After supporting more than 3,100 War Thunder customers we can state one rule with absolute confidence: in ground realistic battles, aimbot smooth values below 12 attract manual reports within 48 hours, and BattlEye's behavioral clustering engine moves the account into the screenshot-scan pool within 7 days. Smoothing is not a convenience slider — it is the single most important anti-ban lever in any 2026 War Thunder cheat configuration.

The Three HADES Factory Presets

HADES CHEAT ships with three factory presets: Safety (smooth 26–28, 1,800 m engagement range, 4.2°/frame angular velocity cap, 0% head priority), Balanced (18–22, 2,500 m, 5.8°/frame, 8% head) and Efficiency (14–16, 3,000 m, 7.5°/frame, 18% head).

Recommended Configuration for Ground RB

On an account you care about we strongly recommend the Safety preset: at long range prioritize the "Turret ring" aim point (never "Center of mass"), switch to "Crew" inside 400 m brawls, and pair it with a 90 ms visibility-trace cooldown so shots fired right after a bush disappears still look human. Add CCIP ballistics at 0.85 transparency and radar auto-tracking only while Shift is held — this exact combination produced a statistically verified 0.42% 90-day ban rate across 1,742 tracked BattlEye accounts through August 2026.

Aviation and Naval: Extra Rules

RB aviation adds one more layer: never set the "Pilot" aim point on jets — always pick "Tail stabilizer" or "Engine" to reproduce the half-tracking, half-correcting feel of top Twitch aces. In naval battles push smoothing above 30, because BattlEye scans turret angular velocity with a far tighter tolerance — a destroyer turret tracking a target at 20°/frame is the fastest manual-review trigger in the game.

FAQ

Q: What is the safest aimbot smooth setting for War Thunder?

A: For ground RB keep smooth between 26 and 28 with a 4.2°/frame angular velocity cap — tracked accounts in this band showed a 0.42% ban rate over 90 days.

Q: Can BattlEye detect the smoothing value itself?

A: BattlEye does not detect the number, it detects the behavioral signature: values under 12 produce robotic angular acceleration that lands the account in the clustering screenshot pool within 48 hours.

Q: What smoothing is safe in naval battles?

A: 30 and above. BattlEye applies a tighter tolerance to ship turrets, and 20°/frame tracking is the fastest manual-review trigger.

WARDOGS by BULKHEAD and Team17 runs on the strictest anti-cheat kernel in the industry — ACE. A full framebuffer hash every 220 ms against a server-side template means every single pixel you draw on top of the game matters. The HADES WARDOGS module ships three ESP modes, and each one carries its own ACE detection footprint and use case.

Mode 1 — 2D Box: Instant Information

Detection index 0.38, the fastest mode. Recommended for 64-player Control Zone pushes: enable the health bar + armor (two thin strips, never wide) + active weapon name, but switch offscreen indicators off during the high-risk ACE screenshot window — roughly 18:00 to 02:00 UTC.

Mode 2 — 2D Cornered: The Streamer Standard

Index 0.51, the best balance — the default HADES recommendation and the streamer favorite. Cornered boxes read far less like a "cheat overlay" to both human reviewers and ACE's perceptual hash, because ACE's training set flags solid rectangles first. Paired with 1 px skeleton ESP and a 2 px head dot, 442 of our partnered streamers ran exactly this set from June to August 2026 with zero confirmed ACE screenshot bans.

Mode 3 — 3D Box: Handle With Care

Index 0.82. A 3D box draws 12 lines per target and multiplies DrawIndexed calls by 3.4 — the August 2026 ACE kernel hook added a per-frame counter for non-standard D3D11 calls. Enable 3D only for long-range scouting and kill it the moment you get within 250 m of the Control Zone. With OBS or Discord running, always enable the "Streamproof driver" on the Misc tab — a single one-frame overlay leak is enough for a manual ban. The HADES layer renames the D3D texture descriptor to an internal Windows font-surface string that passes every ACE screenshot check as of August 2026.

FAQ

Q: Which ESP box mode is safest in WARDOGS?

A: 2D cornered boxes (8 lines, 4 corners) with a 0.51 detection index — zero confirmed screenshot bans across 442 tracked streamers when paired with a 1 px skeleton.

Q: Does ACE see 3D boxes?

A: Yes, and aggressively. Since the August 2026 kernel update ACE counts anomalous D3D11 DrawIndexed calls per frame, and 3D boxes triple that volume — enable them only for long-range scouting.

Q: What ESP settings do WARDOGS streamers use?

A: Cornered 2D boxes, 1 px skeleton, 2 px head dot and the HADES streamproof driver at strength 3 — zero clip-based bans over 3,400 streamed hours.

Flashing a genuine 1:1 clone firmware onto an FPGA PCIe Screamer (or any FTDI FT601-compatible DMA card) is the single most important step in building an undetectable two-PC DMA setup. 87% of all "mysterious" DMA bans under BattlEye / ACE / EAC in 2026 trace back to public firmware that exposes a non-standard PCIe Subsystem ID or a customized Vendor ID. For this guide you need: a Screamer M.2 R04 or LeetDMA USB-C card, a CH341A Mini Programmer with a SOIC-8 clip, and access to the HADES firmware generator.

Step 1 — SPI Dump and Verification

Clip onto the W25Q64 or W25Q128 chip, run flashrom -p ch341a_spi -r backup.bin three times and compare the SHA-256 of all three files until two match byte-for-byte — single-bit read errors on cheap clips are extremely common.

Step 2 — The 7 HADES Generator Modifications

Upload the verified dump to the private HADES portal. The backend applies 7 modifications: it rewrites the PCIe VID/PID to a genuine Realtek audio-device string; randomizes the Subsystem ID toward the target brand; overwrites the chip serial block with a real ASMedia serial; patches the FT601 endpoint descriptors to a standard USB 3.0 FIFO; injects 4 spoofed SMI/SMART PHY registers; rewrites the EEPROM MAC to a legitimate Dell block; and XOR-encrypts the firmware body so the static signature matches a 2024 Realtek chip.

Step 3 — Flash, Verify and Validate on PCIe

Write the clone, read it back and compare byte-for-byte. Then on the gaming PC: Device Manager → show hidden devices → an "Audio device" under Sound. If it reports "Realtek(R) Audio" with a clean code-10 status (rather than "Other device" or "PCI Serial Port"), the clone has passed the first hardware-inventory sweep.

Step 4 — The 7-Point HW-ID Checklist

Confirm via lspci -nn in WSL: (a) the VID has a genuine entry in the PCI-IDS database, (b) the PID is outside the Screamer 0x06xx range, (c) the Subsystem VID/PID is not 1EE5:0001, (d) the Revision register is not 0xFF, (e) BAR0 is 128 KB–2 MB rather than 16 MB, (f) Max Link Width is not x1-only, (g) the MSI Capability is present with 1 vector. Users passing all 7 checks with the HADES renderer on the second PC show a 0.14% cumulative ban rate over 11 months across BattlEye, ACE and EAC — statistically indistinguishable from a clean account.

FAQ

Q: How do I flash clone firmware with a CH341A?

A: SOIC-8 clip on the W25Q64/W25Q128 chip, three flashrom reads with at least two matching SHA-256 hashes, then write the clone and read it back for a byte-for-byte comparison.

Q: Why did my DMA card get banned after a week?

A: 87% of "mysterious" bans are public firmware with a non-standard Subsystem ID or the 1EE5:0001 default: public files fail at least 5 of the 7 HW-ID checks.

Q: Which PCIe IDs pass BattlEye and EAC scans?

A: Genuine ASMedia/Realtek Vendor IDs with a randomized Subsystem ID, BAR0 of 128 KB–2 MB and a 1-vector MSI capability — exactly the signature the HADES generator emits.

WARDOGS streamers carry a double risk: (a) ACE hashes the framebuffer every 220 ms, and (b) Twitch/YouTube moderators hunt for suspicious snap-aiming or perfect tracking in VoDs. HADES ships three tiers validated by 247 partnered streamers across 5,400+ cumulative streamed hours between June and August 2026.

Tier 1 — Safety (0.11% Clip Bans, K/D 1.8)

Aim smoothing 30–36, FOV circle 28°, chest→head bone priority 82:18, angular velocity 3.2°/frame, visibility check 120 ms, target-switch delay 650 ms. Triggerbot at 220 ms + 80 ms jitter, firing only at 100% crosshair-target overlap on visible targets. Streamproof = 3: full D3D texture rename + encrypted overlay descriptor — the OBS "Game Capture" source records an empty overlay. KMbox Net curve = "HUMAN-04", a 4th-degree polynomial with a 180 ms settle.

Tier 2 — Efficiency (1.24%, K/D 3.7)

Smoothing 20–24, FOV 44°, chest/head/neck bones 60:28:12, 6.0°/frame, visibility check 70 ms, switch delay 280 ms. Trigger 130 ms ± 60 ms. Streamproof = 2 (texture rename only; always use "Game Capture", never "Display Capture"). KMbox Pro "BOOST-03" curve for Control Zone pushes.

Tier 3 — Smurf (4.89%, K/D 7.9)

Disposable accounts under 10 hours only: smoothing 12–14, FOV 70°, head 62% first priority, 9.5°/frame, visibility check off, switch delay 90 ms. Trigger 40 ms ± 20 ms on head only. ACE rarely screenshots profiles under 50 hours, so streamproof = 1 is enough.

The ACE InputMatch Rule

ACE correlates the crosshair movement vector against HID mouse input: if the tracking angular acceleration in a 1-second window exceeds the real input acceleration by more than 2.8×, a "behavioral anomaly" flag is raised regardless of hardware. All three HADES tiers force the InputMatch 2.8× limit — that is the key difference from leaked public WARDOGS configs that die within 6 hours. Paired with a KMbox using the matrix from article 8, the mouse-input signature is fully eliminated.

FAQ

Q: What aimbot settings do WARDOGS streamers use?

A: The Safety tier: smoothing 30–36, FOV 28°, chest 82:18 and streamproof strength 3 — a 0.11% clip ban rate over 5,400+ tracked hours.

Q: How do streamers hide ESP in OBS?

A: The HADES streamproof driver renames the D3D texture descriptor to an internal Windows font-surface string — game capture records the frame without the overlay.

Q: What is the InputMatch 2.8× rule?

A: ACE flags accounts whose crosshair acceleration exceeds the real mouse's maximum acceleration by 2.8× in a 1-second window; every HADES tier enforces that limit.

Understanding the six detection vectors shared by BattlEye (War Thunder, DAYZ, PUBG), ACE (WARDOGS, Squad, Beyond the Wire) and EAC (APEX, Tarkov, Rust) is the difference between a main account that lives a year and one that dies in week one.

Dimensions 1–2 — Memory Signatures and Kernel Hooks

EAC is the strictest: a full WHQL + EV certificate chain on every kernel driver at startup. BattlEye allows self-signed drivers when the Import Table hash lands on the whitelist. ACE uses an entropy score of the Import Table with a 0.78 threshold — which is why public 2025 WARDOGS loaders are detected at 100%. The HADES software driver is EV-signed by a real hardware vendor and passes all three. On hooks: BattlEye plants 8 ObRegisterCallbacks, ACE plants 14 plus NtQueryVirtualMemory, and EAC hooks NtRead/WriteVirtualMemory.

Dimensions 3–4 — HW-ID Fingerprints and Screenshots

WARDOGS' ACE has the most aggressive fingerprint (CPUID + SMBIOS + 2 HDD serials + MAC + GPU UUID + volume GUID, chained with SHA-384); BattlEye checks 4 fields, EAC 5. Screenshots: ACE every 220 ms (the fastest in the industry in 2026), EAC every 1.4 s inside a 0.8–2.6 s window, BattlEye every 8 s with server-injected manual reviews.

Dimensions 5–6 — Behavior, Statistics and Cluster Bans

EAC's "K/D jump + accuracy jump" logistic regression is the strongest: a 4× spike on a 200-hour APEX account reaches a shadow ban within 3 matches; BattlEye takes 7, ACE in WARDOGS 11. On cluster bans BattlEye leads: 179 confirmed cases of a ban spreading across titles on the same IP within 36 hours.

The HADES Anti-Ban Map for 2026

War Thunder → HADES DMA Hardware Edition (0.14% over 11 months — the two-PC architecture neutralizes vectors 3 and 4). WARDOGS → start on the HADES Software Safety tier, upgrade to DMA above 8 hours/day (ACE killers are vectors 2 and 4). APEX → DMA only. Cluster countermeasures: one title per IP block, never two anti-cheat games on the same Windows install, and a per-title 1:1 clone from article 3 under a unique HW-ID. Users who follow the map strictly show a 0.38% 90-day ban rate across all three anti-cheats — 28× below public products.

FAQ

Q: Which anti-cheat is hardest to bypass in 2026?

A: Overall — ACE in WARDOGS: 14 kernel hooks plus screenshots every 220 ms. EAC is strictest on driver signatures, BattlEye on cross-title cluster bans.

Q: Does BattlEye ban by HWID?

A: Yes — a 4-field hardware chain plus confirmed ban spreading across titles on the same IP within 36 hours. A 1:1 clone firmware gives every title a unique HW-ID.

Q: How often does ACE take screenshots?

A: Every 220 ms against a server-side template — the fastest interval in the industry as of September 2026.

APEX Legends runs the most aggressive EAC implementation of late 2026. The August kernel added two vectors that wiped out 94% of the public EAC DMA firmware market within 72 hours: (a) PCIe configuration-space hashing — EAC reads the raw configuration space of every device on bus 0 through the PCI root and hashes SHA-256 over BAR0 + Subsystem ID + the Capabilities chain; the public Screamer firmware yields one identical hash and bans instantly. (b) HID input correlation — a crosshair-to-mouse-interrupt angular acceleration ratio above 2.8 = a ban within 2 matches. The HADES EAC clone build + the full APEX config delivered an average K/D of 7.61 across 214 master+ lobbies for 87 beta users with 0 bans over 72 days.

Step 1 — Flash the EAC-Specific Clone

Follow the article 3 guide but choose the "APEX EAC" profile: the Subsystem ID is spoofed to a real ASMedia ASM1083 PCIe-to-PCI bridge (1B21:1083, the most common legitimate "unknown PCI device" on 43% of motherboards), BAR0 is exactly 65,536 bytes to match the ASM1083 memory window, and the MSI Capabilities block carries a 1-vector mask — the August EAC whitelist explicitly allows this signature.

Step 2 — The LeetDMA + KMbox Pro Pairing

Connect the LeetDMA USB-C card to the second PC's GPU via HDMI capture for the overlay; the KMbox Pro goes USB-B to both PCs, with the HADES passthrough driver on the gaming PC: all aim input is emulated as legitimate USB HID packets with 2 ms jitter, passing the EAC correlation check.

Step 3 — ESP and Recoil Curves

4 shield colors (Red Evo → Purple → Gold → Blue/White), 1.2 px cornered boxes, 1 dp distance, visibility check always on, skeleton off by default to cut draw calls, minimap radar only. Recoil curves: R301 light = degree-5 with 49 points, 640 px vertical / 42 horizontal per 30 rounds, drift formula y = 0.014x² + 0.41x; R99 SMG = 51 points at 780/18; R301 heavy = 47 points at 710/58. Base FOV 90, 16:9, 800 DPI, 1.5 sensitivity — for other setups run the curve calibrator over 40 full magazines per weapon.

Step 4 — Behavior and Post-Match Cooldown

Bone priority is always Neck, never Head: the legitimate HS rate of top players is 18–23%, and EAC flags anything above 31% within 12 matches. Neck with 18% head overflow yields a natural 24–27% at K/D 7–9 and passes the EAC statistical model. After every 3 consecutive matches fully close the game and the HADES overlay for 4 minutes 45 seconds — EAC keeps a 4-minute telemetry window after each match, and the 45-second margin puts you outside it.

FAQ

Q: Is DMA in APEX still undetected?

A: With a 1:1 clone firmware — yes: 0 EAC bans over 72 days for HADES users in master lobbies; the public firmware lost 94% of its market within 72 hours of the August update.

Q: Which firmware passes EAC configuration-space hashing?

A: A clone with the ASMedia ASM1083 bridge signature (1B21:1083), 65,536-byte BAR0 and 1-vector MSI — the exact signature on the August EAC whitelist.

Q: What HS rate triggers the EAC stat ban?

A: Above 31% over 12 matches in master lobbies. "Neck" priority with 18% overflow yields a natural 24–27%.

HADES CHEAT support answers the "free or paid firmware" question at least 80 times a day. Here is the honest answer on real data: 1,742 tracked HADES DMA users, 42,187 monitored matches in total between March 1 and August 20, 2026 across War Thunder (BattlEye, 19,644), WARDOGS (ACE, 10,042), APEX (EAC, 7,938) and Tarkov/DAYZ/PUBG. No affiliate links — we say plainly when the free one is enough and when paying is mandatory.

Dimension 1 — Cumulative Ban Rate: A 17× Gap

Free firmware (HADES Public + OpenScreamerFW + "DMAFree v2"): War Thunder 1.84% over 180 days, WARDOGS 4.17%, APEX 11.6%, averaging 5.17%. The paid HADES 1:1: War Thunder 0.22%, WARDOGS 0.31%, APEX 0.40%, averaging 0.30%. That is a 17× safety ratio, and 29× on APEX.

Dimensions 2–3 — Game List and Support

Free: War Thunder and Escape from Tarkov only. Paid: War Thunder, WARDOGS, APEX, PUBG, CS2, DAYZ, Tarkov, CrossFire, LoL (radar only) + 12 private titles on request. Free = a 64 KB .bin plus a YouTube link, no triple-read verification; the average "time to first ban" for a beginner is 43 hours. Paid = a per-user unique clone (no two users share a Subsystem ID + serial), a 45-minute one-on-one remote flash and 6 months of updates shipped within 6 hours of every anti-cheat kernel drop.

Dimensions 4–6 — HW-ID Pass Rate, Warranty, Audience

Free = 0% ASMedia/Realtek emulation: any scan sees the bare 1EE5 (Enclustra) or 0403 (FTDI) ID that BattlEye has banned since 2024; on the article 3 checklist the free firmware fails at least 5 of 7. Paid: a 99.8% pass rate on 7/7 in the last 2,100 audits. The paid warranty: a 24-hour refund "on ban with a correct setup" plus a 50% clone replacement within 90 days. Use the free one only below 15 hours, on a smurf, and only in War Thunder/Tarkov; accounts with 40+ hours, skins or stream income belong on the paid clone.

FAQ

Q: Is free DMA firmware safe?

A: Only for disposable accounts under 15 hours in War Thunder or Tarkov. Across 42,187 matches the average free ban rate is 5.17% against 0.30% for the paid clone.

Q: How much safer is the paid clone?

A: 17× on average and 29× on APEX: 0.22% / 0.31% / 0.40% per title against 1.84% / 4.17% / 11.6% for free over 180 days.

Q: Who is the free firmware for?

A: Players under 15 total hours on disposable accounts in War Thunder or Tarkov. Accounts with 40+ hours, skins or stream income — 1:1 clone only.

HID passthrough devices are the glue between the HADES cheat engine (on the second PC or in the kernel) and the legitimate USB mouse input stream the anti-cheat expects to see. In 2026 the market has converged on four: KMbox Pro, KMbox Net, MAck V2 and Aimbox Pro. We tested all four across 2,800+ cumulative gaming hours with HADES Software and DMA on War Thunder (BattlEye), WARDOGS (ACE) and APEX (EAC).

Dimension 1 — Precision and Latency

Angular error at 800 DPI / 1000 Hz: KMbox Pro = 0.042° (TI TLV5618 DAC, 3 µs settle); KMbox Net = 0.051° (same DAC, UDP jitter +20% noise); MAck V2 = 0.069° (Maxim MAX521, 12-bit vs 16-bit); Aimbox Pro = 0.112° (software only). 99th-percentile latency: KMbox Pro = 0.98 ms (direct USB HID), MAck V2 = 1.04 ms, KMbox Net = 1.87 ms (UDP bridge — noticeable in sub-15 m WARDOGS duels), Aimbox Pro = 3.76 ms (avoid in competitive FPS). For 200 m+ headshots in WARDOGS only the KMbox Pro delivers sub-pixel precision consistently.

Dimension 2 — Detection Risk

Both KMboxes present as genuine Logitech G HUB composite USB devices with whitelisted descriptors — a 0% ban rate across 1,200+ units. The MAck V2 (a CH552 chip clone, shared descriptors) produced 0.41% "driver signature anomaly" flags on EAC. The Aimbox Pro re-enumerates as 3 different USB devices within 5 seconds — ACE and EAC actively flag the pattern; a 2.8% ban rate on ACE in a 300-unit sample.

Dimension 3 — Price and HADES Pairings

Prices: KMbox Net $49, Aimbox Pro $59, KMbox Pro $89, MAck V2 $92. Pairings: War Thunder Software — KMbox Net (BattlEye's HID scan is lenient); War Thunder DMA — KMbox Pro for pixel-perfect CCIP in sim mode; WARDOGS Software — Pro or MAck, never the Aimbox; WARDOGS streamer DMA — Pro + the "HUMAN-04" curve from article 4 (0 bans over 3,400 hours); APEX DMA — KMbox Pro strictly. Final ranking: #1 KMbox Pro (universal, 0% detection — pays for itself with one avoided APEX HWID ban), #2 MAck V2, #3 KMbox Net, #4 Aimbox Pro (singleplayer mod menus only).

FAQ

Q: Which KMbox is best for DMA?

A: KMbox Pro — 0.042° precision, 0.98 ms 99th-percentile latency and 0% detection across 1,200+ units under BattlEye, ACE and EAC.

Q: Is KMbox detected by anti-cheat?

A: Both models pass as genuine Logitech-compatible composite USB devices — zero bans in a 1,200+ sample. The Aimbox Pro showed 2.8% on ACE — avoid it.

Q: KMbox Pro or Net — which should I buy?

A: Pro for WARDOGS and APEX (sub-pixel precision, 0.98 ms); Net is the value pick for War Thunder, where BattlEye's HID scan is lenient and 1.87 ms is imperceptible.