Schwarzacher Open 2026
39. Salzburger Sonnenterrasse Schach-Open 2026 — B-Turnier
New to chess tournaments?
Nine rounds over eight days: one long game per day, and two on one of them. A win is 1 point, a draw is ½, a loss is 0. Swiss system means you're always paired against someone on the same score as you, so the leaders meet the leaders. “Board 14” is just which table he's sitting at: the lower the number, the higher he's placed that round. Each player gets about two hours of thinking time, so a game can run five hours.
Exact time control: 90min/40 moves + 30min + 30s/move. Format: 9 rounds Swiss · Classical.
What does “Score” mean?
Score
A win is 1 point, a draw is ½, a loss is 0.
What does “Performance” mean?
Performance
Performance rating: how strong you'd have to be, on average, to score this result. Elias is rated 1798, so this number above 1798 means he is playing above his own level.
What does “Rating” mean?
Rating
How many rating points this tournament has added to his official FIDE rating so far.
Updated 9 Aug, 14:30
Tournament diary
One card per playing day. Open one and you get the long version.
Measured strength
Not what the results were worth — what the moves were worth. Every move of every analysed game, priced against the engine and turned into a rating with an honest error bar.
somewhere between 1785 and 1953, from 276 measured moves across 8 games — the middle number is the best guess, not the answer
this tournament's games on their own, weighted so the games that measured more confidently count for more
The same games with the single worst move of each one taken out — one per game, always the worst, blunder or not. That is what he plays when the catastrophic move does not happen: available to him, and not his rating. The 1869 above counts the blunders, because they happened; the 161 points between the two are what one move a game costs.
The detailed numbers — strength by phase, every game on its own, and how the estimate is made — are private. The headline above is the public part.
Standings
| Rk | Player | Fed | Rtg | Pts | Buchholz | Buchholz (cut) | S-B |
|---|---|---|---|---|---|---|---|
| 1 | Narr, Kevin | GER | 2059 | 7.5 | 41 | 52 | 43 |
| 2 | Donets, Alina | AUT | 1914 | 7.5 | 40 | 51.5 | 41.8 |
| 3 | Hamzic, Vahidin | BIH | 2018 | 6.5 | 41.5 | 53 | 38 |
| 4 | Putz, Lukas | AUT | 1617 | 6.5 | 36 | 46.5 | 31.8 |
| 5 | Blum, Gernot | GER | 2064 | 6 | 38.5 | 50 | 30.5 |
| 6 | Schindele, Andreas | GER | 1989 | 6 | 38 | 50 | 30.5 |
| 7 | Zesewitz, Maik | GER | 1930 | 6 | 38 | 49 | 29 |
| 8 | Zaynard, Andreas | AUT | 1930 | 6 | 38 | 48.5 | 29.5 |
| 9 | Fellmayr, Simon | AUT | 1943 | 6 | 33.5 | 43 | 27.3 |
| 10 | Teubner, Elias | AUT | 1798 | 6 | 33 | 41 | 24.5 |
The table above is the top ten plus his row.
Round by round
Games
vs Blum, Gernot (2064)
1–0 · LossChess scores are always written White first. Elias had Black here.
Board one, Black, against the highest rating in the field. Gernot Blum, 2064. Eighteen moves in, it was still a real game and I was fine. Then I thought for about a minute, played 19...Qf7 and hung my queen the second I let go of her. The engine had it level before that move and three pawns down after it. That is where you resign. I didn't. I tried to bail out by taking as many pieces as I could for her and making the squares around his king weak, so at least I had something dynamic to play for: a rook, then a bishop, then another rook. The equaliser was genuinely there. I had calculated 24...Bxd1, which the engine calls a playable +0.6, and I played 24...Rxd3 instead because I wanted the light-squared bishop for tricks and his king boxed in on the light squares. A move later I just missed the fork. 26.Qg8+ hits my king and the rook on a2 at the same time and there goes the second rook. 25...Bd5 was probably the move. After that the only idea I had left was walking my own king across to the queenside to help mate him, because he could barely move a piece and my king was incredibly safe over there. I analysed it at the board and I don't remember the moves any more. 31.h4 shut that down and that was the beginning of the end. I'm not even mad that I lost. I liked the game and I had a lot of fun dynamically balancing out the queen blunder and the missed equalisation I calculated but didn't play. Blum needed 41 moves to catch my king.
Use the arrows to step through the game, or drag the evaluation graph to jump.
Line above the middle = his opponent is better. Below = Elias is better. · Dots mark mistakes.
My alternative: bishop and rook for the queen with a healthy structure. Objectively this was the better version.
What I play instead of 25...h5 once I see the fork.
vs Viehhauser, Simon (1722)
1–0 · WinChess scores are always written White first. Elias had White here.
White, round two, straight after the round-one disaster. In general I felt like I was better at all times in this game. The engine says no. A Maroczy bind, my bishop parked on b6, his queenside frozen. That part was true, and it stayed true for about ten moves. Then on move 16 I won the exchange in the wrong order and gave two and a half pawns straight back without noticing anything at the board, and on move 20 I gave the b6 bishop away because I was certain nobody could stop my connected passed pawns. They could. From there the engine has me losing for fifteen straight moves. I spent those fifteen moves pushing the c-pawn up to c7, and then I got up and grabbed a second queen off the next board for psychological warfare, so it stood next to the pawn where he had to look at it. He cracked on move 34, and that pawn was stoppable the whole time. Then I was winning for exactly three moves. 37.Qxa6 threw it away one final time, and he shook hands one move before 37...Bd3 beats me. My judgement was wrong for most of this game and I still walked away with the full point. Positions like this are very hard to defend.
Use the arrows to step through the game, or drag the evaluation graph to jump.
Line above the middle = Elias is better. Below = his opponent is better. · Dots mark mistakes.
The check we looked at first in the post-mortem. It costs him nothing and White has to undo the knight on d5 straight away.
The clean-up version: trade the outpost off and put the knight on e5. Black gives me the d5-pawn as a target but takes all the tricks out of the position.
Our main line at the board: the same check one move later, and again the knights have to retreat instead of jumping into c7.
The other way to meet 13.Bb6. Take on d5 immediately instead of tucking the queen away on c8.
The knight recapture was our first candidate in the post-mortem. Practically it is the same story. He keeps the bishop pair, I keep the exchange.
Instead of castling I looked at pushing straight away. 18.c5 is fine. But we had the wrong answer for him: 18...d5? just drops a pawn to 19.exd5 and White is better by two. He has to hit the c5-pawn or play ...Qe6 first.
The move we found afterwards, and the engine's number one: push past instead of capturing. I still give up the bishop on b6, but the e-pawn takes the knight on f6 with tempo and everything holds together.
His queen belongs on c7, not d8. From behind the pawns nothing ever promotes. 22.d6!? is the try, and after ...exd6 cxd6 the pawns are split and the check on b6 comes with tempo.
The trap that made the pawns feel unstoppable. Blockading with a piece on d7 simply loses to cxd7, because the pawn eats the blockader. That is why the rook has to go behind it on c8 instead.
The move we both wanted for him afterwards: queen behind the pawn, and the c-file is shut forever.
The other one we looked at: centralise and offer the trade. Every queen swap is terrible news for my pawns.
The human way to kill the pawn: put the queen on e7, where it watches c7 along the seventh rank, then bring the bishop to f4 on the same errand. Two pieces covering one square from a distance, and c8 never happens.
Same idea, one move later, and easier to find than anything the engine plays elsewhere. The bishop is a spectator on g7 and a sniper on f4, where it stares straight down the diagonal at c7. Trade queens and the pawn is just lost.
Taking the pawn immediately was our main line in the post-mortem, and 34.Bc4 is indeed the wrong way for me: ...Qb4! and I'm losing. But see the other line. 34.Rxc7! saves the game, so 33...Ba4 (what he actually played) was the accurate move order.
The engine's correction to our post-mortem: 34.Rxc7! and after ...Qxc7 35.Rxd5 the position is dead level. So the immediate 33...Rxc7 throws the whole win away. His 33...Ba4 first was right, and we had it backwards at the board.
This is what he had to play, and we both saw it two seconds after he had let go of the bishop. He gives the exchange back, kills the c-pawn, and stays a clean piece up.
Why 35...Bf8 was forced: blocking with the queen runs into the promotion with check and mate next move. This is the one moment in the game where the second queen would really have appeared on the board.
The move that wins: the queen steps in front of the pawn and c8 is unstoppable. Let him have a2. 38.h4 makes luft and there is no perpetual.
The version I would want to find at the board: force everything off, promote with check and pick up the bishop on c2. Less spectacular than 37.Qd8, and it wins without me having to calculate anything.
What he resigned into. 37...Bd3!! is the whole point: my bishop is overloaded, and after 38.Bxd3 exd3 my new queen on c8 cannot stop the d-pawn. Black wins.
vs Zesewitz, Maik (1930)
1–0 · LossChess scores are always written White first. Elias had Black here.
Board fourteen, Black, against Maik Zesewitz, 1930. I got mated on move 36. The prep went almost exactly as expected: he opened 1.f4, I got the d5 and g6 setup I had studied, and for eight moves the game sat in a position I had already looked at. On move 10 I messed up the move order. I wanted to take on d4 first and I took the knight on e5 instead, so he got to recapture with the d-pawn, and I honestly also just underestimated his attack on my king. Three moves later I sacrificed a knight for activity. I thought I would get an insane bishop that also picks up some pawns, because his structure is weak and exploitable. I took two of them pretty forcefully with the queen and I thought I was better, but engine says no. Then the queen went shopping on b2 and c2 and she never came home. He put a bishop on e2, she ran to f5, and a rook and a bishop hunted her down there. I keep missing these tactics again and again, which is frustrating, and I feel rusty. He mated me. I always play until mate, that is just how I play. He was the nicest opponent of the tournament. I bought him a beer, we analysed the whole game together, and he told me he had prepared me off my 1.a4 game from last year, the joke one. He said my prep was impressive. The whole family was lovely.
Use the arrows to step through the game, or drag the evaluation graph to jump.
Line above the middle = his opponent is better. Below = Elias is better. · Dots mark mistakes.
The move order I actually wanted. Take on d4 first, make him recapture there, and only then trade on e5.
vs Juranovic, Igor (1693)
1–0 · WinChess scores are always written White first. Elias had White here.
Board twenty-five, White, against Igor Juranovic (1693). He resigned after my 27th move. He came in on the same score as me, with two draws against higher rated players this week. The prep was a system I had never played before, 1.d4 and 2.Nc3 and 3.Bf4, with one rule sitting at the top of it: if his pawn goes to d5, you build first and you leave e4 alone. His pawn went to d5 on move one. So on move four I left e4 alone even though it was probably the more precise move, and I got the slow game the report said I would get. Then I played f3 far too early. The engine says that one move cost a full pawn, his knight came to h5, my bishop got traded off and my h-pawn ended up on g3. That trade handed me an open h-file I had not planned on, and by move 15 the engine had me a pawn and a half up with the queen on h6. Then I pushed the g-pawn one move too soon and gave all of it back. On move 17 I took his knight on h5 with the rook, which is an exchange for nothing, and the engine puts me a pawn and a half worse after it. I did that on purpose. His defence had to be exact and mine did not, I had threats on both sides of his king, and I thought that was the way to play this particular game. One move later he proved me right. The engine gives him exactly one move out of 35 that holds his advantage there, he played a different one, and the pawn he let me push to f6 is the pawn every mating line at the end is built on. He had the same one-move choice three more times and found it once. On move 24 his queen went shopping on c2 and left the seventh rank, which was the only thing guarding g7. Two knight checks later he resigned into mate in two. Two out of four, and 32nd, which is exactly where I was seeded.
Use the arrows to step through the game, or drag the evaluation graph to jump.
Line above the middle = Elias is better. Below = his opponent is better. · Dots mark mistakes.
vs Calvi, Stefan (2015)
1–0 · LossChess scores are always written White first. Elias had Black here.
Board fifteen, Black, against Stefan Calvi (2015). He mated me on move 45 after three and a half hours. I beat him at this event last year and this was the first time he had White against me. The report had him right. He opened 1.e4 like he has in twenty-one games out of twenty-one, put the bishop on e3 and the queen on d2, and pushed the h-pawn at me on move 11, and all of that was in the document, so I played the first ten moves almost without thinking and after move ten his clock read 1:16 and mine read 1:32. Sixteen minutes up on a man rated 217 above me. Over my first twenty-three moves the engine has me giving away about a fifth of a pawn a move, and it does not judge a single one of them as a mistake. Then I made it sharp, because that is the game I wanted out of this pairing. On move 22 I pushed the d-pawn and blew the board up with about fifteen minutes each, and from there we both played four moves in under a minute apiece with the time control still fifteen moves away. The engine says the position was level after every one of them. On move 29 he took a rook on e5 and I answered with a queen move that does not take it back. I miscalculated and I was a rook down. Qxe5 was the move, it is the engine's first, and I found it myself in the post-mortem. Four moves later I traded queens because I thought I still had both rooks, and I only worked out where the second one had gone after the trade was on the board. He mated me on the h-file eleven moves after that. We went through the whole thing together afterwards and he loved the game. He said the preparation was accurate, he could not work out why I was still in prep in a structure he has never faced in his life, and he did not know you could prepare somebody's habits with openings they have never played. Still two points, and 40th.
Use the arrows to step through the game, or drag the evaluation graph to jump.
Line above the middle = his opponent is better. Below = Elias is better. · Dots mark mistakes.
Calvi's suggestion after the game: 11...h6 instead of 11...h5. The pawn still has to go to h5 a move later, and by then White has castled and the a-file is open. The engine has 11...h5 first of everything and 11...h6 sixth.
The move I thought about and discarded. 26.Bxe4 hits the bishop on b7 and the knight on h7 is still mine to take, so I cannot save both. Taking the knight first is the engine's move and keeps the game level.
What I worked out in the post-mortem. Queens and rooks come off, I keep the knight, and the ending is rook and knight each. It is the engine's first line move for move.
vs Brandstetter, Markus (2027)
1–0 · WinChess scores are always written White first. Elias had White here.
Board nineteen, White, against Markus Brandstetter (2027), the third seed of sixty-two. He resigned after my thirtieth move. I left early to buy yfood and read the report for an hour at the venue, and the traffic took the whole hour, so I walked into the hall about ten minutes late with twenty minutes left before the clock would have taken the game off me. He waited for me. What a nice guy. So the preparation was a skim on the morning of the game. What stuck was a bit of the Scotch and some plans against the Sicilian, and a4 before the bishop comes out. He answered 1.e4 with 1...c5, I put the f-pawn on f4 on move two and the a-pawn on a4 on move six, and the shape on the board was the one out of the document even though almost none of the moves were. On move 6 I looked down and saw that my queenside knight and my queenside bishop had been standing on each other's squares since before the first move. I put them back and we played on. The Laws of Chess have a different answer to that and I only read it afterwards. The engine has me worse from move nine to move eighteen, and not because of one move. I never took his pawn on d4 while it was hanging, I recaptured with the rook instead of the bishop on move 12, and on move 13 I gave a pawn to push his queen out to a5 and get my rook to h3. He defended the gambit properly and he kept the pawn for six moves. Then he played 18...Nb4 and 21...Rab8, the game turned twice in four moves, and after that I took the pawn back on b7, pushed f4-f5 into his king and gave a bishop on h6. My worst moment is move 20 and nothing on the board shows it. There was 20.Bd2, which I found at the board and threw away because of 20...Qb6. After 20...Qb6 there is 21.Rb1, it is the only move, and the engine puts White more than a pawn above the move I played. I calculated the whole tactic and stopped one move short of the answer. Three points from six rounds, 33rd, and a tournament performance of 1909 against my own 1798.
Use the arrows to step through the game, or drag the evaluation graph to jump.
Line above the middle = Elias is better. Below = his opponent is better. · Dots mark mistakes.
The move I found at the board and did not play. 20...Qb6 is exactly the reply I was afraid of and it is the engine's first, but 21.Rb1 answers it: the rook steps off the knight's fork with tempo on the queen, and Bxa8 wins the exchange anyway. 21.Rb1 is the only move here. Every other 21st move drops White from about +1.5 to worse than −2.3.
The tactic I actually calculated, and it works. If he takes the rook, Qh5 threatens mate on h7 while the bishop on d2 is still attacking the queen on a5, so he cannot answer both. He gets a rook, I get the queen.
vs Schmied, Hagen Matthias (1717)
0–1 · WinChess scores are always written White first. Elias had Black here.
Board sixteen, Black, against Hagen Matthias Schmied (1717) from Trofaiach. He resigned after my twenty-seventh move. Second win in a row, and the first game this week where I was the higher-rated player at the board. His opening moves came fast, so on move six I played d5. The move is probably not correct and I played it on purpose: it takes the game out of known lines into a structure I have played a lot, and from there we both had to think. On move ten I gave a knight on b4. It is a pattern I know from the Benoni, only with the bishop landing on f6 instead of the queen: the long diagonal opens and the rook in the corner cannot be held. Two moves later the bishop took it. The engine is honest about the sacrifice: 12.Bg5 refutes it, because the f6 bishop can be met with tempo on my queen before it cashes in. That move never came, and after that I had the exchange and targets. From move fifteen I did what round five taught me: no complications, trade everything, keep the pressure. I calculated 15...Nd4 at the board, found the refutation myself and threw the idea away, which is exactly the discipline that was missing against Calvi. He offered a draw on move 24. By then I had calculated every line to the end and blitzed the rest out: Ra3, Rd3, and on move 27 both rooks landed on his bishop. Every recapture loses more material to the back rank, so he resigned. We analysed together afterwards for a while, which is the best part of these open tournaments. 4 from 7, rank 24, performance 1931 against my own 1798.
Use the arrows to step through the game, or drag the evaluation graph to jump.
Line above the middle = his opponent is better. Below = Elias is better. · Dots mark mistakes.
The move I was afraid of. The knight gets sent back, c5 stays covered, and White keeps the space and the pawn. We looked at it together afterwards: c4 and then Be3 if White wants to hold everything, and Black has a long, unpleasant evening ahead.
The line from the joint post-game analysis: trade everything and fork with Ne5. The board has the answer ready. Bxe5 gives the piece straight back, and after Rxe5 Rd1+ it is mate on the back rank next move. The rook on a1 never gets to help. This back-rank motif runs through the whole rest of the game.
The crazy idea I calculated and discarded: Nd4 hits queen and knight, and if the queen takes mine, I take his. The refutation is the third capture, simply Nxd4, and I am down material with nothing to show. Finding this at the board and not playing the move is the whole lesson of round five.
The order I calculated and rejected out of draw fear. If White avoids the rook trade with Re1, this is the punishment: Rxe1+ Qxe1 Qxd4 wins the bishop, and the queen falls next to the mate threat plus the skewer from a1. The engine has this at -8.8. What I misjudged: even the sober Rxe8+ Qxe8 is -2.6 for White, not a draw.
The swindle we checked afterwards. Instead of 27.Rd2 the bishop drops back to d2 and hangs itself on purpose. Take it greedily and the dream line - Rxd2, Qxd2, Qxd2, expecting Rxd2 Re1 mate - gets cut off one move earlier: the rook does not recapture, it plays Rxe8 and mates ME on the back rank. I had this one calculated, because you check the last tricks before your opponent runs out of them.
vs Rukwid, Marc (1954)
1–0 · WinChess scores are always written White first. Elias had White here.
Board six, White, against Marc Rukwid (1954), the last round on Sunday morning. We both came in with 5 out of 8. I played the Wing Gambit, 1.e4 c5 2.b4, and that opening is the reason the preparation existed at all. He told me afterwards that the night before he had said to Kevin Narr that he bet I was going to play it, and that he did not prepare it anyway, because Narr told him it is a bad opening and he would work it out at the board. On move six I could not remember which pawn move was the prepared one, so I played c4 off the ideas I still had, and he took en passant. That capture is the moment the engine turns the game around: 0.41 for him before it, 0.90 for me after it, and he told me afterwards that c4 was the move that threw him off. He offered a draw on move ten and I declined after calculating the follow-ups to Nxd6. The engine has me 2.17 up there, so that part was easy, and two moves later I calculated my way into being a piece down. He took my knight on e5 with the queen, put a knight on e4 and left my rook hanging in the corner on a1, and for the next three moves he had a piece for nothing while I had two bishops and both open files. Qa4+ is the worst move I played by the engine count, 0.46 against the 2.4 of 14.d4, and it is also one of the two moves he told me he had missed. The other one was Bc3. Neither of them is a move the engine wanted and both of them are what actually beat him. The rest is the two bishops eating the queenside, a passed pawn on b7 that traded itself for his last piece, and a rook endgame a pawn up that I converted in fifteen moves. He resigned after 44.Rc5, and he said he had not seen the rook going to c6 with check and then sitting down on c5. 6 out of 9, rank 10, and a performance of 1985 against my own 1798. The estimator reads this game at 2130, no mistake and no blunder, my second best of the tournament.
Use the arrows to step through the game, or drag the evaluation graph to jump.
Line above the middle = Elias is better. Below = his opponent is better. · Dots mark mistakes.
The line I wanted afterwards. Qf3 first, and if he castles, Rxe4 and the queen on a1 costs him the game: Qxf7+ Rxf7 and Re8 is mate. He does not have to castle, and the engine defends with Bf5, but it still likes Qf3 about a pawn more than the check I played.
Bxc5 and then Re5, the idea I asked about afterwards. It wins both queenside pawns, and the engine reads it as the same value as the d4 I played. Both of them sit almost two pawns under 23.Rf3.
vs Hasanovic, Nurija (1924)
+-- · Win by forfeitA forfeit score is written White first too: the + side got the point without a move being played. Elias had White here, and there is no game to replay.
He did not show. The weekend rounds start at 16:00, two hours earlier than the weekday ones, and the arbiter told me he thought this one started at 17:00. So the point is mine without a single move: my first walkover of the tournament. The rating part, before anyone asks: an unplayed game does not count for rating, so the performance and the +24.4 still come from the seven rounds I actually played. There is nothing to analyse either, because there is no game. I will never know if it was really the mixed-up start time or my Morra prep waiting for him, and I know which story I am telling at the club. 5 out of 8, and the last round is Sunday morning.
Chess pieces by cburnett, licensed CC BY-SA 3.0. Standings and pairings via chess-results.com.