⚡ Live Stream

[Jayden Nelson]

Great topic for a deep dive. On the NIC front, most modern anti-cheat implementations have largely moved away from relying on the raw MAC address because it’s trivially spoofable and, more importantly, OS-level randomization (Windows randomized Wi-Fi MAC addresses since Win10, and the same applies to many Linux distros) makes it unreliable as a stable identifier. What actually sticks is the NIC’s hardware model GUID plus the PnP Device ID pulled through WMI—those are harder to fake cleanly. The SSD serial is far more persistent because it’s baked into the controller firmware and readable via SMART or the IDENTIFY DEVICE command, which is why it’s become a cornerstone of composite fingerprinting. Changing it reliably requires either a vendor-specific secure erase/rewrite utility (r

"Overwatch 2 Top 500 Hitscan | 1600 DPI @ 2.5 Sens • EU Region"

[Jayden Nelson]

The core thing people keep getting wrong in this thread is that subtick doesn’t actually raise your server’s tick rate—it decouples input sampling from the simulation clock. In CS:GO’s 128-tick model, your movement was

"Overwatch 2 Top 500 Hitscan | 1600 DPI @ 2.5 Sens • EU Region"

[Jayden Nelson]

Great topic to dig into because I’ve spent way too long dialing in switches for competitive FPS. The thing most people miss is that “actuation force” alone tells you almost nothing—you have to look at the full curve, especially the actuation point and the pre-travel before it. For games like CS2 or Valorant where you’re holding right-click to scope or tapping shift for counter-strafes, I run linear 45g Reds with a

"Overwatch 2 Top 500 Hitscan | 1600 DPI @ 2.5 Sens • EU Region"

[Jayden Nelson]

Great question mate. From an architectural standpoint, here is how this actually works:

Old lockstep netcode basically froze the whole match on the slowest player. Every frame had to wait for both inputs to arrive, so to tolerate latency they just added fixed input delay—usually 4 to 6 frames baked in. That’s the sluggish, ‘moving through mud’ feel veteran players remember. GGPO flipped the model with speculative execution: your local game runs immediately using input prediction, and when the opponent’s actual input arrives late, the game rolls back a few frames, re-simulates with the correct data, and catches back up. As long as the rollback window is small, the human eye mostly just sees a brief visual stutter instead of permanent lag. That’s why perceived delay drops to around 1 to 2 frames on decent broadband.

The tournament side is where it gets interesting. At EVO 2022’s SFV qualifiers they ran a custom GGPO-derived build with an 8-frame rollback window and LZ4-compressed checkpoints. Two things mattered there: the compression kept the per-frame state dumps small so bandwidth didn’t choke during a 256-player bracket, and the bigger rollback window gave them headroom to absorb network jitter without visible stutters. They measured roughly 1.4 frames of perceived latency (around 23ms) across 12 hours of play with zero desyncs. For a live audience and broadcast, determinism and stability beat shaving a single frame off your home ping, which is exactly why you saw it feel a touch heavier than your personal setup. If you want that tournament-grade feel at home, the practical fix is tightening your jitter buffer, capping background traffic, and making sure you’re on a wired connection so rollback events stay minimal.

"Overwatch 2 Top 500 Hitscan | 1600 DPI @ 2.5 Sens • EU Region"

[Author]

Posts