RangeSharp
RangeSharp
Step-by-step guide to drilling preflop ranges with spaced repetition. Pick a spot, train, review, repeat.
You've heard that spaced repetition works for poker. You want to try it. Here's exactly how to do it, step by step.
Start with a single spot. Don't try to learn BTN open, CO vs BTN 3-bet, BB defense, and SB squeeze all at once. You'll overwhelm the algorithm and yourself. Each spot has dozens or hundreds of hand-action pairs. The algorithm needs to learn your retention for each one. More spots means more items, means less repetition per item, means slower progress.
Good first spots: BTN open raise, or BB vs BTN defense. They're common. You'll use them every session. And they're contained. One decision point. One range to learn. BTN open is about 40% of hands. BB vs BTN defense is maybe 45%. Both are finite. You can get through a full review in one session.
I started with BTN open. It's 40% of hands. Manageable. And when I got it down, I had something I could use immediately at the table. Every session has button opens. I could feel the improvement. That kept me motivated for spot two.
Before you train, understand why each hand is in or out. Don't just memorize a grid. Know that A5s opens because of nut flush potential, playability, and wheel straights. Know that 72o folds because it has no equity and no playability. The edges matter. When you get a borderline hand wrong, understanding the logic helps you remember.
Open the range in your tool. Look at the 13x13 matrix. See which cells are raise, which are fold. Notice the patterns. Suited hands in, offsuit hands out (in early positions). The cutoff between call and fold. The cutoff between raise and fold. Those lines are what you'll be tested on.
Decision mode: you get dealt a hand. You choose the action (fold, call, raise, etc.). You get immediate feedback. Right or wrong.
In FreeRangeLab, you load the spot (e.g., BTN Open Raise), select Decision mode, and start. Hands come one at a time. You click your answer. The algorithm records the result. If you're right, that hand goes into the "review later" queue with a longer interval. If you're wrong, it comes back sooner. Maybe next session. Maybe the next day.
A typical session: 20-50 hands depending on the spot and how many the algorithm thinks you need to see. Some hands you'll see again because you got them wrong before. Some are new. The mix is automatic. You just answer.
When you get something wrong, look at it. Don't click through. Pause. Why did you think 98s was a fold? It's a raise. Look at the range again. See where 98s sits. Remember the logic. Suited connector, can make straights and flushes, plays well in position. Next time it shows up, you'll have a fighting chance.
The AI debrief (if your tool has it) can help. It explains why the correct action is correct. Use it for the hands you mess up. The algorithm can't fix a fundamental misunderstanding. You have to do that part yourself.
The algorithm schedules your reviews. You don't decide what to practice. It does. Wrong answers from today will show up tomorrow or the day after. Right answers will drift to longer intervals. Your job is to show up and do the session.
I do mine in the morning. Ten to twenty minutes. Coffee, trainer, done. It's part of the routine. The algorithm always has something for me. Sometimes it's a lot of new hands. Sometimes it's mostly reviews from previous sessions. I don't think about it. I just run the session.
Most spaced repetition tools track per-spot mastery. FreeRangeLab uses a 0-4 scale per spot. 0 means you haven't trained it or you're terrible. 4 means you've got it cold. The goal is to see that number climb over weeks.
Check your dashboard periodically. Which spots are stuck at 1 or 2? Those need more work. Which are at 3 or 4? You can ease off or add a new spot. The data tells you where to focus.
One position at a time. Don't jump from BTN to UTG to BB in the same week. Master one. Then add the next. The algorithm works better when you're not mixing too many spots. Your brain does too.
Focus on the edges. Obvious folds (72o, 83o) and obvious raises (AA, KK, AKs) are easy. You don't need to drill them. The borderline hands are where the money is. K9s, QJo, 65s, A4s. The hands that could go either way depending on the spot. Those should get extra attention. When you get one wrong, spend a few seconds locking in why.
Use Recall mode sometimes. Decision mode tests you one hand at a time. Recall mode tests whether you can paint the whole range from memory. You see the 13x13 grid. You highlight which hands are in the range. It's harder. It tests a different kind of recall. I use it once I'm solid in Decision mode for a spot. It catches gaps. "I know 98s is a raise but I forgot 87s." That kind of thing.
Start small. Your first session might be 15 minutes. That's fine. You're building the habit. The algorithm will accumulate enough data to schedule you properly after a few sessions. Don't try to do everything on day one.
You open the app. You have 23 hands due for review (the algorithm calculated this). You hit start. First hand: T7s on the button. You think. Suited, connected, plays well. Raise. Correct. Next hand: K3o. You think. King with a bad kicker, offsuit. Fold. Correct. Next hand: 54s. You hesitate. Small suited connector. Raise. Correct.
Then you get 96s. You say fold. Wrong. It's a raise. You look at the range. 96s is in. You note it. Suited, can make straights, discounted. Next time.
Fifteen minutes later you're done. Twenty-three hands. You got four wrong. Those four will show up again soon. The nineteen you got right will drift to longer intervals. You close the app. Tomorrow you'll do it again.
Over two or three weeks, the spot starts to feel automatic. You see K8s on the button and you know. No thinking. That's when you add the next spot.
Starting with too many spots. They add UTG, HJ, CO, BTN, SB, BB, and six different 3-bet configurations in week one. The algorithm is overwhelmed. The user is overwhelmed. Nothing sticks. Add one spot. Master it. Then add another.
Skipping the study step. They want to jump straight to training. "I'll learn by doing." You won't. You'll memorize wrong things. Or you'll memorize without understanding, and when a similar spot comes up you'll have no framework. Always study the range first. Know why hands are in or out. Then train.
Ignoring the wrong answers. They get 54s wrong, see the correct answer, and move on. No processing. The hand shows up again in two days. They get it wrong again. The algorithm keeps bringing it back. They get frustrated. The fix: when you get something wrong, stop. Look at the range. Say out loud why the hand is in or out. Give your brain a chance to encode it.
Training only when they feel like it. Spaced repetition requires consistency. The intervals only work if you show up. If you train Monday, skip Tuesday through Friday, then do a two-hour session Saturday, you've broken the system. Daily beats sporadic. Even 10 minutes daily beats 70 minutes once a week.
Spaced repetition is boring. It's not exciting. It's repetitive. But it works. The same method that gets medical students through boards and language learners through fluency gets poker players through preflop. You show up. You answer. You come back when the algorithm tells you to. Over time, the ranges stick.
I've been using it for a few months. My BTN open accuracy is solid. My BB vs BTN defense is getting there. I still make mistakes. But I know when I'm guessing now. That's progress. And the algorithm is patient. It'll keep bringing back the hands I mess up until I get them. No judgment. Just repetition at the right time.
The weird part: you stop thinking about the trainer as "study." It becomes maintenance. Like brushing your teeth. You do it. It takes 15 minutes. You move on. The ranges stay in your head. That's the goal.
One more tip: if you're using FreeRangeLab, the paste-to-import feature is worth using. You can pull ranges from PioSolver, GTO Wizard, or other solvers and get them into the trainer in seconds. No manual entry. That lowers the friction for adding new spots. When your solver session produces a new range, import it immediately. Start training it the next day. Don't let it sit in a file somewhere. The faster you get a range into the spaced repetition system, the faster it sticks.
Master your preflop ranges.