The words of betting and of KickValue, in plain words.
1X2
The result market: 1 = home win, X = draw, 2 = away win.
Accumulator (combo)
Several bets on one ticket. All must win; the odds multiply.
Bankroll
The money set aside for betting, that you can afford to lose.
BTTS
Both teams to score. Yes if each team scores at least once.
Calibration
When the model says 60%, does it happen about 60% of the time? A chart on the Record page.
Chasing
Betting bigger to win back a loss. The fastest way to lose a bankroll.
Closing line / CLV
The last odds we saw before kick-off (the middle price of many bookmakers, one per brand). For every match with a model pick our robot reads the prices once more about 10 minutes before kick-off (since 23 Oct 2026; until then our last read was on average about 4 hours before), so it is close to the last price, not the price at the whistle. CLV = how much better our logged odds were than that. Beating it on average is the earliest sign of an edge.
Correct score
A bet on the exact final score, e.g. 2–1.
Double chance
Two of the three results in one bet (1X, X2, 12).
Draw no bet
Back one team; a draw returns your stake.
Drawdown
How far the bankroll fell from its highest point. Every strategy has some.
Edge
Model chance minus bookmakers' chance (margin removed), in points. A pick means the price looks too high to the model, not that the outcome will happen. Our threshold for a pick is 3 pts. On 5 seasons of past odds, a bigger edge meant a bigger loss: the market was right more often. From 10 points the pick carries far from the market.
EV (expected value)
Average profit per unit if you made this bet many times. +0.05 = 5 cents per unit.
Fair odds
1 ÷ model chance. The price at which the bet is neither good nor bad.
Handicap
An imaginary head start given to one team (+1.5) or taken from it (-1.5) before reading the score.
Hit rate
Share of settled picks that won. Meaningless without the odds.
Kelly
A formula for the stake that grows a bankroll fastest, if the chance is right. The site suggests no stake since 21 Sept 2026: the backtest says the model chance is not.
Line
The number in a totals bet: 2.5 goals, 9.5 corners, 3.5 cards.
Logged pick ()
A pick written to the database before kick-off, with its odds at that moment: the middle price of many bookmakers (one per brand), not a fair price. It can never be edited.
Margin (overround)
The bookmaker's cut: the chances of all outcomes add to more than 100%.
Market chance
The bookmakers' chance for this outcome: from the middle price, with their margin taken out (their prices add up to more than 100%). It is what the edge is measured against: model chance minus market chance = edge.
Median odds
The middle price of many bookmakers, one per brand. Every "Odds" on the site is this price unless it says otherwise: Best = the highest bookmaker, Fair = 1 ÷ our model's chance, no margin. We use the middle so one odd bookmaker cannot create a fake edge.
Now ()
Today's odds (the middle price of many bookmakers, one per brand), chance and edge for a pick still to be played. They can move until kick-off. The logged price stays next to them: Logged @ 4.50 · now 4.80 ▲.
Over / Under
More than, or fewer than, the line. Over 2.5 goals = 3 or more.
Poisson
The maths the model uses to turn expected goals into the chance of every score.
ROI
Return on investment: profit ÷ total staked. -2.4% = you lost 2.4 cents per unit.
Settled
The match is over and the pick is marked won, lost or void.
Stake
The amount you put on a bet. In units here.
Stop-loss
How far your bankroll may fall from its peak (20% unless you change it). Past it, My bets asks for a pause; 24 hours later an "I reviewed my bets" button starts the count again from that day.
Unit
1% of your bankroll. Every stake in the app is in units.
Value
When the model's chance beats the market's by more than the threshold. An opinion, logged so it can be judged. Judged on 5 seasons: the model loses, so the site says "model picks", not "value".
Variance
Luck. The reason a good bettor loses some months and a bad one wins some.
Void
The bet is cancelled and the stake returned (draw no bet on a draw, postponed match).
xG
Expected goals from the chances: how many goals a team's shots were worth, from where and how each one was taken (API-Football's number for each match). 1.8 xG and 0 goals = good chances, missed. It shows how a team played, not who won, and it is never a forecast. No number from the feed = –.
We show the losses too.
18+ only. Analysis only: you decide and bet yourself. Bet only money you can lose.