus. A tank attack or major territorial incursion is much less likely.

The main reasons I could be wrong are that the Lithuanian drone may later be conclusively attributed to Russia or Belarus, effectively resolving the event; intelligence behind Poland’s warning may reflect a concrete operational plan rather than a general assessment; or Russia may calculate that a deniable strike carries manageable escalation costs. Conversely, a ceasefire, stronger NATO deterrence, clear attribution of the Lithuanian incident to a non-Russian actor, or Russia’s continued fear of direct confrontation could keep the probability closer to 10–15%.

I estimate a 19% probability that the event will occur by November 30, 2026. This is somewhat below the research document’s 22% central estimate because the resolution criteria are unusually demanding: the incident must be kinetic, attributable to Russia or a Russian-directed Belarusian proxy, deliberate rather than accidental, and plausibly intended in part to test NATO’s Article 5 commitment.

The historical base rate for Russia deliberately attacking NATO territory or assets is very low. Since the start of the Ukraine war, Russia has repeatedly operated near NATO borders and has caused several dangerous spillover incidents, but it has generally avoided an unmistakable attack on a NATO member. This reflects the high costs of direct confrontation with the Alliance, Russia’s ongoing resource commitments in Ukraine, and NATO’s strengthened air-defense and surveillance posture.

The risk is nevertheless materially above the normal baseline. The armed drone entering Lithuania from the Belarusian direction, the explosive payload, Russia’s strikes near the Polish border, repeated Baltic military encounters, and the Polish warning about possible deniable Russian attacks all support a meaningful probability of a limited kinetic probe. A small drone or missile incident is much more plausible than a tank incursion or sustained territorial attack.

A rough decomposition is: approximately 25% for any deliberate Russian or Russian-directed kinetic incident affecting NATO territory, personnel, vessels, or military assets during the remaining period; multiplied by roughly 75% that such an incident would ultimately be judged sufficiently deliberate, attributable, and connected to testing NATO rather than being accidental, criminal, Ukrainian, or merely ambiguous. This produces an estimate near 19%.

The main upward factor is the possibility that the Lithuanian incident is later attributed to Russia or Belarus and judged intentional. The main downward factors are the short remaining time window, the lack of confirmed attribution, Russia’s strong incentive to preserve deniability and avoid a direct NATO confrontation, and the possibility that NATO’s visible response already provides deterrence. I assign only about 2–4% to a major or sustained Russian attack; nearly all of the estimated probability concerns a limited drone, missile, maritime, or proxy incident.

I could be wrong chiefly if the Lithuanian drone investigation reveals clear Russian or Belarusian direction, or if Polish intelligence reflects a concrete operational plan rather than a broad warning. Conversely, the probability would be overstated if the Lithuanian incident proves non-Russian or accidental, if improved NATO defenses prevent further incursions, or if Russia prioritizes avoiding escalation while continuing to rely on hybrid operations, which are excluded from this forecast.

I estimate a 12% probability that the event, as narrowly defined, occurs before December 1, 2026. This is slightly below the research document’s 14% central estimate because the latest prediction history extends through September 16, while the remaining time window is now shorter.

Base rate: Since the start of the Ukraine war, Russia has repeatedly conducted dangerous operations near NATO territory, but it has not deliberately launched a clearly attributable conventional attack on a NATO member. The historical base rate for an intentional Russia-NATO kinetic attack is therefore very low. The current war raises the risk above the normal peacetime level, but not to a high-probability event.

Key factors:

1. Opportunity and exposure — upward pressure. Russian missile and drone operations against Ukraine, reconnaissance flights, Belarusian military integration, and dense activity around the Baltic and eastern flank create multiple opportunities for a limited strike, spillover event, or deniable probe. This supports a probability meaningfully above 5%.

2. Russian intent — moderate upward pressure. Moscow has shown willingness to conduct coercive and hybrid operations inside NATO countries and to test Western responses. A small, ambiguous drone or missile attack is more plausible than a major invasion. However, there is no clear evidence of a current decision to initiate a NATO confrontation.

3. Deterrence and escalation costs — strong downward pressure. A deliberate attack would risk NATO military retaliation, further Western military support for Ukraine, and loss of diplomatic leverage. NATO’s strengthened eastern-flank posture also makes surprise and deniability more difficult.

4. Current military indicators — downward to neutral. There is no reported large Russian buildup against the Baltic states, no clear offensive preparation, and no newly identified qualifying incident. Exercises and reconnaissance activity are concerning but are not, by themselves, evidence of imminent attack.

5. Diplomacy — modest downward pressure. Continuing U.S.-Russia diplomatic engagement creates an incentive for Moscow to avoid an unmistakable attack, although the lack of a durable ceasefire means this effect is limited.

6. Time remaining — downward pressure. With roughly ten to eleven weeks left, there is still enough time for a small drone or missile incident, particularly during a renewed Russian air campaign, but insufficient time for a large planned offensive to develop without warning.

Scenario weighting: I assign roughly 8–10% to a deliberate limited drone or missile attack that could plausibly be interpreted as a test of NATO, about 1–2% to a more overt or sustained attack, and a small additional probability to a Belarus-linked or proxy action that meets the attribution and intent threshold. These categories are not fully independent, so I do not simply add their maximum values; the combined estimate is approximately 12%.

The principal reason not to forecast higher is that Russia can pursue most of its coercive aims through hybrid activity, airspace pressure, reconnaissance, and deniable operations without crossing the kinetic threshold. The principal reason not to forecast lower is that repeated spillover incidents and ambiguous drone activity create a realistic pathway to a deliberately limited strike disguised as an accident.

I could be wrong if intelligence has not detected preparations for a deniable operation, if diplomatic talks collapse abruptly, or if Russia concludes that NATO would respond weakly to a small incident. Conversely, the event may not occur even after a serious border incident because attribution or intent remains ambiguous, which is especially important given the forecast’s strict resolution criteria.

Clarifying the event: this requires more than a Russian or Belarusian weapon entering NATO airspace or causing accidental spillover. It requires a deliberate, attributable kinetic attack before December 1, 2026, with evidence that it was intended at least partly to test NATO’s Article 5 response. A small drone or missile strike would be the most plausible form; a tank incursion or sustained attack is much less likely.

Historical base rate: Russia has repeatedly conducted dangerous military and hybrid activity near or inside NATO territory, but since NATO’s creation there has been no clearly deliberate conventional Russian attack on a NATO member. The base rate for an intentional attack is therefore very low. The current Russia-Ukraine war raises the risk of miscalculation and limited probing, but the strict intent and attribution requirements materially reduce the probability of resolution.

Key upward factors include Russia’s continuing high-volume missile and drone campaign against Ukraine, repeated airspace incidents and reconnaissance flights near the Baltic states and Poland, Belarus’s military integration with Russia, Russian-attributed hybrid operations inside NATO, and the possibility that Moscow would prefer a small, ambiguous test rather than a major invasion. These factors make a limited deniable drone or missile incident plausible.

Key downward factors are the severe escalation costs of a clearly attributable attack, Russia’s continuing military commitment in Ukraine, strengthened NATO air defenses and surveillance, the absence of evidence of a major Russian buildup against the Baltic states, and ongoing diplomatic engagement. NATO’s demonstrated tendency to distinguish spillover incidents from deliberate attacks also reduces the likelihood that an ambiguous event will satisfy the resolution criteria.

The recent forecast history moved from 8% to 13%, then stabilized around 11–12%. The research document’s 14% central estimate appears somewhat aggressive because it gives substantial weight to ambiguous incidents and hybrid activity that are explicitly excluded. I therefore place the probability modestly above the latest 11% estimates but below 14%, at 12%.

I could be wrong if diplomacy collapses, Russia launches an unusually large autumn air campaign, or Moscow exploits a drone incident near Belarus or the Baltic region to test allied political cohesion. Conversely, an apparently serious incident may remain unresolved as to attribution or intent and therefore fail to qualify. The probability of a major sustained Russian attack is far lower, probably around 1–3%; most of the 12% estimate is concentrated in a small, deniable drone or missile strike.

The event is narrowly defined: it requires a deliberate, attributable Russian or Russian-directed kinetic attack against NATO territory, personnel, vessels, or military assets before December 1, 2026, with evidence that it was intended at least partly to test Article 5. Accidental spillover, debris, ambiguous airspace violations, and hybrid operations do not qualify.

The historical base rate for an intentional Russian kinetic attack on NATO territory is very low. Russia has repeatedly operated near NATO airspace and conducted hybrid activities inside NATO countries, but it has generally avoided a clearly attributable conventional attack that could trigger a direct NATO response. This strongly limits the prior probability.

The main upward factors are the continuing high-volume Russian air campaign against Ukraine, increased drone activity near Baltic airspace, Russian reconnaissance flights, Belarus’s military integration with Russia, and the possibility of a small deniable strike against a military or logistics target. A limited drone or missile attack would be more plausible than a tank incursion or major territorial operation. Existing spillover incidents also create opportunities for Moscow to disguise a deliberate action as an accident.

The main downward factors are the severe escalation costs for Russia, NATO’s strengthened eastern-flank defenses, the lack of evidence of a major offensive buildup, Russia’s continued military focus on Ukraine, and ongoing U.S.-Russia diplomatic engagement. A deliberate attack would risk unifying NATO and producing consequences that Moscow might not be able to control. The strict requirement to establish both Russian responsibility and deliberate intent further reduces the probability compared with the chance of an accidental or ambiguous incident.

Recombining these factors, I estimate roughly an 11% probability. This is slightly below the research document’s 14% central estimate and consistent with the latest historical forecast of 11%, because the remaining time window is short and there is no reported new qualifying incident or clear preparation for one. I would assign roughly 1–3% to a major or sustained attack, with most of the 11% concentrated in a small drone or missile incident involving Poland, the Baltic states, Romania, Bulgaria, Finland, or a Belarus-linked launch vector.

The estimate could be too high if diplomatic contacts develop into a durable ceasefire or if Russian leaders remain deterred by NATO’s response posture. It could be too low if negotiations collapse, Russia intensifies attacks near NATO borders, Belarus hosts new Russian strike assets, or a future incident causes casualties or major damage and is rapidly attributed to Russian forces.

I estimate an 11% probability that the event, as narrowly defined, occurs before November 30, 2026. This is slightly below the research document’s 14% central estimate and broadly consistent with the recent prediction history of 8%, 13%, and 12%.

1. Clarifying the event: This is not simply a drone crossing, debris impact, airspace violation, sabotage operation, or accidental spillover. It requires a deliberate kinetic attack by Russia, Belarus acting under clear Russian direction, or a Russian-controlled proxy, with sufficient evidence that the operation was at least partly intended to test NATO’s Article 5 response. That attribution-and-intent requirement materially lowers the probability.

2. Historical base rate: Since the creation of NATO, deliberate conventional Russian attacks on NATO territory have been extremely rare to nonexistent. Russia has repeatedly used coercion, reconnaissance, airspace violations, cyber operations, sabotage, and proxy activity without crossing into an unmistakable armed attack. The base rate for a deliberate attack is therefore very low, although the war in Ukraine and Russia’s increasing risk tolerance raise it above normal peacetime levels.

3. Deliberate Russian decision: The probability that Moscow chooses to conduct a limited test is meaningfully above zero, perhaps around 15–20% over the remaining period. Russia may see value in probing NATO cohesion, especially with ambiguous drones or attacks routed through Belarus. However, a deliberate strike would risk military retaliation, stronger NATO deployments, and loss of diplomatic leverage. Russia can pursue most of its pressure objectives through hybrid means, which remain less costly and are currently preferred.

4. Operational opportunity: The probability of an incident opportunity is elevated, perhaps around 25–35%. Russia is conducting intensive drone and missile operations against Ukraine, while reconnaissance flights, exercises, counter-drone activity, and military deployments are increasing around NATO’s eastern flank. This creates opportunities for miscalculation, spillover, or a deniable operation. The July missile impact in Poland and Baltic drone activity demonstrate the mechanism, but neither established deliberate Russian intent.

5. Attribution and resolution: The probability that a future incident would be credibly attributed to Russia or a Russian-directed Belarusian proxy, and judged deliberate rather than accidental or Ukrainian-origin, is only moderate. Moscow could exploit ambiguity, but ambiguity also makes it difficult for the event to satisfy the forecast’s resolution criteria. This is a major reason not to equate the probability of any kinetic incident with the probability of this specific event.

6. Time window and deterrence: Approximately twelve weeks remain in the supplied assessment framework. That is sufficient for a small drone or missile incident but too short for a major planned invasion absent obvious preparations. NATO’s Eastern Sentry measures, increased air defense, allied exercises, and multinational forces raise the cost of an attack and improve detection. Active U.S.-Russia diplomacy also provides a temporary incentive for restraint, though its effect is limited because no durable settlement has emerged.

The most likely qualifying scenario is a small, ambiguous drone or missile strike against eastern NATO territory, a military or logistics facility, or a vessel, potentially involving Belarus. A major territorial incursion remains much less likely, approximately 1–3%.

I would increase the estimate sharply if there were a confirmed Russian launch from Belarus, a strike on a NATO military installation, multiple coordinated incursions, explicit Russian threats against a member state, or emergency NATO consultations followed by force dispersal. I would reduce it if diplomacy produced a durable ceasefire, Russia reduced military activity near NATO borders, or Moscow and NATO adopted effective incident-management measures.

The main ways this forecast could be wrong are: an apparently accidental incident could later be shown to have been a deliberate probe; Russia could exploit a diplomatic breakdown to escalate rapidly; Belarus could enable an operation without clear public command evidence; or repeated spillover incidents could culminate in a misidentified and rapidly escalating exchange. Conversely, the estimate could be too high if Russia remains focused on Ukraine and continues to prefer deniable hybrid pressure over an action that risks direct war with NATO.

Clarification: I interpret the event narrowly. It requires a deliberate Russian, Belarusian, or clearly Russian-directed conventional kinetic attack against a NATO member before December 1, 2026, with enough evidence to distinguish it from accidental spillover, debris, an airspace violation, or a hybrid operation. The attack need not be large, but intent to test or probe NATO’s response must be reasonably supported.

Historical base rate: Since Russia’s full-scale invasion of Ukraine, there have been repeated dangerous incursions, missile spillovers, reconnaissance encounters, and ambiguous drone incidents near or inside NATO territory, but no clearly deliberate Russian kinetic attack on NATO. The base rate for a major attack is therefore very low; the base rate for a small, deniable incident is higher but still limited.

Factor estimates:

• Opportunity and operational exposure: moderately high. Russia’s continuing drone and missile campaign, activity near NATO airspace, Belarusian military integration, and dense Baltic exercises create many opportunities for an incident during the roughly 84-day window.
• Russian intent: low to moderate. Moscow has incentives to probe NATO cohesion and already conducts aggressive activity below the overt-war threshold, but a deliberate kinetic attack would carry substantial escalation costs and could strengthen NATO unity.
• Capability: high. Russia and Belarus possess the relevant missiles, drones, surveillance, and launch infrastructure. Capability is not the main constraint.
• Deterrence and defensive posture: strongly downward. Eastern Sentry, allied deployments, air defenses, surveillance, and the prospect of a unified response make a purposeful attack risky.
• Attribution and resolution: downward. Even if a weapon enters NATO territory, proving that it was deliberately targeted by Russia or a Russian-directed proxy—and that it was intended to test Article 5—may be difficult. Many plausible incidents would resolve as accidental spillover or remain ambiguous.
• Diplomacy and Russia’s Ukraine commitments: modestly downward. Current U.S.-Russian engagement and Russia’s continued focus on Ukraine reduce incentives for an unmistakable attack, although failed talks could reverse this effect.

Recombining these factors, I estimate roughly a 15% chance of some Russian-related kinetic spillover or ambiguous incident affecting NATO territory or assets during the remaining period, but only about a 12% chance that such an event will meet the stricter deliberate-attack and intent criteria in the resolution.

This is slightly below the research document’s 14% central estimate. The main reason is that the event’s intent and attribution requirements are unusually demanding: a missile impact or drone crossing alone should not count. A major, sustained attack remains far less likely, probably around 1–3%.

I could be wrong if diplomacy collapses, Russia launches a major new air campaign near NATO borders, Belarus becomes an active launch platform, or Moscow believes Washington would respond cautiously to a small ambiguous strike. Conversely, the estimate would fall materially if a credible ceasefire emerges or if NATO and the United States issue especially clear, coordinated deterrence commitments.

I estimate a 13% probability, modestly above the prior 8% forecast but below the upper end of the research document’s 12–18% range.

1. Clarifying the event: The event requires more than a Russian airspace violation, accidental spillover, debris impact, sabotage, or cyber operation. It requires a deliberate kinetic attack by Russia, Belarus under clear Russian direction, or a Russian-controlled proxy against NATO territory, personnel, vessels, or military assets before December 1, 2026. Because the attack must be interpreted as at least partly intended to test NATO’s response, ambiguous incidents may not qualify.

2. Historical base rate: Since NATO’s expansion and the end of the Cold War, Russia has repeatedly conducted coercive and hybrid operations against NATO members but has avoided an unambiguous deliberate conventional attack. The base rate for a direct attack is therefore very low. Recent Russian behavior raises the near-term risk, but does not overturn that historical pattern.

3. Factors increasing the probability: Russia is conducting intensive missile and drone operations near NATO borders; there have been repeated airspace incursions and a missile landing in Poland; Russian-linked sabotage inside NATO countries indicates willingness to operate there; Belarus provides a plausible route for deniable action; and Moscow could view a small drone or missile strike as a way to probe NATO political cohesion without launching a major war. The remaining period also includes enough time for a limited incident to occur during a Ukrainian strike campaign or regional military exercise.

4. Factors reducing the probability: A clearly attributable attack could trigger Article 4 consultations, military retaliation, further NATO deployments, and much stronger Western support for Ukraine. Russia is already heavily committed in Ukraine and can pursue many of its coercive objectives through sabotage, cyber operations, disinformation, and airspace violations without crossing the kinetic threshold. NATO’s strengthened eastern-flank defenses also increase the chance of rapid detection and attribution. A tank attack is especially unlikely because it would require conspicuous preparations and carry high escalation risks.

5. Recombination: I assess roughly a 15–18% chance of some deliberate Russian or Russian-directed kinetic incident against NATO territory in the remaining period, but reduce that estimate because the resolution criteria require sufficiently clear intent and attribution. This produces a central estimate of about 13% for the specific event.

6. Overconfidence check: The estimate should not be treated as precise. A small drone or missile incident could be classified differently depending on evidence about targeting, command responsibility, and intent. The probability is also highly sensitive to one sudden crisis, such as a serious border incident, a Belarus-related military move, or a major escalation in Ukraine.

7. Why this could be wrong: I could be underestimating Putin’s willingness to accept escalation, the possibility of an unauthorized or proxy action, or the chance that Russia deliberately exploits an ongoing Ukrainian strike to conceal a test of NATO. Conversely, I could be overestimating the significance of recent incidents: they may reflect reckless but still bounded behavior, while the strong costs of an overt attack continue to deter Moscow. A major sustained attack remains far less likely, probably around 1–3%, but a limited ambiguous drone or missile strike is the principal pathway to a 'yes' resolution.

Clarification: The event is a deliberate, attributable Russian or Belarusian-proxy kinetic attack on the territory of a NATO member before November 30, 2026, using drones, missiles, or ground forces. Hybrid activity, accidental spillover, isolated airspace violations, and ambiguous incidents that do not clearly constitute an attack are excluded.

Historical base rate: The historical base rate is very low. Russia has repeatedly used cyber operations, sabotage, disinformation, airspace incursions, and other gray-zone tactics against NATO countries, but it has avoided an overt kinetic attack on NATO territory. NATO deterrence, the risk of a much larger war, and Russia’s need to prioritize Ukraine all weigh against such a decision.

Key factors:

1. Russian incentives to test NATO: Moderate upward pressure. Moscow may see value in probing alliance cohesion, especially if it perceives U.S. disengagement, political division, or weak willingness to respond. A small drone or missile strike could theoretically be intended to create ambiguity and test the threshold for collective action.
2. Military availability and strategic priorities: Downward pressure. Russia remains heavily committed to the war in Ukraine. A deliberate attack on NATO would expose Russian forces and infrastructure to potentially overwhelming allied retaliation, making it a costly way to achieve political objectives.
3. NATO deterrence and alliance response: Strong downward pressure. Even if Article 5 consultation is not automatic, an unambiguous armed attack would likely trigger a substantial allied response. Russia has strong reasons to believe that a kinetic attack would be difficult to contain.
4. Gray-zone escalation: Moderate upward pressure. Continued drone incursions, sabotage, and other provocations increase the chance of miscalculation or gradual escalation. However, the event definition requires a clear kinetic attack, not merely a continuation of these activities.
5. Belarusian involvement: Slight upward pressure but limited. Belarus could provide geographic access or deniability, particularly near Poland or the Baltic states. Nevertheless, proxy use would not reliably prevent NATO from attributing responsibility to Russia, so it offers only limited protection against retaliation.
6. Time window and warning: The remaining period through November 2026 provides opportunities for escalation, but a deliberate operation would probably generate substantial intelligence warning through force movements, logistics, exercises, and political preparation. This reduces the likelihood of a surprise attack.

Recombination: The supplied reference estimates roughly 9% for a limited kinetic action on the Baltic flank and about 17% for a broader Russia-NATO military clash. This event is narrower than the broader clash estimate but somewhat broader than a major Baltic operation because it includes a small drone or missile attack and a Belarusian proxy. After accounting for the low historical base rate, Russia’s current constraints, and the strong deterrent effect of NATO, I estimate an 8% probability.

Why this could be wrong: The probability could be materially higher if the Ukraine war ends on terms Russia interprets as a victory, Russian forces regenerate quickly, U.S. political commitment to NATO weakens, or a crisis involving Belarus creates a plausible pretext for intervention. It could also be higher if repeated incursions produce an accidental or deliberately calibrated kinetic strike. Conversely, a durable ceasefire, stronger NATO deployments, clear allied red lines, or continued Russian military exhaustion would reduce the risk substantially. The main uncertainty is that a low-level attack could be designed to remain ambiguous, making classification under the event’s definition difficult.