According to Lichess (September 2026, N = 524,350 games), you score 47.3% as Black with Petrov's Defense. You answer with and hit the e4 pawn instead of guarding yours on e5. Strong players go 3. 4. : you kick the knight first, then win the pawn back. At your level, 36 in 100 opponents dodge the trade with 3..
Lichess games from September 2026 (30,577,162 rated 900 to 1700) and 31,791 over-the-board games rated 2300+ from the last 12 months. Data updated: October 14, 2026.
43 named lines in this family; 2 with their own page.
The opening's share in each Lichess rating band.
After 1. 2. , most club players at your level don't take on e5. 36 in 100 play 3., which guards e4, and 18 in 100 play 3.. Only 28 in 100 go for 3.. Stronger players flip that: 3. is half of all online games at 2200+, and 58 in 100 in over-the-board events at 2300+. Your side changes too. After 3., 34 in 100 of the public reply and 22 in 100 grab . Just 17 in 100 choose , which rises to 72 in 100 among strong players.
| Move | Lichess 900–1700 | White's score | Online, 2200+ | Over the board, 2300+ |
|---|---|---|---|---|
| 35.8% | 53.3% | 26.4% | 14.7% | |
| 27.7% | 52.5% | 49.9% | 58.2% | |
| 18.5% | 53.0% | 4.3% | no data | |
| 7.3% | 51.1% | 3.2% | 2.9% | |
| 6.8% | 54.1% | 15.4% | 24.2% | |
| 1.2% | 51.2% | no data | no data | |
| no data | no data | 0.2% | no data |
| Move | Lichess 900–1700 | White's score | Online, 2200+ | Over the board, 2300+ |
|---|---|---|---|---|
| 34.3% | 47.1% | 11.5% | no data | |
| 22.3% | 61.8% | 11.1% | 4.1% | |
| 17.2% | 52.4% | 72.3% | 95.9% | |
| 15.1% | 50.0% | 3.3% | no data | |
| 6.7% | 53.2% | no data | no data | |
| 1.7% | 52.2% | 1.9% | no data |
| Move | Lichess 900–1700 | White's score | Online, 2200+ | Over the board, 2300+ |
|---|---|---|---|---|
| 78.8% | 53.8% | 90.1% | 84.3% | |
| 9.2% | 50.1% | 3.5% | no data | |
| 5.9% | 49.3% | 3.8% | 7.4% | |
| 3.2% | 47.3% | 2.7% | 8.3% | |
| 1.5% | 39.3% | no data | no data | |
| 0.9% | 37.3% | no data | no data |
| Move | Lichess 900–1700 | White's score | Online, 2200+ | Over the board, 2300+ |
|---|---|---|---|---|
| 87.1% | 52.7% | 99.8% | 100.0% | |
| 4.9% | 59.0% | 0.2% | no data | |
| 2.7% | 62.8% | no data | no data | |
| 2.0% | 55.7% | no data | no data | |
| 1.3% | 59.4% | no data | no data | |
| 0.8% | 64.4% | no data | no data |
| Move | Lichess 900–1700 | White's score | Online, 2200+ | Over the board, 2300+ |
|---|---|---|---|---|
| 50.6% | 52.0% | 6.2% | 4.1% | |
| 13.5% | 51.2% | 8.2% | 13.0% | |
| 13.5% | 56.9% | 52.6% | 47.4% | |
| 9.2% | 51.0% | no data | no data | |
| 6.9% | 56.3% | 27.3% | 21.8% | |
| 3.5% | 49.9% | 1.9% | 9.6% | |
| no data | no data | 2.7% | 4.1% |
The Petrov meets an attack with an attack. White goes after your e5 pawn, and instead of defending it you go after e4 with . After 3. White has already collected, and it's tempting to collect straight back with . The trouble is the e-file. If White's queen reaches e2, it hits your knight on e4, which nothing protects. And if that knight steps aside, the e5 knight jumps to c6 with a discovered check and attacks your queen on d8. After you score 38 in 100.
The strong players' sequence clears that knight out of the way first. attacks it with a pawn, and 4. is the retreat 90 in 100 of them choose. Now there's no white knight on the e-file to unmask a check, and wins the pawn back with material level. That's why the move order matters: the pawn goes to d6 before you capture.
Your knight on e4 sits in the center, but nothing defends it yet. So the plan continues with pawns. 5. stakes out the center, and props up the knight and gives you a central square of your own. When 6. attacks the knight, that pawn already covers it. With you develop the bishop, clear f8 and get ready to castle kingside. The structure is symmetrical: pawns on d4 and d5, and an open e-file for both sides' rooks.
At your level, the most common answer to is 5., in half of all games. It's the same e-file idea: the queen pins your knight to your king, and nothing defends it. The knight can't move, so protect it before you finish developing.
If White doesn't take on e5, check who guards e4. After 3., nobody does, and picks up the pawn. After 3., the knight covers it, so there's nothing to win on e4 and your priority is protecting your own pawn on e5.
| Move | Strong games | Chosen by |
|---|---|---|
| 14,458 | 50.3% | |
| 7,264 | 73.4% | |
| 5,328 | 89.7% | |
| 4,774 | 99.9% | |
| 4,758 | 52.3% | |
| 2,574 | 88.7% | |
| 2,259 | 97.8% | |
| 2,198 | 42.9% |
A mistake. After 3., neither e4 nor e5 has a defender, and develops without cashing in. White cashes in first: 4. 5., and the pawn kicks your bishop while backing up the knight. After 6. you're a pawn down. You score 46.8 in 100 here, against 49.9 with .
22 in 100 play it in this position. It gives away 2.1 pawns of advantage. It scores 46.8%; scores 49.9%.
A mistake, though it pays off at your level. hands over the e5 pawn for development: this is the Stafford Gambit, which has its own page. You score 52.9 in 100 with it, more than with . Against someone who knows 4. 5., the d3 pawn covers e4 and you're simply a pawn down.
34 in 100 play it in this position. It gives away 1.6 pawns of advantage. It scores 52.9%; scores 47.6%.
A mistake. With 3. guarding e4, your knight no longer threatens anything, and abandons e5. After 4., hits c3 and e5, and 5. 6. levels the pawns. In return, White keeps both bishops. At your level it rarely gets punished: you score 48.5 in 100, slightly more than with .
23 in 100 play it in this position. It gives away 1.2 pawns of advantage. It scores 48.5%; scores 47.0%.
A mistake. After 3. 4., the bishop eyes f7, and guards e5, which the knight already covers. 5. adds a second attacker on f7, and after 6. 7. you've dropped a pawn. Play : if the c3 knight takes it, forks c4 and e4. You score 46.6 in 100 instead of 58.1.
8 in 100 play it in this position. It gives away 1.6 pawns of advantage. It scores 46.6%; scores 58.1%.
An inaccuracy. After 3., the e4 pawn is hanging, and lets the chance to take it slip. 4. covers it and the game slows down: 5. 6. , with the c3 pawn ready to support a push to d4. The cost is small: 47.2 in 100 against 49.9 with .
30 in 100 play it in this position. It gives away 0.6 pawns of advantage. It scores 47.2%; scores 49.9%.
A blunder. After 3., watches g5 but leaves e5 hanging. 4. grabs the pawn and lines up knight and bishop against f7. hits the bishop, but 5. takes that pawn too, and after 6. you're two pawns down. It leaves you on 39.7 in 100, against 49.9 after .
5 in 100 play it in this position. It gives away 3.3 pawns of advantage. It scores 39.7%; scores 49.9%.
| White | Black | Games |
|---|---|---|
| 198,407 | ||
| Not castled | 82,607 | |
| Not castled | Not castled | 74,552 |
| Not castled | 61,734 | |
| 37,685 | ||
| 33,435 |
| White | Black | Games |
|---|---|---|
| 336 | ||
| 54 | ||
| Not castled | Not castled | 42 |
| Not castled | 29 | |
| 24 | ||
| 21 |
1. 2. 3. · 148,641 amateur games · 13 over the board
3. guards e4 with the knight. It's the public's most common reply, with 148,641 games.
1. 2. 3. · 126,363 amateur games
3. leaves the e4 pawn undefended, and wins it: 126,363 games.
1. 2. 3. · 36,317 amateur games · 151 over the board
3. hits e5 with a pawn and leaves e4 without a defender: 36,317 games.
1. 2. 3. · 33,814 amateur games · 15 over the board
You grab before kicking the e5 knight: you score 38 in 100.
1. 2. 3. 4. 5. · 9,349 amateur games · 12 over the board
After , 5. pins your knight to your king: 9,349 public games.
1. 2. 3. 4. 5. · C42
3,017 amateur games · White's score: 55.9%
1. 2. 3. 4. · C42
2,499 amateur games · White's score: 50.2%
1. 2. 3. 4. 5. · C42
2,493 amateur games · White's score: 51.5%
1. 2. 3. 4. 5. · C42
1,728 amateur games · White's score: 50.6%
1. 2. 3. 4. · C42
1,564 amateur games · White's score: 49.2%
1. 2. 3. 4. 5. · C42
1,242 amateur games · White's score: 56.6%
1. 2. 3. 4. · C42
851 amateur games · White's score: 47.4%
1. 2. 3. 4. 5. · C42
638 amateur games · White's score: 50.0%
1. 2. 3. 4. 5. · C42
406 amateur games · White's score: 54.4%
1. 2. 3. 4. 5. 6. · C42
227 amateur games · White's score: 50.7%
1. 2. 3. 4. 5. · C42
200 amateur games · White's score: 54.2%
1. 2. 3. 4. · C42
158 amateur games · White's score: 50.9%
1. 2. 3. 4. 5. 6. 7. · C42
82 amateur games · White's score: 65.2%
1. 2. 3. 4. 5. 6. · C42
79 amateur games · White's score: 53.2%
1. 2. 3. 4. 5. 6. 7. 8. · C42
55 amateur games · White's score: 63.6%
1. 2. 3. 4. 5. 6. 7. 8. · C42
23 amateur games · White's score: 47.8%
1. 2. 3. 4. 5. · C42
17 amateur games · White's score: 38.2%
1. 2. 3. 4. 5. 6. 7. 8. · C42
12 amateur games · White's score: 83.3%
1. 2. 3. 4. 5. 6. 7. 8. 9. · C42
4 amateur games · White's score: 25.0%
1. 2. 3. 4. 5. 6. 7. 8. 9. · C42
1 amateur game · White's score: 100.0%
1. 2. 3. 4. 5. 6. 7. 8. · C42
1 amateur game · White's score: 100.0%
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. · C42
1. 2. 3. 4. 5. 6. 7. 8. 9. · C42
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. · C42
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. · C42
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. · C42
1. 2. 3. 4. · C42
At your level it does fine: you score 47.3 in 100 in public games, almost the same as the 47.2 in 100 strong players get with it online at 2200+. White starts with a small edge, about 0.42 of a pawn. The catch is move order: after 3., play before , or the e-file costs you.
Meet 1. with and 2. with , hitting e4. If White takes with 3., don't recapture right away: kicks the knight, and after 4. you take with . Then 5. 6. , with the d5 pawn supporting your knight and castling ready.
At 900 to 1700, the Three Knights Game with 3. leads with 148,641 public games, followed by the Italian Variation with 3. at 126,363. Then come the Stafford Gambit, 52,753 games after 3. , and the Modern Attack with 3., 36,317. The Damiano Variation, taking at once, has 33,814.