Home
Blog
Football predictions done yourself: how to read a match before betting

Football predictions done yourself: how to read a match before betting

Football predictions done yourself: how to read a match before betting
Author
Gustavs Ērkšķis-Kreicbergs
Updated
17.11.2026

A football prediction you make yourself is a probability, not a guess at the winner. If the home side creates chances worth 1.6 goals on average and the away side 1.0, a simple model puts the home win at 51.3%. Bookmaker odds of 2.05 imply 48.8%, so on paper you have +5.3%. But if the home figure is really 1.4, the home win drops to 46.1% and the same bet becomes -5.4%. This piece is about how to get to that number and how precise it has to be. There are no tips or picks here.

Form: results against chances

Results over the last five to ten games show what happened; expected goals (xG) show how many chances a team made and how many it allowed.

In the explanation by Opta, the data company, xG measures the quality of a chance as the likelihood that a shot is scored, on a scale from zero to one. The model takes in the distance and angle to goal, the goalkeeper's position, defensive pressure, the type of shot and the pass that led to it. A team's xG for a game is the sum of all its shots.

Why does it matter? A team can win four of five games with a negative xG difference. Then the wins came from rare chances that went in and opponents' chances that didn't. On the table it looks like form; it rarely lasts.

Why five to ten games? Over five games one lucky strike from 30 metres moves the picture far more than over ten. Go past ten and the sample often reaches back across injuries, a new coach or a transfer window, and you're rating a different team. We usually read both numbers side by side: results to understand the table, xG created and allowed to understand the play.

Line-ups and injuries: 75 minutes before kick-off

In the Nations League each team must submit its line-up at least 75 minutes before the game, and until then every injury report is only a probability.

Article 52.04 of the UEFA Nations League regulations requires each team to submit its match sheet at least 75 minutes before kick-off, and the first 11 named on it start the game. A prediction for today's match is only finished once the line-ups are out.

How do you turn a missing player into a number? We look at what share of the team's xG he created himself or set up. If the side averages 1.6 with its main striker and made around 1.4 in recent games without him, the model gets 1.4. The opening example shows what those 0.2 do to the value of a bet.

Fixture load and travel

Two games in three days with a flight in between is a normal November for national teams, and it shows in the line-ups.

UEFA's calendar had Latvia scheduled for two games: away to Montenegro on 12 November and at home to Armenia on 15 November. Three days, a trip across Europe and a coach deciding whom to rest. Clubs face the same in weeks with a European game on Wednesday and a league game on Saturday.

We write down the rest days for both teams. We have no exact formula for how many goals a day of rest is worth, and we don't trust anyone who sells one. Fixture load mainly changes the line-up, so we count it through the line-up rather than as a separate percentage.

Motivation at the end of a group

The Nations League league phase ends on 17 November, and in the last rounds one team is playing for promotion while the result changes nothing for another.

In UEFA's update of 12.11.2026, League C group winners go up to League B, runners-up go to the play-offs, and third and fourth stay in League C. Latvia had 1 point after four rounds in Group C2, and UEFA wrote: "Latvia will remain in League C." In their last two games, Latvia's league place was no longer at stake.

What does that do to a prediction? First, a coach can use such games to give young players a run, so the line-up matters even more than usual. Second, the bookmaker can read the table too, and the obvious motivation is already in the odds. You only gain if you read motivation better than the market, for example by seeing that a team that only needs a draw has no reason to attack.

From numbers to probability

The simplest way to get 1X2 probabilities from xG is the Poisson distribution: assume each team scores independently, at its own average.

In our example the home average is 1.6 goals and the away average 1.0. The Poisson formula gives the probability of each number of goals. For the home side, 0 goals is 20.2%, 1 goal 32.3%, 2 goals 25.8%. For the away side, 0 and 1 goal are 36.8% each. Multiply the two teams' probabilities for every scoreline and add up the home wins, the draws and the away wins, and you get:

  • home win: 51.3% (fair odds 1.95);
  • draw: 25.0% (4.00);
  • away win: 23.7% (4.23).

In a spreadsheet the POISSON.DIST function does this in ten minutes. If the model feels too heavy, a simpler route works too: write down your own percentages for the three outcomes so they add up to 100. What matters is that your view becomes a number you can compare.

Checking it against the odds

A bet has value only if your probability multiplied by the odds is more than one.

Say a bookmaker offers 2.05 / 3.50 / 3.70 on this game. These are not any bookmaker's prices, just an example. The implied probabilities are 48.8%, 28.6% and 27.0%, a total of 104.4%. The 4.4% above a hundred is the bookmaker's commission, the same logic as the house edge we covered in casino game odds.

  • Home win: 0.513 × 2.05 = 1.053, so +5.3%, or EUR 0.53 on average for every EUR 10.
  • Draw: 0.250 × 3.50 = 0.875, so -12.5%.
  • Away win: 0.237 × 3.70 = 0.875, so -12.5%.

Now the same sum with slightly different inputs:

  • home 1.6, away 1.0: home win 51.3%, a bet at 2.05 gives +5.3%;
  • home 1.4, away 1.0: 46.1%, -5.4%;
  • home 1.6, away 1.2: 46.7%, -4.4%;
  • home 1.4, away 1.2: 41.6%, -14.8%.

An error of 0.2 goals in one input is bigger than the whole edge. So we expect a range from a prediction rather than one precise number, and only bet when the odds are good across that whole range. How much to stake once you've found an edge is in our piece on value bets and the Kelly criterion.

Where this method goes wrong

The Poisson model doesn't see red cards, the weather, or that a team in front plays differently.

The model assumes the two teams' goals are independent and the averages hold for the whole game. Football isn't like that: a team leading 1-0 often sits back, and the team behind takes risks. So for us the model's numbers are a starting point, adjusted for line-ups and the table.

If betting on football is new to you, start with our sports betting basics on bankroll and odds. The football markets and sportsbooks available in Latvia are in our football betting section.

Play within your means

No prediction is a reason to stake more than you planned. Set a budget before you start and stop when it is spent. Our responsible gambling page lists tools that help, and you can add yourself to the State Revenue Service self-exclusion register at any time. The Centre for Disease Prevention and Control addiction support line, 67037333, is anonymous. Betting is for people aged 18 and over only.

Gustavs Ērkšķis-Kreicbergs, SpinsPilot author
iGaming industry expert
Gustavs is the former COO of the casino "Spēlēt.lv" with more than 6 years of experience in iGaming marketing. He writes about industry news, its history and how casino offers are actually built.
Read more

Best casinos to play at

Frequently asked questions

How do I make my own football prediction?
What is xG in football?
When are football line-ups announced?
How do I turn odds into a probability?
Can a football prediction be accurate?

Become a pro

Read more →
Poker

Master the most popular card game around, where skill counts for as much as luck. Learn the hand rankings, basic strategy and how to read your opponents so you can sit down at the table with confidence.

Roulette

This guide has everything you need to know about improving your odds at roulette while enjoying this thrilling casino classic.

Blackjack

Master the casino's most popular card game, where skill meets luck. Learn basic strategy, card-counting techniques and how to maximise your odds against the dealer in a classic table game that has entertained players for centuries.

Baccarat

Baccarat is one of Europe's oldest card games. Today it ranks alongside poker and blackjack as a casino classic. Players around the world love baccarat - it is easy to learn and the rounds are fast and exciting.

wise
Wise
usdt
Tether
usdc
USD Coin
trustly
Trustly
swedbank
Swedbank
solana
Solana
sepa
SEPA
seb
SEB
paysera
Paysera
paypal
PayPal
neteller
Neteller
muchbetter
MuchBetter
mifinity
MiFinity
luminor
Luminor
litecoin
Litecoin
indexo
Indexo
ethereum
Ethereum
dogecoin
Dogecoin
citadele
Citadele
cardano
Cardano
bitcoin
Bitcoin
apple-pay
Apple Pay
crypto
Crypto
google-play
Google Pay
master-card
Mastercard
payafe
Paysafecard
revolut
Revolut
skrill
Skrill
visa
Visa