Preflop Edge — Terms of Service

🎯 Grading & Scoring

Action Accuracy: Did you pick the exact best action (Raise/Fold/Call)? If the best action is "Raise" and you picked "Fold", that's wrong — even if folding was a close call. This is the strictest metric. Action Accuracy = (exact matches / total decisions) × 100
Decision Quality: A looser metric that counts "acceptable" mistakes as correct. If the best action is "Raise" but "Call" only loses 0.10 SB, that counts as correct here. This rewards understanding that some decisions are close calls. Decision Quality = (optimal + acceptable / total decisions) × 100
EV (Expected Value): The average amount you win or lose per decision, measured in SB (small bets). A +1.00 SB decision gains 1 small bet. A -0.50 SB decision loses half a small bet. EV is the mathematical foundation of poker strategy. EV = (% win × pot won) - (% lose × amount risked)
EV Delta: The difference between the action you took and the best action. A delta of -0.15 SB means your choice cost you 0.15 small bets compared to the best play. This is what determines severity.

Optimal Best action. No EV lost. EV delta ≥ -0.02 SB

Acceptable Small mistake. EV delta -0.02 to -0.15 SB

Leak Consistent mistake. EV delta -0.15 to -0.40 SB

Blunder Big mistake. EV delta < -0.40 SB

Why "Acceptable" counts as correct in Decision Quality: Poker has many close spots where two actions have nearly identical EV. If raising wins 52.3% and calling wins 51.8%, the delta is tiny (0.05 SB). We don't want to punish you for picking the "slightly worse" option when both are reasonable. This is how real poker works — the best players disagree on marginal spots.

A leak is a pattern of mistakes in the same direction. The coach detects these:

Fold-vs-Raise: You fold hands that should be raised for value or fold equity. Symptom: too tight. Fix: trust the equity and raise marginal hands.
Fold-vs-Call: You fold hands with enough equity to limp. Symptom: too tight in heads-up pots. Fix: completing the SB is cheap — take the discount.
Call-vs-Raise: You limp hands that should be raised. Symptom: too passive. Fix: raise strong hands to build the pot and gain fold equity.
Raise-vs-Fold: You raise hands that should be folded. Symptom: too aggressive. Fix: save chips on trash hands — fold is free.
Raise-vs-Call: You raise hands that should limp. Symptom: over-aggressive on marginals. Fix: keep the pot small with borderline hands.

Equity: Your hand's % chance of winning the pot at showdown vs specified opponents. AA vs KK heads-up ≈ 82%. Equity is calculated via Monte Carlo simulation — dealing thousands of random boards and counting wins.
Monte Carlo Simulation: A statistical method for estimating equity when exact calculation is too slow. We deal 10,000 random boards, count wins/ties/losses, and average. More simulations = more accurate. Equity ≈ (wins + 0.5 × ties) / total_simulations × 100
Vs Random Equity: Your hand's equity against N opponents playing random cards. This is what we pre-compute for the 169×8 grid. Real opponents don't play random, but this is the baseline.
Fold Equity: The % chance your opponent folds when you bet/raise. In the Play vs Bot trainer, you can adjust this with the fold equity slider. Higher fold equity makes raising more profitable. Effective Equity = Base Equity + (Fold Equity × Pot)
Effective Equity: Your equity adjusted for fold equity. If your base equity is 45% and you have 20% fold equity, your effective equity is higher because you win the pot immediately 20% of the time.

Hand Key (e.g. "AKs", "72o", "TT"): Two-card hand notation. First letter = higher rank, second = lower. Third letter (if present): s = suited (same suit), o = offsuit (different suits). Pairs have no third letter.
Examples: AKs = Ace-King suited, AKo = Ace-King offsuit, 72o = 7-2 offsuit, TT = pair of Tens.
Ranks: From high to low: A (Ace), K, Q, J, T (Ten), 9, 8, 7, 6, 5, 4, 3, 2. In Short Deck, 2-5 are removed (only A-6 remain).
Suits: ♠ Spades (black), ♥ Hearts (red), ♦ Diamonds (red), ♣ Clubs (black). Suited hands have two cards of the same suit, giving a higher chance of making a flush.
Canonical Hand: The 169 unique Hold'em starting hands (13 pairs + 78 suited + 78 offsuit). We group all 1,326 possible two-card combinations into 169 canonical types. Short Deck has 81 canonical hands (9 pairs + 36 suited + 36 offsuit).

SB (Small Blind): The first position. You post 0.5 SB forced bet. In the Play vs Bot trainer, you are always SB. This is the hardest position because you act first on every street.
BB (Big Blind): The second position. Posts 1 SB forced bet. In the trainer, the bot plays BB. If you fold, BB wins without a fight.
UTG (Under the Gun): First to act after the blinds. In N-opponent training, this means N players left to act behind you. More opponents = tighter range.
HJ (Hijack), CO (Cutoff), BTN (Button): Later positions with fewer opponents behind. Later = wider range because you see what others do first.

Texas Hold'em: 2 hole cards, 5 community cards (flop, turn, river). Standard 52-card deck. Make the best 5-card hand from your 2 + the 5 community. 169 canonical starting hands.
Short Deck (Six Plus Hold'em): Same as Hold'em but with cards 2-5 removed. 36-card deck, 9 ranks (A-6). 81 canonical starting hands. Flushes are rarer (and rank higher than full houses). Trips rank higher than straights. AA vs KK is only ~73% (vs 82% in Hold'em) because the deck is denser with high cards.
PLO (Pot-Limit Omaha): 4 hole cards, must use exactly 2. More hand combinations, higher equities. Coming soon.

Module: A poker variant package containing: equity data, hand list, grading engine, and grid layout. Each module registers with the PE_MODULES framework. Switch variants in the dropdown at the top of the page.
Leak Map: The 13×13 (or 9×9 for Short Deck) grid on your dashboard showing per-hand accuracy. Green = solid (>85%), Yellow = marginal (55-85%), Red = leak (<40%), Gray = not enough data.
Leitner Spaced Repetition: A learning system that schedules hands for review based on your accuracy. Get a hand right → see it less often. Get it wrong → see it more. 5 boxes with increasing intervals: immediate, 4h, 1d, 3d, 7d. The "Fix My Leaks" mode drills hands you've gotten wrong recently.
Fold Equity Slider: In the Play vs Bot trainer, adjust how often the bot folds when you raise. 0% = bot always calls. 50% = bot folds half the time. Higher fold equity makes raising more profitable because you win without showdown.