- GTO is a virtual game where every player is perfect. Nobody can beat it, and it beats nobody by much.
- MDA is the real game: how real players play, measured on 379M real hands.
- Where they agree, play GTO. Where they don’t, the gap is your profit.
- You need both: GTO is the baseline, and MDA says where to leave it.
What is the difference between GTO and MDA?
GTO is the theory of a perfect game. MDA is the measure of the real one.
GTO and MDA, side by side
| Compare | GTO | MDA |
|---|---|---|
| Answers | What can nobody exploit? | What do real players do? |
| Built from | A solver, with game theory and no hands | Hundreds of millions of real hands |
| Opponents | Perfect, virtual | Real, by room, stake and player type |
| Wins | Never loses, wins little | Wins the most, where the pool deviates |
| Limit | Blind to mistakes | Needs hands, because thin spots are noisy |
pokalab.comA solver plays against a copy of itself: both players are perfect, so every mistake is assumed away. That makes GTO the safest strategy and the right baseline. But real players are not copies of a solver, and every spot where they deviate is money GTO leaves on the table.
Is GTO or MDA better?
GTO makes your bluff-catcher worth 0. Real players make it worth money.
What the same river call earns, against GTO and against real players
In big blinds per call, after a 75% pot bet, on GGPoker.
pokalab.com- Against GTO, the call is worth −0.16 bb: the solver bluffs just enough to make it break even.
- Against NL50 regs, it is worth +0.95 bb: 34% of their bets are bluffs, against 29% for GTO.
- The answer: GTO tells you the call is close, and MDA tells you it wins.
GTO never answers “how much can I win here?”: by design, it makes the opponent’s choices worth the same. MDA answers it, because it measures what the opponent really does. The full spot, in both seats, is in What is MDA.
Where do real players differ from GTO?
Not everywhere, and that is the point.
How often NL25 regs fold to a river bet, next to GTO
Regs against regs in a 2-bet pot, on GGPoker.
pokalab.com- After two checks, regs fold 68.5% to a big blind lead, GTO 39.9%. Bet those rivers.
- After a c-bet and a checked turn, regs fold 59.3%, GTO 47.1%.
- Facing a triple barrel, regs fold 57.0%, GTO 57.8%. The regs play it like GTO.
The pool is not uniformly bad. Facing three barrels, NL25 regs fold almost exactly like the solver: there, GTO is already the best play. When the action was checked, they give up far too often: there, the profitable play is far from GTO. Only real hands tell the two apart.
When should you play GTO?
When the pool plays like GTO, or when you don’t know who you face.
- The pool is close to GTO — like the triple barrel above. Deviating only costs you.
- An unknown opponent — GTO cannot lose against anyone.
- A thin spot — too few hands for MDA to be sure. Read a broader spot, or ask Pokai.
Everywhere else, with a measured gap on enough hands against the pool you actually play, move toward the exploit, and only as far as the gap.
How to use both
Learn the baseline, read the gap, practice the exploit.
- Build your GTO baseline — Solutions and the Trainer.
- Exploit your opponents — MDA: the pool’s fold, call and raise next to GTO’s, and what each play earns.
- Practice your exploits — MDA Solutions: the solver re-run against the real pool, and the MDA Trainer.
Questions
What is the difference between GTO and MDA?
GTO (game theory optimal) is the strategy nobody can exploit, computed by a solver in a virtual game where every player plays perfectly. MDA (mass data analysis) measures how real players actually play each spot, on a very large database of real hands. GTO is the theory; MDA is the real game.
Is GTO or MDA better?
Neither alone. GTO is the baseline that cannot lose against anyone; MDA shows where real players leave it, and by how much. Against GTO, a river bluff-catcher's call is worth about 0; against NL50 regs in the same spot, it earns +0.95 bb per call.
Should I study GTO or exploitative poker?
Both, in that order. Learn the GTO baseline first, so you know what an unexploitable strategy does. Then use MDA to find the spots where your pool deviates, and move only there. Where the pool plays like GTO, keep playing GTO.
Is GTO dead because of MDA?
No. MDA needs GTO: a gap means nothing without the baseline it is measured against. And in many spots the pool plays close to GTO — there, GTO stays the best play.
All the numbers
NL25 regs facing a river bet
| Spot | Line | Regs fold | GTO folds |
|---|---|---|---|
| BTN c-bets, checks back the turn, bets the river | XBC-XX-XB | 59.3% | 47.1% |
| BTN bets the river after flop and turn check-check | XX-XX-XB | 74.2% | 57.0% |
| BTN triple barrels | XBC-XBC-XB | 57.0% | 57.8% |
| BTN checks back the flop, bets turn and river | XX-XBC-XB | 59.6% | 57.4% |
| BB leads the river after flop and turn check-check | XX-XX-B | 68.5% | 39.9% |
| SB c-bets, checks the turn, bets the river | BC-XX-B | 56.4% | 38.9% |
| SB triple barrels | BC-BC-B | 51.5% | 40.8% |
| BB bets the river after flop and turn check-check (vs SB) | XX-XX-XB | 72.1% | 48.5% |
pokalab.comWhat we measured
- The call — MDA Explorer, NL50 GGPoker, regulars vs regulars, 100 bb, line
FFFRFC-XBC-XX-X, the button’s 75% pot river bet; Call EV = what a hand that beats only bluffs earns by calling, in big blinds. Read on 2026-10-28. - The river spots — the MDA module’s own rows, GGPoker NL25 regs facing regs in heads-up 2-bet pots, 80–120bb; GTO = Pokalab’s solutions. Re-measured October 28, 2026.
- Hands — real-money hands only, never simulated, never user uploads.
What this does not say
- One pool at a time — NL25 GGPoker is not NL500 PokerStars. Read the pool you play.
- Frequencies, not hands — which hands to bet is the next step, in MDA and MDA Solutions.