I was born in Canberra in 1956. My mother was a great one for books, and when I was about nine years old she borrowed a beginner’s guide to chess. I eagerly devoured the book (after reading it first) and later taught my younger brother to play. We would also play chess with the kids next door. Soon afterwards, my mother found a book called Mate in Two Moves at a fete, and she bought it for the princely sum of two pence. The book, written by Brian Harley, was my introduction to the world of chess problems, and it is still one of my most treasured possessions. Occasionally I would attempt to solve the weekly chess problem in The SunHerald, but it would always take me hours.
Chess problems became an everpresent interest. I attempted to compose my own, but with little success. Eventually I began to improve, but it was not until my late twenties that I produced anything worthwhile. I was strongly influenced by the Harley book, so I concentrated on twomove chess problems of the type that were popular in the 1930s, which is when the book was written.
1. Geoff Foster British Chess Magazine 1987 

Mate in 2 
In Problem 1, any move by the black queen will relinquish control of one of the squares b7 and e4, allowing the knight to mate. No other black piece can move, so a waiting move will solve the problem (this type of problem, in which every black move is provided with a mate, is known as a complete block). A promising try is 1.Rg5?, but Black can play 1…Qh1!, retaining control of both b7 and e4. The other thematic try is 1.Ra5?, but this opens the bfile, so that 1…Qb1! now guards b7. The key is the surprise retreat 1.Rh4! (waiting), giving the black king a flight square on e5. If 1…Kxe5 then 2.Sb7 (not 2.Se4, which would allow the king to escape to d4). This problem was very difficult to compose. One of the main challenges was in finding a safe square for the white king, but fortunately he can be placed on a6, although this also requires a black pawn on a7. The white pawn guards d5, but it has another use – after 1.Ra5?, it prevents 1…Qd3 from being a check, which would be a second refutation of the try and so ruin the problem. The bishop guards c7 and e7, and also guards f6 after 1…Kxe5. The rook on b5 guards c5 and also guards e5 after the key move, giving a battery mate after 1…Kxe5.
2. Geoff Foster The Problemist 1985 1st Hon. Mention 

Mate in 2 
Problem 2 is another complete block. A move of the black queen is met by 2.Sb5 or 2.Qxd2; a move of the knight on a5 will lead to 2.Qc4; a move of the other knight is answered by 2.Qc3; while the pawn moves interfere with the queen’s guard of b5, being met by 2.Sb5. There is no waiting move available, so White plays 1.Se5!, threatening 2.Qc5 (a complete block in which the key makes a threat is known as a blockthreat). The key grants a flight to the black king, so disrupting most of the set mates. The key is also a threefold sacrifice of the knight. If the pawn or queen captures the knight then this brings back the set mates 2.Sb5 and 2.Qxd2 respectively, while 1…Kxe5 is met by the threat 2.Qc5. Black can also defend with the knight on a5 (e.g. 1…Sb7), but this is answered with the changed mate 2.Sc6 (2.Qc4 no longer works, because of the black king’s flight square on e5). The other knight can also defend against the threat (e.g. 1…Sxb4), which is answered by the set mate 2.Qc3.
My main field of interest was twomovers, but in the 1970s I became increasingly fascinated with seriesmovers (a type of problem in which one side plays a series of moves). Seriesmovers were very popular with Australian composers, and I used to marvel at the ones by Brian Tomson in Chess in Australia. My first seriesmovers were serieshelpstalemates (Black plays a series of moves to reach a position in which White can stalemate in one move. Black may not inflict check, except possibly on the last move of the series. Black may not move into check during the series. The order of Black’s moves must be strictly forced, otherwise the problem is unsound.) The following problem was my most ambitious and successful serieshelpstalemate.
3. Geoff Foster The Problemist 1986 

Serieshelpstalemate in 68 
In the above problem, if Black were to play 1.h3 then White could reply with Sxf3, which would stalemate Black were it not for the white pawn on b6. Therefore Black must capture this white pawn, but it must be done without giving check to the white king (which is not allowed in a seriesmover) or disrupting the intended stalemate. The only piece which can do this is the black king, which must move through the “cage” of pieces to escape via h3, after which it can capture the pawn and make the return journey to f1. The solution is: 1.Qh3 2.Bh2 3.g1(Q) 4.Qgg2 5.Bg1 6.Rh2 7.Qh1 8.Q3g2 9.Rh3 10.Bh2 11.Kg1 12.Qf1 13.Qhg2 14.Kh1 15.Bg1 16.Rh2 17.Qh3 18.Rg2 19.Bh2 20.Rg1 21.Qfg2 22.Rf1 23.Bg1 24.Kh2 25.Qh1 26.Q3g2 27.Kh3 36.Kxb6 45.Kh3 46.Kh2 47.Qh3 48.Q1g2 49.Kh1 50.Bh2 51.Rg1 52.Qf1 53.Rg2 54.Bg1 55.Rh2 56.Qhg2 57.Rh3 58.Bh2 59.Kg1 60.Qh1 61.Qfg2 62.Kf1 63.Bg1 64.Rh2 65.Qh3 66.Rg2 67.Q3h2 68.h3 Sxf3=.
A few years after the above problems were published I had a career change, becoming a computer programmer. I was paid to sit in a cosy chair in a warm office, working on problems of a different type. This suited my temperament well, because I could tune out from the world, only having to communicate with the computer. Chess problems didn’t seem very important, because I gained great satisfaction from my job.
After a long period during which I had little interest in chess problems, Ian Shanahan sent me a new twomover he was working on. In his problem the key move created a battery, and I realized that a similar scheme could be used to show changed play. My first settings were simple affairs, but gradually I found that an amazing amount of richness could be built into the matrix. I sent the finished problem to The Problemist, where it won a welcome prize. Then later it received the prestigious Brian Harley Award, for the best twomove composition by a Commonwealth composer published in Britain during the years 200102. This meant a great deal to me, because it was Brian Harley’s book that had introduced me to chess problems all those years ago. Here is the problem.
4. Geoff Foster The Problemist 2001 4th Prize [Brian Harley Award 200102] 

Mate in 2 
The try 1.Sf7? threatens 2.Qf6. Black has several defences which prevent this threat, but most of these unguard a square next to the black king, allowing mate by the white queen: 1…Sxh3,Sxd3,Sxg4 2.Qxe4; 1…Rd5 2.Qxd5; 1…Qxf4 2.Qxc3; 1…bxa4 2.Qc4. However, there is no mate after the defence 1…Rh6! A similar try is 1.Sg6?, after which Black can defend with 1…Rf8! The key move is 1.Qf6!, which forms a battery aimed at the black king. There are four threats which are separately forced: 1…Rf8 2.Sf7 and 1…Rh6 2.Sg6 bring back the white tries as mating moves, 1…Sxh3 removes the white guard of e3 forcing 2.Sc4 and 1…bxa4 forces 2.Sxd7 in order to guard c5. There are other defences which prevent all four threats. 1…Rd5 selfblocks d5, allowing the doublecheck 2.Sc6. In addition, 1…Qxf4 opens the bishop’s line to e3, allowing 2.Sf3, another doublecheck. There are also a pair of recapture mates: 1…Sxd3 2.Sxd3 and 1…Sxg4 2.Sxg4. The white knight thus travels to each of the eight possible squares and there are six changed mates. Other play is 1…Rb7,Re7,Rd6+ 2.Q(x)d6 and 1…Bb8 2.Qb6.
While composing this problem I would send each new version to Ian Shanahan, who would provide advice and encouragement. It was Ian who had the idea of placing the white rook on h3, so that 1…Sxh3 would force 2.Sc4. I offered to make Ian the problem’s coauthor, but he generously declined. This is just as well, because joint compositions are not eligible for the Brian Harley Award. Thanks Ian!
5. Geoff Foster (after C. Goldschmeding) The Problemist 2003 

Mate in 2 
The above problem is another complex affair. The black king has two flights, both of which have set mates: 1…Kf4 2.Bxg3 and 1…Kd4 2.Bxc3. Any move by the white knight will leave black in check from the white rook on e2, but will also give the black king two more flights, making four in total (four diagonal flights are known as star flights). The knight has five promising moves, but only one of these provides a mate for each of the flights. The other four knight moves are refuted by different flights, with changed play throughout. Checking first moves are usually frowned upon, but in this case there is compensation in the granting of two flights and the amazing amount of changed play. Try 1.Sxc5+? 1…Kf4 2.Sd3; 1…Kd4 2.Sb3; 1…Kd6 2.Sxb7; 1…Kf6! Try 1.Sxc3+? 1…Kf4 2.Bd2; 1…Kf6 2.Sxd5; 1…Kd6 2.Sb5; 1…Kd4! Try 1.Sxg3+? 1…Kd4 2.Bf2; 1…Kf6 2.Sh5; 1…Kd6 2.Sf5; 1…Kf4! Try 1.Sf2+? 1…Kd4 2.Qd3; 1…Kf6 2.Sg4; 1…Kd6! Key 1.Sxg5+! 1…Kf4 2.Sh3; 1…Kd4 2.Rh4; 1…Kf6 2.Sh7; 1…Kd6 2.Sxf7. There are 17 different mates after moves of the black king. After 1.Sf2+? Kf4, there is a choice of two mates: 2.Sd3 and 2.Sh3. Such a choice of mates is known as a dual and is a fault (especially as the mates occur elsewhere in the play), but the scheme was bound to break down somewhere! The key 1.Sxg5+! leads to the surprise mate 1…Kd4 2.Rh4, which is especially welcome because it gives that rook a function in the postkey play. The mate is hard to spot; in fact it was found by a computer while I was testing an earlier version of the problem!
6. Geoff Foster Koldijk Memorial Tourney 2007 1st Prize 

Mate in 2 
In the above problem, a move by the black rook would open the white bishop’s line to d6 and e5. This would seem to allow three doublepin mates (2.Rc5, 2.Rd6 and 2.Sb6), but each of the rook’s defences unpins a different black piece, so preventing two of the mates. A choice of three mates is known as a triple, and the problem shows a theme known as triple avoidance. 1.Qb3! (threat 2.Qxb5) 1…Rxb3 2.Rc5; 1…Rd3 2.Rd6; 1…Rf3,Rxg2 2.Sb6; 1…b4 2.Qxc4.
All through this period my job was very important to me. In fact, I thought it was the only thing keeping me sane. However, in 2008 I decided that my chess problem activities didn’t leave time for a job, so something would have to go! I decided to retire from work, and surprisingly I haven’t missed it at all. Now chess problems are the only thing keeping me sane. To finish, the following problem holds the task record for the most number of interference unpins in a seriesmover.
7. Geoff Foster StrateGems 2009 

Serieshelpstalemate in 20 
1.Rd6 2.b5 3.b4 4.b3 5.Sb6 6.Rd5 7.b2 8.bxa1(B) 9.Be5 10.Sg4 11.Be4 12.Bd6 13.Sc4 14.Rf5 15.Sge5 16.Bc6 17.Be7 18.Rd5 19.Sd6 20.Bxe8, fxe8(Q). The final position is a quintuplepin stalemate!