The Tipster Problem
Twitter is full of "player prop tips" accounts. Most post selections without context: "Haaland 2+ shots on target @ 2.10". No reasoning. No methodology. No accountability.
Here's why this approach fails long-term:
No Value Context
A tip at 2.10 means nothing. Is that good value? What should the fair price be? Without comparing to true probability, you're gambling blind.
Cherry-Picked Results
Tipsters post winners, hide losers. A 60% strike rate sounds good until you realize the average odds were 1.50 — negative EV overall.
No Position Awareness
A midfielder playing as a winger has different shot expectations. A right-back asked to push forward changes tackle predictions. Position matters.
Static Analysis
Last season's stats don't account for new managers, formations, or fitness. The market moves — static analysis doesn't.
What to Look For Instead
The difference between losing bettors and sharp bettors isn't prediction accuracy — it's process.
Value Percentage
Don't just know the odds — know if they're good odds. A 10.00 shot on a player with 5% true probability is terrible value. A 3.00 shot on 40% probability is excellent.
Recent Form (Not Season Stats)
Last 5-10 games matter more than season averages. A striker returning from injury won't match his pre-injury xG immediately.
Position Context
Primary vs alternative position changes everything. A natural striker playing on the wing will have fewer shot opportunities.
Minutes Played
A player averaging 60 minutes per game has different expectations than a 90-minute starter. The model must account for this.
Opponent Adjustment
Shot totals against Burnley's deep block differ from shots against City's high press. Context matters.
The Cerebro Model Approach
Proppr's Player Bot uses the Cerebro statistical model — hard-coded mathematics, not AI. Here's what makes it different:
Data Window
Last 5 or Last 10 games
Position Modeling
Primary + alternative positions
Model Confidence
60-70% on value plays
Markets Covered
10+ per player
Markets That Offer Edge
Not all player prop markets are created equal. Some have lower bookmaker margins, meaning more edge potential:
The Super Sub Edge
Bookmakers undervalue substitutes. Fresh legs against tired defenders at 70th minute create value. The /supersub command surfaces bench players with edge — prices the market ignores.
Use It As a Tool, Not a Tipster
As Jay (Prof. X), the Proppr creator, says:
"The main thing with the bot is it's NOT a tipster, taking every alert it sends isn't the way to go, but use it as a tool to point you in the right direction you can't go wrong."
The bot identifies mathematical value. You decide whether the context supports the bet. That's the difference between a tipster and a research tool.