When should you commit to a colour in draft?

Practice · draft strategy

Draw a note and draft by it — three boosters, forty-two picks, at a table of eight. The note names the method for choosing each of your two colours, so at every pick the question is not which card you want but which card the note takes, and at the end you see the deck the method built.

The two methods, and what a note is for

The letter on each card. A is the top quarter of its colour — the bombs that win a game on their own sit at the top of it, though not every A is one. B is a solid playable, C is filler, and D is a card that does not belong in the deck at all, not even as filler: too dear to cast in a limited game, or waiting on conditions that rarely come, or needing a combination you cannot build in forty cards. A nonbasic land carries a letter too, by the same test; a basic land is not graded. The note picks by quality, and this letter is the quality it means, not the rarity, which only says where the card sits in the pack.
Early. Steering the table. You commit to a colour at the start and keep taking it, its good cards and its ordinary ones, giving up better cards in other colours, so that the players on your left never see enough of it to move in. You gain synergy from the first pick and pay for it with flexibility.
Late. Reading the table. You commit to nothing and watch what comes round: what the players on your right are not taking in picks two to four, or what the whole table has left alone by picks five to eight. You gain raw card quality and pay for it with your first pick and with early synergy.
The three notes. Early/Early is a fixed plan and a gamble on your own reading of the set. Early/Late is the usual one, and the safe one, because each half covers the other’s weakness. Late/Late is the other gamble: the best cards on the table, if the table is uneven enough to give them to you.
Three boosters, and the middle one goes the other way. The first pack passes to your left, the second to your right, the third to your left again. That is the real rule of a draft, and it decides who is upstream of you: in the first and third packs the player on your right feeds you, in the second it is the player on your left. A colour that dried up can come back when the direction changes — which is exactly when a late note has something new to read.
And then the deck. Forty cards is twenty-three playables and seventeen lands, so a draft that comes home with twenty-three cards in its colours has built one. The summary says whether yours did, which is the only honest test of the note you were given.
And then it plays. At the end, the deck the note built plays the decks the three players beside you came home with — three hundred games against each, both sides opening on seven cards and twenty life — and you are told the share of games it wins and where that puts you among the four. The note is a method, not a prediction, and this is the only way to see whether the method worked. The same measurement over a hundred and fifty tables says what each note averages, so your table can be set against the method rather than against nothing.
Every note is defensible. What is not defensible is holding one and then picking as though you held another.