Predicting Premier League scores is easy to do badly. Pick the team you think will win, give them two goals and the other side none, repeat ten times, and you will finish mid-table in your predictions league every season. Finishing top takes something else: an understanding of how often each result and each scoreline actually happens, and the discipline to predict what is likely rather than what would be fun.
This guide is about that discipline. It covers the base rates that should anchor every prediction, the trade-off between exact scores and results, how to read a fixture list, and the habits that separate the players who win from the ones who blame the referee.
Over a typical Premier League season, roughly two in five matches are home wins, about a quarter are draws and the remainder are away wins. The average match produces just under three goals. A scoreline with five or more goals in it happens in fewer than one match in ten, and a team scoring four or more is rarer still. Those numbers move a little from season to season, but not enough to change the strategy: a prediction that assumes a rout is fighting the odds before a ball is kicked.
The most common final scores, season after season, are 1-0, 2-1, 1-1 and 2-0, followed by 0-1, 1-2 and 2-2. Between them those seven scorelines account for well over half of all results. 0-0 is less common than most people assume - fewer than one game in twelve - and 3-0 is more common than 4-0, 5-0 and 6-0 put together. If your predictions never look like a list of 1-0s, 2-1s and 1-1s, you are leaving points on the table every week.
With the standard 5 points for an exact score and 2 for the result, the question in every match is the same: which scoreline gives the best chance of the 5 without risking the 2? The answer is almost always the most common scoreline for the result you expect. If you think the home side wins, 2-1 or 1-0 is the pick; if you think it is a draw, 1-1; if you fancy the away side, 0-1 or 1-2. Reaching for 3-1 because it feels like the kind of game that finishes 3-1 costs you the exact-score chance without improving the result chance at all.
The corollary is to stop chasing five-pointers in matches you do not understand. A mid-table game between two sides in average form is a coin toss with three faces; predict the most common draw or the narrowest home win and accept that you are playing for the 2 points there. Save the confident scoreline calls for the fixtures where you genuinely know something: a strong home side against a team that has stopped scoring, a club playing three days after a European trip, a relegated side with nothing left.
The fixture list tells you more than the form table. Things to check before predicting a round:
A quarter of all matches are draws, so a player who never predicts one is wrong a quarter of the time by construction. The trick is knowing which quarter. Draws cluster in games between sides of similar strength, in derbies, in matches between two defensive teams, and late in the season between two clubs with nothing to gain. When you do predict a draw, 1-1 is the right scoreline far more often than 0-0 or 2-2: it is the most common draw by a distance, and a 1-1 prediction still collects the result points if the game finishes 2-2.
Once a match kicks off, every prediction for it is revealed, and that grid is the best coaching you will get. Look at what the players above you in the table predicted and how it differed from your call. The consistent winners tend to have fewer big scorelines, more 1-0s and 2-1s, and a higher proportion of draws than the rest. Look too at where the whole league agreed: if everyone predicted the same result and it went wrong, nobody lost ground, but if you were the only one to call the upset you gained on everybody at once. Being contrarian is expensive when you are wrong and valuable when you are right, so spend it on the fixtures where you have a real reason.
A predictions league is a marathon, and the table at the end reflects consistency more than brilliance. A few habits that add up:
The ways points are usually thrown away:
1-0, 2-1, 1-1 and 2-0 lead the list season after season, followed by 0-1, 1-2 and 2-2. Between them those seven scorelines cover well over half of all results, which is why a predictions league table is usually topped by the player whose predictions look like a list of them.
Roughly a quarter of matches in a typical season. A player who never predicts a draw is therefore wrong about a quarter of the time by construction; the skill is picking the games between evenly matched or cautious sides where a 1-1 is the likeliest outcome.
Both, in one prediction: choose the result you expect, then the most common scoreline for it (2-1 or 1-0 for a home win, 1-1 for a draw, 0-1 or 1-2 for an away win). That maximises the chance of the exact-score points without giving up the result points.
Rarely. Goalless draws happen in fewer than one match in twelve, and a 1-1 prediction collects the same result points whenever the game finishes level. Save 0-0 for two genuinely low-scoring sides in a game neither needs to win.
Move away from the crowd in the last few rounds: predict the upsets and the less common scorelines, because matching everyone else's predictions cannot close a gap. If you are leading, do the reverse and predict with the pack.