Why We Built a Poker Coach That Talks Back

If you have ever tried to learn from a GTO solver, you know the problem. You fire up PioSolver or GTO+, spend five minutes configuring the node tree, and what you get back is a 169-cell grid of percentages. It is correct. It is also incomprehensible unless you already know what you are looking at.

We built RangeIQ because the existing tools are right, but they teach wrong. A solver gives you the answer. A coach teaches you how to understand it. RangeIQ is the coach.

The problem: solvers that speak in grids

Traditional solvers are powerful, but they have three flaws that make them poor teachers:

• They are mechanical. You pick a spot, wait for the solve, and stare at a matrix of frequencies. No context, no explanation, no sense of why QJs is a mixed open on the button.
• They are clunky. Setting up a single spot takes time. Comparing two stack depths means running two solves. The interface was built for researchers, not players trying to fix their game between sessions.
• They do not tell you what you got wrong. After hours of browsing nodes, you might notice that you fold AJo too often from the cutoff. Or you might not. There is no feedback loop, no drill that targets the leak, no score that tells you what it costs you in big blinds.

The result: most players buy a solver, look at a few spots, and go back to guessing. The data is there, but it stays locked behind a UI that was never designed for learning.

What a real coach does differently

A good poker coach does three things a solver cannot:

1. Understands your question in plain language. You do not say "show me the BTN opening range at 100bb." You say "is QJs a button open?" A coach translates intent into action.
2. Grounds every number in real data. A coach tells you QJs is a mixed open on the button, raising about 60% of the time and folding 40%, because the solver says so. Not because they feel like it. Not because they memorized a chart. Because the solve is right there.
3. Turns mistakes into drills. You fold AJo on the button once. The coach remembers. Next drill, it shows up again, spaced out so the fix actually sticks. Your Leak Score drops. You can see the progress in big blinds saved.

Grounded, not guessing

This is the part that matters most. We built our own solver from the ground up based on CFR+ (Counterfactual Regret Minimization), the same family of algorithms that power professional GTO study tools. Every number RangeIQ gives you: ranges, frequencies, equities, EV costs in big blinds, is backed by math, not hallucinated by a language model.

When you ask "how do I play AK against a 3-bet?", the coach calls a data tool that reads the actual solve for that spot. Then a grounding audit checks the answer: did every percentage, combo count, and bb cost come from a tool? If yes, it ships. If not, the answer is retried, or degraded to a numberless response rather than guess.

No other study app does this. Most "AI coaches" are chatbots with poker knowledge baked into their training data. They sound right until they are wrong, and you cannot tell the difference. RangeIQ is different because every number is verified against a real solve before it reaches you.

From question to drill to fix

The coaching loop is what makes RangeIQ stick:

• Ask a spot in plain English. "Is QJs a button open?"
• See the exact range on a 13x13 grid, with colors for raise, call, fold. Tap any hand for its frequency breakdown.
• Get scored on your biggest leaks. The Leak Score adds up your standing mistakes in big blinds so you know exactly what they cost you.
• Drill the hands you keep getting wrong. RangeIQ builds a stratified deck that prefers your leak hands, so you practice exactly what you need to fix.
• Track progress over time with spaced repetition (FSRS) and Bayesian mastery tracking. Fixes actually stick.

A fraction of the cost

A real GTO coach charges $100 to $300 per hour. A solver subscription runs $50 to $100 per month, and you still have to teach yourself how to use it. RangeIQ gives you grounded coaching, targeted drills, and a leak tracker for $12.99/month or $99.99/year. The free tier gives you 30 actions on day one (enough for several coaching questions and a drill set) and scales down as you become a regular.

There is no feature paywall beyond the daily action limit. Chat coach, ranges, drills, courses, Daily Spot, progress tracking, and analytics are all available to free users. Premium unlocks unlimited actions.

The AI-native difference

In an AI-native world, the right question is not "can a chatbot answer poker questions?" Of course it can. The question is: can it guarantee the answer is right?

RangeIQ's answer is the grounding audit. Every response is checked against real solve data before it reaches you. If the coach cannot verify a number, it will not state one. This is not a chatbot with poker flavor. It is a solver with a coach's voice.

That is why we built it. Not because solvers are bad, but because they are incomplete. They give you the map but no guide. RangeIQ gives you both: the exact data from a real solve, and a coach that walks you through it in plain English.