What is the best move in chess?

Introduction

The question "what is the best move in this position" is something that is quite widely asked and it is something that is crucial to every chess player. After all, it is a question that must be asked on pretty much every position, as we would not like to play moves that are subpar. The anwser to such question seems simple, though. Don't we have access to a godlike silicon entity, which indicates the best move? How could the best move be anything else besides the move indicated by Stockfish (for example) at high depths? This is something that I thought hard about, and I would like to introduce you to the "Four Variable Rule", which says that no, the engine's top move is not, necessarily, the best.

What is a best move in chess?

During a chess game, when we ask the question "what is the best move in this position?", we are actually asking "what is the move that I should play in this position?". These are two different questions, even though it might not seem so at first glance. The objectively best move will be, indeed, the engine's top move at very high depths. However, we do not care for the objectively best move at all. What we want to is to eventually win the game, as surely as possible, and bring home the full point without any hiccups. This means choosing the move that is best for those means, instead of the fastest win, the only thing a chess engine considers.

This means that we should make the distinction between the absolute best move in a position, and the move that gives the best practical chances, in the same position.

The Four Variable Rule

The Four Variable Rule, as I like to call it, is quite simple, in fact. It states that the objective evaluation of the move is only one of Four main considerations (variables) that a player must take into account, when considering a chess move. The four variables are the following:

1. What is the objective value of this move?
2. How much time on the clock will I need to spend in order to calculate the required depth of this move?
3. How likely is it that I will mess up the resulting position? / How likely is it that I missed something in my calculations?
4. How straightforward is the plan for the opponent after my move?

Lets have a look at each of them. Question One is simple. It is simply how good we think the move actually is, objectively (you can think about this as the closest as possible to Stockfish's top suggestion).

Question 2 is self-explanatory as well. If we have two moves, one that takes the queen for free in one move, or one that might give mate in 7 tricky moves, which one will take up more of your time to solve? This should also be taken into consideration to pick the best move.

Question 3 is split into two questions, which are interconnected. They make you evaluate how sure are you of your line, and if the risk involved is worthy. Going back to the previous example, if you can capture a queen in one without any repercussions, or you can calculate a hard 7 move line, with many unforced moves and continuations, to give checkmate... Which one should you choose? This third question, along with the others argues that it should definitely be the queen, even though Stockfish might scream at you.

Finally, Question 4 is Question 3, in reverse. Sometimes, it is quite important to understand how easy will it be for the opponent to play the best moves, after our move. However, we should use this with care, and this is mostly useful in positions that are going south already. In general, we should not play moves to which we see a refutation, just because we think the opponent will not see them. In special scenarios though, where we are lost, or the "best" move is obviously a draw and we need to win, going for small tricks can be the way to go.

Examples

1.FM Santos, José - IM Sousa, André | Portuguese Championship 2022

The first example depicts Question 4 in action. It was played between me, as Black, and José Santos, a strong FM from Portugal, in the last round of the 2022 Portuguese Championship. For added context, keep in mind that if I got the same result as another IM playing the Championship, I would win my fourth title, so a win would guarantee the first place.

What was the best move for White on move 20 after all? I strongly argue that it was 20.Na4!!, as I seriously suspect that without it, the game would be a clean win for Black.

2. GM Fier, Alexandr - IM Sousa, André | Menorca Open 2025

The second example was played between a Brazilian GM and myself, during a recent tournament in Menorca. The interesting position arises at a moment when I am under some pressure.

So, what was the best way for me to neutralize the pressure? I strongly think that it was my move, and not 20...f6. The objective quality of the moves were the same - both eventually "0.00s" - but while in one I can just play simple moves quickly, in the other I have to calculate complicated variations with pitfalls, which fall directly into Question 2 and 3.

3.GM Shankland, Sam - IM Sousa, André | Menorca Open 2025

The third example is also from Menorca. It is a bit different from the previous ones. In this one, the best move is indeed the engine's best move, but my move is also evaluated as -2. Is my move any bad? I think so. See for yourself.

In the end, the objective evaluation of both moves was the same - winning for Black. However, how much time did I need to make 36...Ne5 work? Not that much, actually. What were the chances that I would mess up if I kept the queens on? Well, since White's king was so much more exposed, definitely lower than after 36...Qf6.

Conclusion

We have discussed the Four Variable Rule and seen some examples of it in action. You may wonder though - what is the usefulness of this in a game? We do not have access to Stockfish during a game (unless you're cheating) so you cannot possibly know the anwser to Question 1 with precision. However, during a game, we can feel if a decision requires a lot of calculation - in Fier, A - Sousa, A, it is clear during the game that if Black does not immediately take the bishop, there must be a concrete "now" way of solving the issues, as otherwise there is no compensation. Other times, it is clear that the "best" move will lead to certain defeat - we must risk something more creative.

In the end though, I think that the best use for this is not during the game, but during the analysis. There is a lot of harm done in the generalized obsession of going over a match with an engine immediately after its finish, both from a learning perspective, but also psychologically. I've seen too many times my students feel "dumb" or "idiots" because they missed "simple things" that the engine spots immediately. This is a whole other topic I might write about in the future, but the Four Variable Rule can be used to use in this post-game analysis frenzy for something more than self-doubt and self-chastisement. It can be used to actually compare what the engine is saying to what we played, and perhaps reflect on whether our move is actually that bad. Sometimes (most of the times, I'd say!) it will be, but there is merit in a good practical decision taken during a real game, and that is easy to forget nowadays.