Lede: Between the bruising lessons of 2024 and the renewed confrontation of 2026, Hezbollah and Israel both raced to adapt - one under fire, the other through rehearsals for a wider multi-front war.
Hezbollah's military thinking is layered, cumulative, and difficult to penetrate. No academic researcher, and not even an intelligence analyst, can credibly claim to have gathered all its components, analyzed them in full, or understood how they interact.
That complexity also shapes how the party learns and extracts lessons. Hezbollah's added advantage, however, is the speed with which it adapts. The clearest evidence is the change in its security and military performance between two wars separated by only 15 months - a period in which the party remained under continuous fire.
The 2024 war forced Hezbollah into a painful review of how it fought, how it deployed, and how much of its command structure could survive when the opening blows came hard and fast. The lessons were not drawn in seminar rooms, but by fighters and commanders who had lived the previous battle, absorbed its losses, and then found themselves preparing for the next one before the smoke had cleared.
This account rests on interviews conducted over roughly a year after the 2024 war with Hezbollah security and military officials. They say no visual or audio documentation can be presented because of the latest "harsh security lessons." Israel was conducting its own review at the same time, using the final months before the renewed confrontation to rehearse for a long, multi-front war that Hezbollah and Iran were watching closely.
Fighting from 2024 to 2026 In Tel Aviv's planning view, the southern Lebanon battlefield is divided into two sectors: eastern and western. Its divisions are deployed accordingly, based on the type of force each sector requires and on near-compulsory entry routes imposed by the terrain - routes that have shaped the battlefield since 1978.
Hezbollah divides the south differently, into three sectors: western, central, and eastern. Responsibility is split between the Nasr Unit, which handles the eastern and central sectors south of the Litani River, and the Aziz Unit, which covers the western sector. The Badr Unit, now heavily discussed again in Israeli commentary during the 2026 confrontation, is responsible for the area north of the Litani.
The Radwan forces were among Israel's central concerns in the previous war, when it demanded that they be pushed north of the Litani. Their presence has returned to Israeli discussion in the current war, with claims in early March that around 1,000 fighters were active. But Radwan is not tied to a fixed territory. It is an elite force that can be redeployed according to the needs of each battle.
The figure of 1,000 Radwan fighters cited by Israel comes from its own claim that 2,500 members of the unit remained combat-capable after the 2024 war, out of an original force of 5,000, most of whom were wounded in the pager and walkie-talkie attacks of September that year.
By Israel's own account, then, another 1,500 Radwan fighters have yet to enter the battle. Hezbollah does not comment on these figures, either publicly or in the private meetings conducted for this article and series.
What stands out in discussions of numbers, however, is a repeated observation made by several planning commanders - including those working on information files - and by field commanders who fought in 2006, Syria, the "support" front, and the 2024 battle of the 'Possessors of Great Strength.' They agree that the large number of fighters placed on alert along the front in 2024 sometimes obstructed operations and contributed to losses and martyrdoms.
A planning commander tells The Cradle: "There is an area that can only hold, for example, eight brothers for defense... Any extra brother is effectively a martyr or wounded. In Uli al-Ba's [Possessors of Great Strength], there was a major rush on several fronts that could not be controlled, and this is what increased the number of martyrs."
A field commander puts the problem more concretely: "During my rounds, I would see excess numbers of fighters to the point that there were not enough trees for them to hide under... The lesson lies in studying the place, understanding the human need, movement lines, and the possibility of camouflage."
By contrast, what stands out in this war - at least in the Israeli narrative - is repeated talk of smaller groups, usually no more than five or six fighters, and sometimes only three or four at forward points, particularly in ambushes. That suggests the lesson was absorbed. In Hezbollah's own accounts, supply and rotation lines for fighters also improved and worked more effectively in the 2026 war.
Many of the villages and towns that witnessed fierce clashes in 2024 returned to the battlefield in 2026, though some names were absent because of the massive destruction Israel inflicted during the 15 months of the previous ceasefire agreement.
Adaisseh, a first-line confrontation point, saw intense clashes in the previous war but not in the current one, while Khiam was central in both. Taybeh and Rabb al-Thalathine in the eastern sector saw medium-intensity clashes in 2024 but became much hotter fronts this time. Beit Lif, the legendary Bint Jbeil, and Ainatha, among others, also stood out more clearly in the current round.
Even so, in both wars, Hezbollah worked to ensure that confrontation remained present along the main axes and within specific villages and towns, even if only to obstruct the enemy, for both symbolic and operational reasons.
Bayyadah, Maqam Shamaa, and the Ramiyeh-Qouzah-Aita al-Shaab triangle in the western sector remained active, as did Yaroun and Maroun al-Ras in the central sector, and Houla and Markaba in the east.
According to the former field commander, it was decided that each area would be handled on the basis that "the brothers would perform their duty there until their last breath," or withdraw from it, based on fire cover from the second and third echelons in the confrontation, using new tools.
"In other words, any spot emptied of resistance will not necessarily be empty of resistance, because there are several means of dealing with the Israeli army there."
As for the decision - whether to remain until the last breath, to hit and run, or to withdraw to another position or facility - it was left to the fighters on the ground to make autonomously and personally.
Lessons of the previous war Because of the severe blows at the start of the war - the pagers and the assassination of commanders - along with the atmosphere of security suspicion around all devices, the disruption of parts of the command chain, and gaps in several missions, the defensive plan did not function well during the first month of the 66-day confrontation.
In the second month, the pattern became clearer. Hezbollah's losses began to fall, while casualties among Israeli soldiers and vehicles increased with greater precision and effectiveness. For this reason, The Cradle's conversations with cadres revealed frustration over the timing of the ceasefire agreement and what followed on the ground.
One commander says: "With our full commitment to our assignment, the truce came at a time when we had begun to keep pace with the front and its requirements... Anyone who works in the military knows how difficult it is to stop when the fighter has regained the initiative. Even our use of qualitative missiles was subject to the organizational decision, not personal assessment."
Based on the details gathered from these sources, the confrontation can be divided into two phases.
During the first month, in the first-echelon villages, there were individual clashes with the occupation, or clashes involving groups that had remained steadfast there. But a full, coordinated defensive plan was difficult to implement. Field improvisation dominated, especially because communication had been cut with many fighters.
In the second- and third-echelon villages, anti-armor fire was difficult because of intensive drone and warplane activity, the absence of Hezbollah air and naval defenses, and Israel's deliberate preemptive bombing of any hill overlooking the battlefield, even if it had not been used, to deny the resistance any benefit from it.
The first month, in short, offered no possibility of establishing a clear defensive pattern, militarily or in security terms. The months of the "support" front during Operation Al-Aqsa Flood had already seen systematic Israeli efforts to prevent Hezbollah from forming any defensive operating pattern.
During the second month, clashes in the first-echelon villages intensified in accordance with the principle of "meeting" the enemy. This did not necessarily mean holding villages and towns in a static way. Fighters would withdraw or lie in wait inside a facility, then return to fight after heavy bombardment.
The effort often resembled martyrdom-style action. This pattern intensified in the western sector, while in the eastern and central sectors, "meeting" the enemy was harder because of exposed terrain and limited means. But whoever survived the bombardment continued to engage once enemy forces approached.
In the second- and third-echelon villages, the work of the Kornet and Almas teams improved noticeably, especially in the western sector - Blat, Zibqin, and other areas - where better visibility and easier firing allowed more tanks to be targeted.
Fighters would suddenly emerge from areas not recorded in the original defensive plan, as an additional security measure, then withdraw quickly. This newly developed tactic proved effective. The sources decline to publish the name of the force that carried it out.
By the second month, fire support from north of the Litani had also intensified and diversified, causing major Israeli losses, many of which were never announced.
The resistance's own scorched earth The final stage of the previous war included details now being disclosed for the first time, both in their nature and scale. Their effects became apparent quickly in the current war after further improvements and new techniques were introduced. When Israel's "scorched earth" approach made direct engagement impossible in many positions before any incursion, Hezbollah began developing its own version within the limits of its capabilities.
Alongside the reconnaissance fed back by fighters on the ground who could still reach command, Hezbollah began using its own strikes to gather live battlefield intelligence. Each hit became a way to identify nearby vehicles, rescue teams, redeploying soldiers, or new positions - and then strike them again. This produced what sources describe as a compound method of action known as "parallel reconnaissance."
For example, when an attack drone is launched against vehicles, its imaging is live, not recorded, and can be jammed - or when an Almas missile is fired, with live transmission and possible recording because it is controlled by fiber optics, that drone or missile transmits a live image of the wider position or adjacent vehicles.
At the same moment, coordinates are passed to artillery and rocket units north of the river: another position, nearby vehicles, rescue forces, or soldiers redeploying. These units need only minutes to enter the coordinates and hit the forces again.
To carry this out, Hezbollah used a two-stage pattern.
The first strike could be carried out by an attack drone with a 10-35 kg warhead, an Almas missile with a 7-15 kg warhead, or a Kornet with a range of 4-10 km. Its task was to land the first hit, inflict losses, and transmit new coordinates.
The second strike could be carried out by artillery of various calibers - 81, 105, 120, or 130 mm, depending on the required range - or by 107 mm rockets, the small Katyusha or Fajr-1 with a range of 12 km, 122 mm Grad rockets with a range of 20-40 km, upgraded Grad rockets with fragmentation warheads that disperse shrapnel before impact, or more advanced missiles such as Fajr-5, Malak-1 and 2, or Fadi-1 and 2. Their purpose was to deliver the second hit, then repeat the cycle.
If the first strike fails to produce confirmed losses, the second is designed to do so. And even when neither lands decisively, the attacking force is still thrown off balance, the advance loses momentum, and fighters along the contact lines gain the opening they need to move or strike.
In the final weeks of the war, Hezbollah used the "second strike/subsequent strikes" pattern in ways Israel did not expect, and which would become the anticipated solution. It used cruise missiles, for example, against forces advancing into Yaroun; Nasr-1 and Nasr-2 missiles against forces in Mays al-Jabal; and Fadi-1 and Fadi-2 missiles, as well as Noor and Qader-1 and 2 ballistic missiles, against forces in Khiam.
According to one operations-room account, the resistance recorded 150 Israeli soldiers killed or wounded in a single strike on Khiam. But the live feed showed that many of the faces appeared "Asian," leading to an assessment that they were mercenaries previously observed during the 2023 standoff over Hezbollah's reconnaissance tents near the occupied Shebaa Farms.
Numerically, Hezbollah recorded hundreds of fires of this kind, with confirmed hits on more than 66 tanks and vehicles by the end of November 2024, in addition to dozens of troop gatherings in open areas or inside houses, where a special type of Kornet was used.
At the time, Hezbollah's official tally from 17 September to 27 November 2024 was 1,666 military operations, including 1,285 rocket attacks, 93 artillery fires, 166 drone attacks, and 86 guided-missile attacks using Kornet and similar systems.
Regarding rocket fire launched from south of the river until the end of the war, one fighter tells The Cradle:
"I had the opportunity to change the launcher's location with every strike, but we decided to take the challenge to the highest level. We would fire from the same point three or four times despite repeated airstrikes on it. We worked in a pattern that made the occupation literally sick of this war ... We had to make it feel that it was repeating the same mission more than once with no result."
The field result was that any Israeli army position, or anything resembling a semi-base, became a direct target and a "danger zone." This explains Israel's inability to occupy some villages and towns, despite the martyrdom of their entire garrisons, such as Adaisseh. It also explains how Hezbollah was able, in the final weeks, to move support forces through to places such as Khiam.
Israel, by contrast, followed a fixed sequence: enter after intensive fire-clearing, rush engineering units in to mine and demolish buildings, take showpiece photographs, then withdraw quickly. If the Israelis failed to kill the resistance fighters inside, they would quickly call in aircraft to bomb the entire building.
Resistance fighters attributed the speed of these airstrikes to the fact that every Israeli company commander had a drone - a Hermes, for example - directly under his command in the sky. This did not recur in the same way in the current war because of the confrontation with Iran, but Hezbollah also had its own solutions.
Hmmm. AI doesn't hallucinate. From what I've gathered it out and out lies and makes up evidence too.
I suspect the "Hallucination" idea is a handy way of making such lying more palatable...to pretend it's a temporary glitch: an aberration rather than an actual part of the program.
A bold statement with no evidence. I read AI experts on this subject and they assure me that they do (hallucinate).
From what I've gathered it out and out lies and makes up evidence too.
They do that too. These are the cheap versions that are not trained well.
Like I said, training AI's can be very expensive. Meanwhile the west is obsessed with production of these 'creatures' by increasing the proliferation of 'data centres' especially in the US. A wrong approach from my perspective. Chinese are better at it in many ways.
similar data together, extrapolating to create BS, depending on how global that data actually is.
The training data is compressed into weighted parameters, not a petabytes large database, rather like an autocomplete function.
One should incant first that the LLM: * Answers exclusively based on the search results * Does not connect the dots, infer or fill in gaps in sources. * If search results don't explicitly state a fact, return that the search results do not contain the answer.
I could tell you black was white I could tell you day was night Not only that, I could tell you why Back then, I could really tell a lie "I won't dignify the attacker's words by repeating them, they are horrific and vile."
on May 12, 2026, 2:37 pm
