We have a raging king hunt going on, so the first order of business has to be to bring more troops to the attack. The Ra2 is doing a great job where he is, so the most promising move is 29...R8a3+. Adding another attacker, driving the black king up the board and keeping the rooks connected.

Okay, admittedly it leaves the Nb4 en prise, but surely that's not an issue when we have two rooks in the hunt.

After 29...R8a3, the white king can run to b4 or d4:

30. Kxb4 Rb3#
30. Kd4 Nc6# or Nc2#

It's where luck plays a part in chess. If 29...Rc2+ appealed to you first you might spend a long time trying to make it work. Fritzie says that it does win, but it takes a while.

On the other hand, life is a lot easier if your eye lit on 29...R8a3+. Then you don't even glance at the complications of 29...Rc2+

That's where I take issue with Kotov. He says that we should pick the most promising line and then examine it completely. Only then should we look at our second most promising line.

I am not so sure. In real life, I suspect that most of us flit from variation to variation. We might start analysing 29...Rc2+, but then realise that it isn't the quick forced mate that the position seems to call for. That might prompt us to put 29...Rc2+ down and look for alternatives. Only if those alternatives fail would we come back to 29...Rc2+ and see if we can make it work.

Initially, I tried to solve it from the White side and kept wanting to move the Rook closest to me (i.e. 29...Rc2+).

However, even the winner of this game had the same problem with passing up a simple win for a more complicated one.

As <coker> notes <Black missed a mate in 2 beginning with 28 ... c3+.> Fritz indicates that with the game move, 28...Ra2+, White can force Black to stretch it out to mate-in-eight (e.g. 28...Ra2+ 29. Kc1 Rc2+ 30. Kb1 c3 31. Rh2 Rxh2 32. Bg2 Rxg2 33. Bf2 Rxf2 34. h5 Rb2+ 35. Kc1 Ra1#).

R8a3+ 2.Kd4 Nc2#

29. R8a3+

Checks are paramount for investigation, because the opponent often has limited or only 1 response. The same could be said for attacks on the Q or loose pieces or pawns, but the defense might have much more leeway in these cases.

This often leads me to investigate positions which lead up to forcing moves that don't work at the moment, but will if my dear opponent cooperates.

Some of my best combinations -- and also Alekhine's, Fischer's, etc. -- come from the most harmless looking positions, but with a sting (forcing sequence) at the end, such as a double/family check even with all the attackers en prise.

Sometimes all you have to do is let your opponent go wrong without harming your position setting up the "sting at the end".