Heavy Flak Effectiveness: A Controlled Test Across Two Turns (Aug 1943, Ruhr)

[matchwood]

**TL;DR:** Ran a test using 8th Air Force to examine flak losses. Found that flak losses were barely discernible from normal ops losses even when flying over very heavily defended targets (Duisburg). Also found that what little flak effect there was largely occurred on the first bombing of the turn, with little to no effect on subsequent raids — by day 2 its effectiveness has collapsed, and from day 4 onwards it produces zero kills. Across days 2-7, bombers over heavily-defended Duisburg died at exactly the same per-sortie rate as bombers over undefended Krefeld (~0.94%). In other words, days 2-7 are effectively "free" airspace for the Allied bomber force regardless of how much flak you've stacked at the target. Also ran a test flying large numbers of 8th Air Force at 10k alt over massed flak — should have been a massacre, but barely any impact. This suggests that flak is not a good reason for the Allied player to run strategic bombing attacks at high altitude.

**Setup:**

Test target: Duisburg (111,179). Defended by 188 × 88mm, 64 × 105mm, 8 × 128mm, plus 45 × 37mm and ~60 × 20mm light AA, organised under Luftflotte Reich. Historically appropriate concentration for a major Ruhr city (Ruhr-wide heavy gun count was ~1,000 in 1943, spread across 4-5 cities).

Control target: Krefeld (110,180), an adjacent Ruhr city with **no flak units assigned** during the test period.

No Luftwaffe fighter intercepts in any of the test battles — Reserve mode not committed against these raids. Allied AI ran 2 bombing raids/day, D1-D7. This was a flak-only test by accident, but it makes the comparison clean.

Two consecutive turns at Duisburg, with the Allied player changing bombing altitude between them:
- **T7 (14 Aug 43): Allies bombing from 22,000 ft**
- **T8 (21 Aug 43): Allies bombing from 10,000 ft**

**The Krefeld baseline is identical to Duisburg.** Marginally higher, in fact. A city with no flak units assigned produced the same per-sortie loss rate as a city with one of the heaviest gun concentrations in Germany — at *either* bombing altitude.

This means the ~0.85-0.90% loss rate we observe at Duisburg is almost entirely **background operational attrition** — the wastage rate the model applies to all bomber sorties regardless of opposition. The flak's contribution above this baseline is, on a per-sortie basis across the full turn, barely measurable.

If background ops at Krefeld's rate (0.94%) applied to Duisburg's sortie count:
- T7 expected: 5,080 × 0.94% = ~48 losses (actual: 42)
- T8 expected: 4,887 × 0.94% = ~46 losses (actual: 44)

So 252 heavy AA guns at Duisburg produced **lower** total attrition than the undefended baseline would predict. The flak's nominal kill count (14 at 22k ft, 8 at 10k ft) is real, but it's not adding to the total at the expected magnitude — the columns are partly substituting for ops losses that would otherwise have occurred.

**The day-1-only pattern: days 2-7 are effectively free airspace:**

**Day 1 produces ~50% of all flak kills on ~17% of the sorties. Days 2-7 produce the other half but at one-fifth to one-seventh the per-sortie rate.** And critically, when you put this next to the Krefeld control (0.94% ops-only), it becomes clear what's happening:

- Day 1 at Duisburg: bombers face real opposition. Loss rate ~0.6-0.7% above the ops floor.
- Day 2 onwards at Duisburg: the per-sortie kill rate from flak collapses to the same level you'd see over an undefended target. By day 4, T8 produced zero flak kills across 788 sorties. T7 day 4 produced zero across 522 sorties.

In effect, **the Luftwaffe flak gets one good shoot per turn, and then the airspace over the Ruhr is free for the rest of the week.** A heavily-defended target on day 5 is mechanically equivalent to an undefended target. The 252 heavy guns at Duisburg may as well not exist for raids 3, 4, 5, 6 and 7 of the turn.

The mechanism is almost certainly ammunition depletion between sub-turns — the 64th(a) Flak Regiment showed Ammo/Need at 550/368, healthy on paper but evidently nowhere near enough to feed multiple 300-400 bomber intercepts across a 7-day turn. Possibly compounded by accumulated gun damage and crew fatigue, but ammo is the most likely primary cause.

The pattern held at both 22k ft and 10k ft — dropping altitude didn't restore flak effectiveness on the back half of the turn, it just shifted some of the day-1 kills from the flak column into the ops column.

**Altitude finding (T7 22k ft vs T8 10k ft):**

When the Allies dropped from 22,000 ft to 10,000 ft between turns, flak kills *fell* (14 → 8), ops *rose* (28 → 36), and total attrition was almost identical. What the Allied player got was a major jump in industrial damage:

- T7 final (after bombing at 22k ft): Resource 59%, Manpower 48%, Railyard 72%, Fuel 0%, Synth Fuel 0%
- T8 final (after bombing at 10k ft): Resource 88%, Manpower 67%, Railyard 69%, Fuel **67%**, Synth Fuel **72%**

If your Jagdwaffe isn't intercepting, the Allied player has every reason to drop altitude. The flak doesn't punish him enough to deter, and his bombing accuracy improves dramatically. Fuel and synthetic fuel were completely untouched at 22k ft and devastated at 10k ft. Historically, a 1,000+ aircraft formation lumbering over the Ruhr at 10,000 ft should have been a massacre — light flak and medium flak alike would have torn through them, plus the heavy guns engaging at near-optimal range. In the test it was a non-event.

**Historical comparison:**

The 8th AF over the Reich in 1943 was losing roughly 5-10% per sortie unescorted, of which flak alone contributed about 0.5-1.5%. Westermann's data gives ~4,000 88mm rounds per kill in 1941-42 rising to ~16,000 by 1944; the 128mm was dramatically more efficient at ~3,000 rounds per kill.

Historically, 252 heavy guns at a Ruhr city would have produced roughly 25-50 kills per week of sustained bombing at 5,000-sortie throughput. The game produces 8-14 in nominal flak kills, of which arguably only the day-1 spike (5-7) is signal above baseline.

So the in-game flak appears to be running at roughly **1/5 to 1/4 of historical lethality** when measured properly against a control — and almost all of that lethality is concentrated into a single day rather than sustained across the week the way real flak corps operated. Historical Ruhr flak fired every day, every raid, throughout 1943-44. The game's flak fires hard once and then goes quiet.

Worth noting that historically, flak's main contribution wasn't outright kills but damage — the 8th AF July 1944 survey attributed 86% of casualties in returning damaged bombers to flak rather than fighters. The game's Ops column may be absorbing some of this, but the Krefeld control shows that ops losses occur at the same rate whether flak is present or not, so it can't be capturing the flak-damage effect either.

**Practical takeaways for Axis players:**

1. **Flak only matters on day 1. Days 2-7 are free airspace for the Allies regardless of how much AA you stack.** This is the single most important takeaway from the test. Plan your turn accordingly — don't expect flak to defend anything except on the opening day.

2. **The Allied player who knows this schedules his valuable raids for mid/late week.** T8's hits on fuel and synthetic fuel — both untouched at 22k ft — only landed from day 5 onwards at 10k ft, exactly when flak had stopped contributing. Expect competent Allied opponents to deliberately save oil/synth fuel/jet plant raids for days 4-7.

3. **Stacking more flak gives you a bigger day-1 punch but doesn't change the day 2-7 cliff.** If you double the guns at a target you'll get maybe twice as many day-1 kills (~14 instead of 7), but you'll still get zero on days 4-7. There's no flak density that turns days 2-7 from "free" into "contested." If you're going to invest, concentrate at the highest-VP targets (Leuna, Pölitz, Brüx, Schweinfurt) to maximise the one day where flak earns its keep.

4. **Railyard repair is air defense.** Damaged railyards choke ammunition resupply between turns, which means even your day-1 spike gets weaker. A bombed Ruhr loses flak effectiveness on subsequent turns even before guns are damaged directly. (Note: T8's day-1 spike was already weaker than T7's — 4 kills vs 7 — possibly because T7's heavy bombing had degraded Duisburg's supply trace.)

5. **Light AA (20mm, 37mm) does almost nothing at Duisburg even at 10,000 ft.** Concentrate it on airbases and front-line LW HQ units where low-level fighter-bombers actually fly.

6. **Fighters on Reserve are the only thing that contests days 2-7.** This is the structural conclusion. Air groups regenerate sortie capacity daily and aren't bound by the per-turn flak ammo cycle — a Reserve-mode Jagdgeschwader can intercept on day 2, day 4, day 6, limited only by morale, fatigue, and ready aircraft. Without fighters, the Ruhr is a free bombing range from day 2 onwards every single turn. The flak day-1 spike is the *only* defensive contribution the city gets if your fighters aren't flying.

**Caveats:**

Two turns is still a small sample for Duisburg, though the Krefeld pool of 14 raids gives the baseline more weight. The Krefeld raids average smaller (~137 sorties/raid vs ~350 at Duisburg), so raid composition may not be directly comparable — possible that larger raids attract better-trained crews with lower ops rates, which would mean the flak signal is being masked. It's also possible some of what's recorded as "ops" at Duisburg is flak-damaged aircraft written off post-mission, in which case the flak/ops split is blurrier than the columns suggest.

That said, the headline finding — that flak produces a meaningful kill spike only on day 1 of the turn, after which days 2-7 are effectively undefended airspace regardless of gun count or altitude — is consistent across both turns and both altitudes, and matches what the Krefeld control would predict.

Interested whether anyone else has run a controlled test like this, particularly anyone who has tried very high flak densities (500+ guns at one target) to see whether the day-1 spike scales linearly, or whether the day-2-onwards collapse is delayed at all with much larger ammunition pools. Also interested in whether the day-1 effect changes with different commander/HQ skill ratings on the flak units.

[LiquidSky]

Hmm...I have to look into this. I do some things differently that might effect the results you are seeing. For example I build depots in every factory hex so they would all have full ammo regardless of rail yard damage. Also, I activate priority repair and attach a construction battalion for all critical damage so I know they are getting repaired every turn. So my Flak only has to lower the bombing enough for the priority repair to fix it. And in general, my factories do tend to repair faster then they get damaged...

I do like to move flak into my forward factories, but my precious admin points have to be spent on more important things first. Building depots, mobilizing divisions moved to Italy....disbanding AA in Sicily before it can surrender.

[matchwood]

I ran a small comparison test on an 8th Air Force raid against Duisburg to see what difference local flak makes. Basically nothing of any consequence.

The first test had flak present in the target city of Duisburg: **2 flak regiments and 2 railway flak battalions**. The second test removed all flak from Duisburg and the surrounding cities. The bombing directive and general setup were otherwise comparable. One important point: the raid was flown at **5,000 feet**.

That should matter. A heavy bomber raid over a major German industrial city at 5,000 feet, with substantial local flak present. I would have expected this to be close to a worst-case profile for the 8th Air Force — if not a bloodbath, then at least a very obvious increase in bomber losses, damaged aircraft, or bombing disruption.

But the results do not really show that.

Flak present:
Sorties: 13,544
Aircraft lost: 33
Aircraft damaged: 528

No flak:
Sorties: 13,996
Aircraft lost: 33
Aircraft damaged: 437

Aircraft destroyed were exactly the same in both tests: **33 lost**.

Where flak did have an effect was damaged aircraft. With flak present, the raid suffered **91 more damaged aircraft**, an increase of about **21%** overall.

Looking specifically at mission aircraft and escorts:

Flak present:
Mission aircraft damaged: 431
Escort damaged: 97

No flak:
Mission aircraft damaged: 332
Escort damaged: 105

So the extra damage is mostly falling on the bombers/mission aircraft, which makes sense. Mission aircraft damage rose from **332 without flak** to **431 with flak**, about **30% more damaged bombers/mission aircraft**.

However, the effect still feels surprisingly modest given the altitude. At 5,000 feet, with 2 flak regiments and 2 railway flak battalions present, I would have expected substantially more than a repair/readiness tax. Instead, there was:

no increase in aircraft destroyed;
only a moderate increase in damaged aircraft;
no obvious evidence that flak disrupted the raid enough to protect the target;
no obvious difference in bombing effectiveness compared with the no-flak run.

Duisburg was still heavily damaged in both cases. The no-flak run may have produced slightly higher visible damage overall, but the target mix varied enough that I would be cautious about claiming a clear bombing-accuracy difference. The broader point is that removing flak did not radically change the result, and keeping flak present did not stop the raid doing major damage.

I also ran some dive-bomber tests against defended targets, and they seem to take high losses. So this may be something more specific to level bombers rather than a general issue with flak.

My tentative conclusion from this test:

Flak is causing damage to level bombers, but it is not shooting down many aircraft and does not appear to be notably influencing bombing accuracy or target damage.

Matchwood

[LiquidSky]

As the Germans I am not concerned if I am not shooting down aircraft. Damaged aircraft don't drop bombs, and that is the real goal.....less damage means less vp's. You will never attrit the allies so much to make them lose more aircraft then they get.

In my game I noticed that the cities that I put depots in that also have flak and are bombed seem to use the flak everyday. If a depot is more then 3 hexes away I know that supply needs trucks to move, and it is quite possible that flak cannot rearm themselves if their are too few trucks and the depot is too far away.

[matchwood]

Thanks LiquidSky — that is a very useful point, and I think the depot/ammunition issue is worth testing properly.

In my Duisburg test I did not specifically build a depot in the city, so it is possible that I was testing heavy flak without optimal local ammunition support. The flak units looked supplied on paper, but that may not be the same thing as being able to sustain repeated large raids across all seven days of the air execution phase. I will run another version with a depot in Duisburg and see whether that changes the day-by-day pattern.

I also agree with the theory that damaged aircraft should matter even if aircraft are not being shot down. If flak is damaging bombers, causing aborts, reducing bombing effectiveness, or preventing aircraft from reaching the target, then that is arguably the more important effect than kills. From the German point of view, the aim is not really to attrit the 8th Air Force out of existence. It is to reduce bombing damage and protect VP targets.

The difficulty is that this effect was not very apparent in my data. In the 5,000 ft Duisburg comparison, flak did increase damaged aircraft, especially mission aircraft, but there was little obvious difference in the damage done to Duisburg with flak present versus no flak. Duisburg was still badly hit in both cases. So while the “damaged aircraft don’t drop bombs” theory makes sense, I am not yet seeing a clear corresponding reduction in target damage.

That may be because my test setup was missing the depot/ammo piece. It may also be because the difference only becomes clear across several turns, once repair, repeated raids, aircraft readiness, and cumulative damage are all taken into account.

I will do another test to see whether the presence of flak materially changes the bombing damage received by the target, rather than just increasing the number of damaged aircraft.

I was wondering whether you, or anyone else, has seen evidence that flak materially reduces city damage, rather than mainly increasing damaged aircraft?

[matchwood]

I ran several Duisburg bombing tests today comparing 8th Air Force raids with default local flak against equivalent raids with local and nearby flak removed. The pattern I am seeing is concerning. At 27,000 feet, local flak does not appear to make a meaningful practical difference to either Allied losses or target damage, while Duisburg Heavy Industry can still be destroyed very quickly. At 15,000 feet, flak does seem to increase bomber losses and damage, but even there it does not materially protect the target. Taken together, the tests suggest that very high-altitude raids may be getting the best of both worlds: low practical vulnerability to flak, while still delivering very fast, highly concentrated industrial damage. Losses at 15k from concentrated flak seem to be very low.

Thanks
Matchwood

[kentkroeckel]

Great insight Mr. Matchwood. I would like to add that in my playing I have found weather and night as an influence. Reason I mention this is that at least half of the year around the Ruhr there is bad weather. Additionally, many Allied aircraft are lost due to flak. I do not have specific numbers but I feel safe in saying that would indicate many bombers would have also been forced to abort. System says that happens. So I think flak can do an effective job in helping curtail Allied bombing.

I would like to know how much weather impacts bombing runs. How many bombers never reach target due to bad weather? Does this make bombing not worth while? I can say from experience that operational losses do increase, just as the system claims. But the big question is, how many bombers fail to reach target due to the specific bad weather?

During light rain, which is just called rain, and light snow, I find it worthwhile to continue to bomb. When there is heavy snow or heavy rain in the sky I do not bomb. Not worth it.

I have found your data on altitude attacks correct. Higher attacks for Allies do reduce losses from flak and aircraft. No surprise in that. But what I am surprised is the amount of damage that is still inflicted on German industry. During the bad weather, I find day bombing to be fine. I also find top altitude bombing for British to be fine. 18,000 feet to 23,000 feet, typically. Americans can bomb at 35,000 feet but I keep it at 25,000 feet to avoid mileage fuel shorting and fatigue. this seems to work fine given the parameter of altitude and bad weather.

Just wanted to add these extra variables into it. And thank you very much for your research Mr. Matchwood.