Track record
Every signal, win or lose
Each value signal is frozen when published and settled after the match. Results assume a level stake of 1 unit at the published price. This is a record of the past, not a forecast: a positive record can turn negative.
Cumulative result, 1 unit per signal · last 90 days
Not enough settled results yet.
Tracking has just started
Signals are settled after their matches finish, so the first results appear within a day or two. Nothing is back-filled: the record only contains signals published before kick-off.
Signals settled (90 days)
0
0 won · 0 lost · 0 void
Hit rate
–
model expected –
Level-stake result
0.00 u
ROI – · avg price –
Closing-line value
–
Brier score –
Calibration
Do outcomes the model gave X% happen about X% of the time?
| Model probability | Signals | Expected | Actual |
|---|---|---|---|
| 0–30% | 0 | – | – |
| 30–45% | 0 | – | – |
| 45–60% | 0 | – | – |
| 60–100% | 0 | – | – |
Small samples swing a lot. Hundreds of signals are needed before these numbers mean much.
By market and sport
| Group | Settled | Hit rate | Units | CLV |
|---|---|---|---|---|
| Match result | 0 | – | 0.00 u | – |
| Over/under | 0 | – | 0.00 u | – |
| Free leagues | 0 | – | 0.00 u | – |
| Pro leagues | 0 | – | 0.00 u | – |
Recently settled
| Date | Match | Signal | Price | Model | Result | Units |
|---|---|---|---|---|---|---|
| Nothing settled yet. | ||||||
How to read this
- Hit rate vs expected: value signals are often on outsiders, so most of them lose even when the model is right.
- Closing-line value (CLV): how the published price compares with the consensus just before kick-off. Consistently positive CLV is the best evidence of skill; profit alone is mostly luck over short periods.
- Brier score: accuracy of the probabilities (0 = perfect, lower is better).
- Real-world results would be worse than shown if you got lower prices than those published.