Discover our technical analysis of the real probability of direct aerial combat between fighter jets in 2025, based on context, doctrine, and figures.

In 2025, the idea of a conventional air combat between two fighter jets seems unlikely. The use of remote strikes, drones, and electronic warfare has greatly reduced these skirmishes. However, in some cases, a direct confrontation remains possible.

The operational context of direct air combat

Since the Cold War, the frequency of air combat has fallen sharply. According to the CSBA database, only 59 air victories were claimed between 1990 and 2002, with the last two in 2001. No other comparable engagements have been documented since then. The main reason is the increasing use of BVR missiles, AWACS systems, drones, and long-range strikes.

Recent conflicts illustrate this trend. In 2025, facing Iran, Israel dominated the airspace using F-35s and F-15s, with no air losses. The Israeli army neutralized enemy air defenses mainly through preparatory strikes. This success shows that direct confrontation remains possible in contexts of low technological opposition, but it does not apply to adversaries such as China or Russia, which have fourth- and fifth-generation air forces, multi-layered SAM defenses, and sophisticated electronic warfare capabilities.

In modern simulations, particularly those from the South Korean ACMI system, machine learning models achieve 83% accuracy in predicting air-to-air impacts using physical variables (distance, speed, relative position). These studies demonstrate that the probability of victory depends on a set of well-measurable factors.

Overall, the probability of a modern aircraft encountering a hostile opponent in close combat remains low, in the range of 0.1% to 1% per mission in a prepared conflict scenario. In a high-intensity environment, with two similar forces, this probability may increase, but remains limited by the effectiveness of remote detection systems.

Factors influencing the probability of close combat

Several parameters determine whether direct combat will take place. First, the technological level of the adversaries: the more AWACS systems, AESA radars, and very long-range missiles both sides have, the more engagements will be conducted BVR, without visual contact. However, in the event of electromagnetic interference or jamming, aircraft may lose their ability to engage at a distance and come closer.

Second, the density of missions and the type of operations. During regional crises such as the India-Pakistan conflict in 2025, engagements took place around limited areas. In such cases, direct confrontation becomes plausible but remains limited to a few encounters.

Thirdly, pilot training. In forces with high turnover rates and a lack of experienced pilots, the ability to avoid skirmishes is reduced. A 2025 report highlights that if an army does not have enough experienced pilots, attrition becomes critical in a conflict with a comparable adversary.

Finally, doctrine. Some armies still design close-range engagement scenarios, despite BVR dominance. Close-range combat exercises are used to train pilots. But modern tactical arrangements favor the suppression of enemy defenses before any contact is made.

According to BVR/WVR simulation models, the most influential variable remains initial spatial separation, followed by relative speed and altitude. Under favorable conditions, a pilot may have more than an 80% chance of hitting the target if the parameters are right: isolated target, favorable position, AWACS coverage.

Quantitative estimate of probability in 2025

To estimate the real probability of a direct confrontation between fighter jets in 2025, it is useful to distinguish between two contexts:

• Asymmetric conflicts (terrorism, insurrection, technologically outdated states)
• Conflicts between comparable powers (NATO vs. China, Russia)

In the former, such as Iraq, Syria, or Iran, the probability of a direct dogfight remains low but not zero. The 2017 incident in which a US F/A-18E shot down a Syrian Su-22 in visual engagement is a typical example. Such confrontations may occur in less than 0.5% of missions, but they are documented.

In the second case, the majority of engagements are conducted BVR, or by ground strikes and missiles, with an estimated probability of visual combat of less than 0.1% per sortie, unless the adversaries are forced to approach each other (jamming, sensor saturation). Furthermore, no active air combat between powers with modern fleets took place between 2000 and 2025.

The probability of a direct engagement can be modeled as P = λ·t·d, where λ is the frequency of missions, t is a proximity factor, and d is a BVR system failure factor. ACMI simulations suggest that d ~ 0.02 to 0.05 when functional sensors and AWACS coverage are active. λ·t is often less than 0.01 in well-prepared real-world conditions. Therefore, P remains very low.

The return of traditional dogfighting is highly unlikely in 2025. It is reserved for scenarios involving BVR system failures, tactical errors, or confrontations between obsolete aircraft. Close combat in a conflict between powers of equal strength is more of a theoretical risk than an expected reality.

However, in localized regional crises, close encounters remain possible. The India-Pakistan incident in May 2025 provides a glimpse of this: some skirmishes took place, but most of the action targeted ground installations, not direct air-to-air combat.

Finally, the rise of autonomous systems and AI-assisted piloting could further reduce exposure to risk. Simulation approaches using CNN-LSTM already make it possible to predict trajectories with an improvement of 32 to 34%, reducing uncertainty.

Get in touch to live a unique fighter jet experience – we fly in France AND YOU CAN TAKE THE CONTROLS!!!