After the Gulf War’s air campaign concluded in February 1991, U.S. commanders were confident that airpower had destroyed the bulk of Iraq’s Republican Guard before the ground offensive ever began. A subsequent General Accounting Office review found they were wrong and identified why. The Republican Guard was among the “least measurable” target categories in the entire campaign. About a third of reported F-117 strikes either lacked corroborating evidence or conflicted with other data, putting the probability of a successful F-117 strike between 41 and 60 percent. Commanders had mistaken reduced enemy activity for the physical destruction of enemy capacity.
That analytical error would recur in later campaigns — from Kosovo to today.
Ten days into the air campaign against Iran, Gen. Dan Caine, chairman of the Joint Chiefs of Staff, told reporters at the Pentagon that Iranian launches of “one way attack drones have decreased 83 percent since the beginning of the operation.” That reduction, he said, showed that strikes on launch infrastructure and production facilities are systematically dismantling Tehran’s ability to threaten the Gulf. Defense Secretary Pete Hegseth offered a similar interpretation, saying, “The numbers staying that low is a demonstration of … efficacy.”
It may be. But the statistic conflates two analytically distinct questions: what Iran is currently doing and what Iran retains the capability and capacity to do. The 83 percent figure reflects a decline in observed launch tempo — a behavior indicator. Treating that change in behavior as evidence that Iran’s drone capacity has been destroyed risks creating a misleading picture of how much of the threat has actually been eliminated.
That distinction matters. If Washington concludes that Iran’s drone threat has been largely neutralized when it has not, it could press harder — militarily or diplomatically — than Tehran is prepared to accept. The result could escalate a conflict already straining Gulf air defenses and threatening global markets.
Behavior Indicators Are Not Battle Damage Assessment
Joint doctrine distinguishes three levels of battle damage assessment. Physical damage assessment measures observable damage to a target. Functional damage assessment estimates how much of a target’s operational capability survives a strike. Target system damage assessment evaluates whether a campaign is degrading the adversary’s ability to fight.
Each level requires progressively more evidence. Physical damage can often be verified directly through imagery and strike reporting. Functional and system-level assessments require additional intelligence. Joint doctrine notes that functional damage “is not always observable directly” while the U.S. Air Force cautions that system-level assessment is a “data-intensive process” that “typically requires weeks to months to accumulate the data.”
The 83 percent figure presented at the Pentagon briefing is not a battle damage assessment at any of these levels. It is what doctrine calls a battle damage indicator — a measurable phenomenon that contributes to a formal assessment but is not a substitute for one. In this case, the decline in Iranian drone launches serves as the indicator, but it does not explain why.
Hegseth acknowledged as much. Battle damage assessment, he said, “takes time.” He is right. The available data support the claim that launch rates have declined. What they cannot yet support is the stronger inference being drawn at the briefing that Iran’s drone threat is being systematically eliminated. A reduced launch rate is consistent with that conclusion, but it is equally consistent with several others.
Why Iran Might Be Launching Fewer Drones
There are at least three other plausible alternative explanations for the decline in Iranian launches. First, it could reflect tactical recalibration. Moscow is reportedly sharing with Tehran drone tactics developed in Ukraine, including coordinated routing strategies designed to evade air defenses, as well as overhead satellite imagery to improve targeting. Tehran could be using this time to learn, adapt, and refine its strategy and tactics.
Second, Iran’s lower daily launch rates could reflect deliberate stockpiling for larger coordinated attacks later. Russian strikes in Ukraine have at times followed a similar pattern. Moscow has repeatedly accumulated ballistic missiles and drones before launching punishing attacks designed to saturate Ukrainian defenses and force the expenditures of interceptors faster than they can be replenished. If Iran is applying the same lesson, the decline in daily launches could reflect stockpiling rather than depletion.
Third, operational priorities may also be shifting toward the Strait of Hormuz. Iran’s Islamic Revolutionary Guard Corps has warned ships not to transit the strait, saying that vessels “could be at risk from missiles or rogue drones.” These are not idle threats. Tehran has reportedly laid a dozen mines in the waterway, and recent imagery has shown cargo ships transiting the strait coming under attack. The U.S. Navy has declined near-daily requests from the shipping industry for military escorts, citing the attack risk.
If Washington attempts to deploy minesweepers — which it lacks — or facilitate the passage of commercial vessels with warship escorts, Iran would likely want its missile and drone inventory in position for that engagement rather than in daily barrages across the Gulf. A reduction in launch rates may therefore reflect repositioning toward the strait rather than any reduction in overall capacity.
Finally, Tehran may have simply concluded that a lower, sustained launch rate is sufficient to maintain coercive pressure on Gulf states while conserving inventory for a conflict that could last months. A strategy of attrition does not require maximum effort every day.
Each of these explanations fits the observed decline in launches.
The pattern of Iranian attacks is more consistent with recalibration than capacity constraints. Over the first ten days of the war, Iranian strikes followed a deliberate escalatory sequence — military installations first, then logistics hubs and communications nodes, and then energy infrastructure. Each category served a distinct coercive purpose. This kind of target sequencing is not typical of an adversary running low on munitions. It is more consistent with deliberate inventory management.
Hegseth hinted at the risk himself, warning, “If the enemy can simply wait and then project power, that’s problematic.”
Why Drone Degradation Is Hard to Measure
Assessing drone degradation requires finding, striking, and verifying targets — three steps the Shahed-136 was designed to complicate. Weighing roughly 200 kilograms, the drone requires no transporter-erector launcher, dedicated launch facility, or fixed infrastructure. It launches from an angled rail mounted on a pickup truck, after which crews can quickly relocate, minimizing exposure to counterstrikes. There are no fixed launch sites, and few observable preparation signatures that distinguish a warehouse full of drones from an empty one. The Shahed is designed to frustrate the targeting methods effective against ballistic missiles, which rely on large vehicles, fixed infrastructure, and longer preparation times that generate detectable signatures.
That said, U.S.-Israeli combined intelligence efforts have had some success against fixed production and storage infrastructure. U.S. Central Command has reported striking more than 6,000 Iranian targets, including multiple drone launch and storage facilities. But much of Iran’s drone production is decentralized and dispersed, complicating efforts to identify and strike the full supply chain. For drones already produced, the challenge is even greater. Shaheds require no specialized storage facilities and generate few identifiable signatures before launch. In many cases, there may simply be nothing observable to target.
The uncertainty surrounding Iran’s stockpiles complicates the problem. Prewar estimates range from several thousand to well above 10,000 drones. That variance alone makes estimating precise degradation nearly impossible. Iran has launched over 2,000 drones since the start of the conflict. Even assuming production has halted entirely, for which there is not established evidence, Iran still likely possesses a large inventory.
History offers a cautionary precedent. During the Scud hunt in 1991, U.S. forces searched for Iraq’s truck-mounted ballistic missiles with the full advantage of air superiority and extensive intelligence support. They still failed to achieve a single confirmed kill against a system that was larger, infrastructure-dependent, and slower to deploy than the Shahed. The pattern has repeated. In Kosovo in 1999, post-war battle damage assessment found that Serbian mobile military assets, including surface-to-air missile batteries and artillery, had largely survived. The Yemen campaign offered a more recent test of the same problem. Over 900 U.S. and U.K. airstrikes between January 2024 and January 2025 targeted Houthi missile and drone launch sites — many of them Shahed-derived systems — but failed to suppress the launches, in part because the systems were dispersed, mobile, and difficult to locate. Against Iran’s Shaheds, this pattern is likely to hold. It is small, mobile, fast to relocate, and designed with decades of subsequent U.S. air operations in mind.
What Analysts and Policymakers Should Be Asking
None of this is an argument that the air campaign has failed, or that degradation figures are fabricated. U.S. and Israeli strikes have likely imposed real constraints on Iran’s missile and drone operations. The reduction in observed launch rates is operationally significant. But the critical question is what that reduction actually indicates — and the 83 percent figure alone cannot answer it.
Analysts and policymakers seeking a more rigorous picture should ask several questions. First, what physical evidence — post-strike imagery, signals intelligence, and human intelligence — corroborates the 83 percent figure beyond the behavioral change itself? Second, what is the current assessed size of Iran’s remaining stockpile, what methodology produced that estimate, and what is the confidence interval? Third, has the drone component supply chain been assessed independently of assembly facility strike data, or does the production degradation claim rest entirely on the destruction of assembly hubs? Fourth, how are Gulf states’ interception rates evolving, and how long can regional air defenses sustain the expenditure of interceptors required to defeat these attacks? These questions distinguish a campaign producing lasting physical effects from one that is suppressing observable behavior while leaving underlying capability and capacity intact.
The Desert Storm findings were not published until six years after the war ended. Systems like the Shahed — designed to obscure production, storage, and launch — make the gap between confident assessment and ground truth even wider. Hegseth was correct about one point: Battle damage assessment “takes time.”
On March 11 — the day after the Pentagon briefing — Iran conducted what it described as its 37th wave of attacks, striking targets across Kuwait, Bahrain, Qatar, the United Arab Emirates, Saudi Arabia, and Oman while also hitting multiple vessels in the Strait of Hormuz. Whatever the 83 percent figure measures, it does not reflect an adversary whose strike capacity — much less its will to fight — has already been neutralized.
Kelly A. Grieco is a senior fellow with the Reimagining U.S. Grand Strategy Program at the Stimson Center and an adjunct professor of security studies at Georgetown University.
**Please note, as a matter of house style, War on the Rocks will not use a different name for the U.S. Department of Defense until and unless the name is changed by statute by the U.S. Congress.
Image: Petty Officer 1st Class Eric Brann via Wikimedia Commons
.jpg){kind=link}