04 Sep '21 17:26>
I'm reading the book How Purdy Won. In Purdy's correspondence game from 1936-7 against G. F. McIntosh (Black) that begins on page 24, the following position was reached after 13...N(a6)c5:
Purdy writes, "And now if 14. exf5 Bxf4 15. Bxf4 Nd3+, and regains the piece with a winning attack." (Purdy doesn't specify which knight would check at move 15.)
However, after 16. Bxd3 Nxd3+ 17. Kf1 Nxf4, White has 18. Qe4+, winning a piece and leaving Black with insufficient compensation. 17...Qa6 (instead of 17...Nxf4) can be met by the cold-blooded 18. Be3! Nxb2+ 19. Kg1, and White is well on the way to consolidating.
Perhaps Black intended to meet 14. exf5 with 14...Nb3!?, which was brought to my attention by the silicon analyst. One point is that after 15. axb3 Qxa1 16. Qxa1 Nc2+, White seems to have nothing better than blocking the f1-bishop by 17. Ke2 (17. Kf2 Bc5+ 18. Kg3 Nxa1 would threaten 19...Nxb3 or 19...Rxd2; 17. Kd1 Nxa1 would threaten 18...Nxb3 or 18...Bxf4). So White should play 15. Qe4+, clearing the rank for 16. Rd1. But 15...Qe5! threatens to win the f4-knight after eliminating both its protectors (by exchanging at d2 and then at e4), leaving White no time for 16. Rd1. So, White would return an exchange by 16. axb3 Nc2+ 17. Kf2 Nxa1 (17...Bc5+ loses even more material after 18. Ke2) 18. Bc4, emerging with two pieces for a rook after Black plays ...Qxe4 and ...Nc2. (See next diagram for the sequence in bold.)
White actually played 14. Nfd5, after which Black finished nicely. Here is the complete game.
Purdy writes, "Black sent the following conditionals: If 19. exd5 Ncd3+ 20. Bxd3 Nxd3+ 21. Ke2 Bb5 22. Qg1 Nf2+ 23. Ke1 O-O, forcing mate or win of the Queen." (See next diagram for this variation.)
Purdy writes, "And now if 14. exf5 Bxf4 15. Bxf4 Nd3+, and regains the piece with a winning attack." (Purdy doesn't specify which knight would check at move 15.)
However, after 16. Bxd3 Nxd3+ 17. Kf1 Nxf4, White has 18. Qe4+, winning a piece and leaving Black with insufficient compensation. 17...Qa6 (instead of 17...Nxf4) can be met by the cold-blooded 18. Be3! Nxb2+ 19. Kg1, and White is well on the way to consolidating.
Perhaps Black intended to meet 14. exf5 with 14...Nb3!?, which was brought to my attention by the silicon analyst. One point is that after 15. axb3 Qxa1 16. Qxa1 Nc2+, White seems to have nothing better than blocking the f1-bishop by 17. Ke2 (17. Kf2 Bc5+ 18. Kg3 Nxa1 would threaten 19...Nxb3 or 19...Rxd2; 17. Kd1 Nxa1 would threaten 18...Nxb3 or 18...Bxf4). So White should play 15. Qe4+, clearing the rank for 16. Rd1. But 15...Qe5! threatens to win the f4-knight after eliminating both its protectors (by exchanging at d2 and then at e4), leaving White no time for 16. Rd1. So, White would return an exchange by 16. axb3 Nc2+ 17. Kf2 Nxa1 (17...Bc5+ loses even more material after 18. Ke2) 18. Bc4, emerging with two pieces for a rook after Black plays ...Qxe4 and ...Nc2. (See next diagram for the sequence in bold.)
White actually played 14. Nfd5, after which Black finished nicely. Here is the complete game.
Purdy writes, "Black sent the following conditionals: If 19. exd5 Ncd3+ 20. Bxd3 Nxd3+ 21. Ke2 Bb5 22. Qg1 Nf2+ 23. Ke1 O-O, forcing mate or win of the Queen." (See next diagram for this variation.)