PickBot Intelligence Lab
See how PickBot thinks.
Explore an interactive demonstration of how market price, implied probability, matchup signals, risk and uncertainty combine when a betting opportunity is evaluated.
Sample analysis for demonstration purposes. Not live odds, not a PickBot record, and not a guarantee of results.
Sample analysis console
Idle — awaiting run
0 / 7
- Market scan
- Probability engine
- Matchup signals
- Risk engine
- Edge detection
- Decision
- Exploration
Sample matchup
SAMPLEHarbor City at Metro Athletic
Metro Athletic -3.5
Fictional clubs in a fictional league. Nothing on this page refers to a real fixture, a real price or a real result.
Market scan
SAMPLE- Spread
- -3.5 -110
- Moneyline
- -150 +130 other side
- Total
- 47.5
- Previous spread
- -2.5
- Movement
- +1.0 pts toward Metro Athletic
Line movement
- Opened -2.5
- Now -3.5
- Edge zero at -5.0
Signal matrix
SAMPLE SIGNALSSix categories, each reduced to one of three states. Three are computed from the sample market inputs, one is a fixed sample value, and none of them is a PickBot production signal.
-
Market structure POSITIVE
Two-sided moneyline holds 3.36%.
-
Matchup POSITIVE
Fixed sample input. Identical at every line on the slider.
-
Price / value POSITIVE
Sample model above the de-vigged price by 4.5 pts.
-
Recent signal POSITIVE
Line moved -2.5 to -3.5, 1.0 pt toward Metro Athletic.
-
Volatility POSITIVE
The move has already consumed 3.0 pts of cover probability.
-
Uncertainty NEUTRAL
Sample margin dispersion 13.2 pts against a 5.0 pt expected margin.
PickBot analysis core
SAMPLEIdle
Awaiting run
- MARKET 3.36% hold POSITIVE
- PROBABILITY 54.5% cover POSITIVE
- MATCHUP sample input POSITIVE
- RISK moderate NEUTRAL
- VALUE +4.5 pts POSITIVE
Edge detection
DEMONSTRATIONMarket, margin removed
50.0%
Sample model estimate
54.5%
Demonstration edge
+4.5 pts
The difference between the sample model estimate and the same market’s price with its margin taken out. It is called a demonstration edge because the model behind it is a stated assumption on this page, not a PickBot production model.
Demonstration decision
NOT A PICKBOT PICKSTRONG PLAY
- PASS
- LEAN
- PLAY
- STRONG PLAY
At this price the sample model and the de-vigged market disagree by more than the demonstration threshold, and the signals mostly agree with each other.
Demonstration edge < 0.75 pts = PASS · 0.75 to 2.25 = LEAN · 2.25 to 3.75 = PLAY · 3.75 pts and above = STRONG PLAY. Elevated risk steps the result down one rung.
This is an illustration of several signals converging, not a recommendation, not a prediction, and not the process PickBot uses to publish a pick.
Implied probability
CALCULATEDReal arithmetic, not a sample value. An American price converts to a probability with one formula, and the two sides of a market always sum to more than 100% — the excess is the book’s margin.
Market implied probability
60.0%
Metro Athletic moneyline at -150
- Metro Athletic -150
- 60.00%
- Harbor City +130
- 43.48%
- Market hold
- 3.36%
- Margin removed
- 57.98%
Market hold is what the book keeps if it balances both sides: one minus the reciprocal of the two probabilities added together. Removing it proportionally gives the price without the margin, which is the number worth comparing a model against.
- Spread priced -110 / -110
- 52.38%
- Spread hold
- 4.55%
- Spread margin removed
- 50.0%
Risk engine
SAMPLEHow much of what is on screen is uncertain. This panel deliberately says nothing about stake size — more uncertainty is a reason for less conviction, never for more exposure.
Risk
MODERATE
35 of 100 on the demonstration index
-
Line volatility 3.0 pts consumed
How much cover probability the market has already taken out of this position by moving.
-
Market disagreement 0.8 pts apart
The gap between the margin the spread implies and the margin the moneyline implies, read through the same sample dispersion.
-
Signal conflict 5 positive / 1 neutral / 0 negative
How far the six signals disagree with one another. Agreement is not accuracy, but disagreement is a reason to hold back.
-
Price sensitivity 1.50 pts per half point
How much of the case disappears if the number moves half a point further.
-
Uncertainty 2.64x the expected margin
The sample margin dispersion set against the sample expected margin. A wide distribution makes any single outcome weak evidence.
The sample inputs mostly agree, but the position is sensitive to the price.
Price matters
What would change this pick?
Move the sample spread. The matchup input does not change — only the number does. Watch the edge, the risk and the demonstration decision follow the price.
Why this pick?
Show the breakdown
Supporting signals
- Market structure — Two-sided moneyline holds 3.36%.
- Matchup — Fixed sample input. Identical at every line on the slider.
- Price / value — Sample model above the de-vigged price by 4.5 pts.
- Recent signal — Line moved -2.5 to -3.5, 1.0 pt toward Metro Athletic.
- Volatility — The move has already consumed 3.0 pts of cover probability.
Counter-signals
- Uncertainty — Sample margin dispersion 13.2 pts against a 5.0 pt expected margin.
- Largest risk factor — uncertainty, at 2.64x the expected margin (66 of 100).
What would invalidate it
- The line reaching -5.0. That is where the sample model's cover probability equals the de-vigged market price and the demonstration edge is exactly zero.
- A further half point. Each half point is worth about 1.50 pts of cover probability at this part of the distribution.
- The sample matchup input falling below 60 of 100, which would remove the one signal that does not move with the price.
- Risk reaching ELEVATED, which steps the demonstration decision down one rung on its own.
Sample analysis for demonstration purposes. Not live odds, not a PickBot record, and not a guarantee of results.
How it works
What the Lab is actually doing
Everything above is one of three things: arithmetic with a single correct answer, a sample input stated in full, or a demonstration threshold printed on the page. There is no fourth category and nothing is fetched from anywhere.
American odds to implied probability
CALCULATED
A negative price divides its own size by itself plus 100. A positive price divides 100 by itself plus 100. One formula, one answer.
Market hold
CALCULATED
One minus the reciprocal of the two sides added together. It is what the market keeps, and it is why a raw price always overstates a probability.
Margin removal
CALCULATED
Each side divided by the two added together. A disclosed, ordinary method — not the only one in use, which is why the page names it.
Line movement
CALCULATED
The current number minus the previous number, and the cover probability that move consumed.
Sample model estimate
SAMPLE INPUT
One stated assumption: the margin is treated as a normal distribution with a 5.0 point mean and a 13.2 point standard deviation. Both numbers are invented for this page.
Cover and win probability
CALCULATED
Read straight off that distribution: the chance the margin exceeds the spread, and the chance it exceeds zero.
Matchup signal
SAMPLE INPUT
A fixed number standing in for whatever a real matchup model would contribute. Constant at every slider position, on purpose.
Signals, risk and decision
DEMONSTRATION
Thresholds chosen for this page and printed on it, so every transition can be reproduced by hand.
What this page is not. No figure here is a live betting line, a PickBot production model output, a PickBot win rate, a PickBot record, an ROI, a bankroll result or a current team statistic. PickBot publishes method rather than numbers until its results ledger can substantiate them, and a demonstration is not an exception to that.
Understand the process.
You have seen how market price, probability, risk and uncertainty work together — and how quickly a position stops being worth taking when the number moves.
Free
Join the free PickBot community
Learn, ask questions, discuss betting concepts and connect with other sports bettors.
Join the Free CommunityMembers
Explore PickBot Inner Circle
For members who want PickBot’s picks delivered to them. Inner Circle is a delivery channel, not a discussion community.
Explore Inner CircleSample analysis for demonstration purposes. Not live odds, not a PickBot record, and not a guarantee of results.