Bots Don't Play Poker. We Make Sure of It.

You know the feeling.

You’re three hours into a session. You’re playing well — reading the table, picking your spots. And then there’s that one player. Always in position. Never a timing tell. Folds in exactly the same amount of time when they’re behind. Bets with machine precision when they’re ahead. You lose a big pot and you sit back and think: is that a real person?

Maybe it was. Probably it was. But the fact that you had to wonder is the problem.

Bots are real. They’ve been documented on platforms far larger than anything in the Texas market. Players report them. Operators acknowledge them. Ban waves happen. And the usual response is a manual support ticket, a review process, and a ban that comes weeks after the damage is done — long after that account ran its edge into your stack.

We think that’s backwards. Fair play isn’t a feature you bolt on after launch. It’s something you design for from day one. Here’s exactly what we’re building.

Behavioral Analytics

Humans are beautifully inconsistent. We hesitate. We mis-click. We take longer on tough decisions and snap-fold the easy ones. We tilt after a bad beat. Bots don’t do any of that — they’re suspiciously consistent in ways that stand out when you know what to look for.

Every account on salty.poker builds a behavioral baseline over time: action timing distributions, decision velocity by hand strength, interaction rhythms between hands. We’re not just clocking how fast you act — we’re building a statistical fingerprint of how you play. When an account’s patterns deviate from human norms in the right combination of ways, it flags for review.

We don’t publish the exact thresholds. That part stays vague on purpose.

Step-Up Challenges

A behavioral flag doesn’t automatically ban anyone — that’s not how we want to operate. False positives are real, and a wrongful ban is a terrible player experience.

Instead, risk flags trigger step-up challenges: CAPTCHA variants and puzzle checks that introduce lightweight friction at the right moment. A human barely notices. A bot fails reliably. The challenge type and difficulty scale with the risk signal — a mild flag gets a mild challenge, a strong flag gets something harder. Most legitimate players never know it happened.

Device and Account Risk Scoring

Bots don’t appear from nowhere. They cluster. Same device fingerprint across multiple accounts. Account creation velocity that doesn’t match organic signup patterns. Activity levels that don’t align with account age.

We track risk at the device and account level, not just the session level. IPQualityScore feeds us real-time device and IP risk scores — flagging VPNs, datacenter IPs, known proxy infrastructure, and devices with prior fraud history across the broader web. That signal combines with our own velocity checks and account age heuristics into a composite risk score that follows an account everywhere it goes on the platform.

A sophisticated bot ring looks very different from a single human player, even when it’s trying not to.

Rapid Enforcement with a Review Path

When enforcement happens, it happens fast — not weeks after a support ticket. Flagged accounts get restricted immediately pending review, not after the damage is done. And there’s a documented appeal path for the rare case where we got it wrong, because we will occasionally get it wrong and players deserve a fair process.

Why This Matters More in a Smaller Pool

Here’s the thing about a regional platform: a smaller player pool means a bot has proportionally more impact. One bot at a $1/$2 table on a platform with 50 active cash game players is a much bigger problem than one bot on a platform with 50,000.

We know that. It’s part of why we treat enforcement as infrastructure rather than an ops task.

One More Thing

We pay close attention to what’s happening across the broader online poker landscape — the health of platforms, the patterns that emerge over time, the things that go wrong before anyone says anything publicly. That visibility informs how we build.

We don’t point fingers. We just build accordingly.

You shouldn’t have to wonder if you’re playing a person. We’re making sure you don’t have to.