Note: The "Today's Picks" feature described in this article has since been folded into Ryujin Radar's single unified prediction model and is no longer offered as a standalone feature. What follows is a record of that period's results. For current AI predictions and tracked hit rates, see the performance page.
Introduction
"Breaking a 100% return rate in boat racing is impossible."
Do you think that too?
Ryujin Radar's "Today's Picks" feature has recorded a cumulative 104% return rate about a month and a half after launch. That means a 4% net profit relative to the amount wagered.
This article publishes the results from 2,577 races and explains what this number actually means.
The results (January 16 – March 1, 2026)
| Metric | Value |
|---|---|
| Period | From January 16, 2026 |
| Races covered | 2,577 races |
| Hits | 772 |
| Hit rate | 30% |
| Wagered (¥100 per race) | ¥257,700 |
| Returned | ¥269,170 |
| Return rate | 104% (+¥11,470) |
This is the result of betting a flat ¥100 on every single recommended race — no cherry-picking, no adjusting amounts. It's a fully mechanical result.
Just how impressive is this?
It clears boat racing's "wall"
Boat racing has a deduction rate (the operator's cut) of roughly 25%. That means betting randomly, with no strategy at all, theoretically converges toward a return rate of about 75%.
| Approach | Expected return rate |
|---|---|
| Random betting | ~75% |
| Typical prediction sites | 70–90% (*many don't disclose) |
| Ryujin Radar standard predictions | 88–94% |
| Ryujin Radar Today's Picks | 104% |
"Today's Picks" turns the tables on the 25% deduction rate and comes out ahead.
What 2,577 races as a sample size means
"Isn't that just luck?" is a fair question.
Statistically, the larger the sample, the more reliable the result. A sample of 2,577 races is large enough to demonstrate a reproducible pattern, not a random fluke.
For reference:
- The probability of flipping a coin 2,577 times and getting "heads 53% of the time or more" is about 1.2%
- In other words, this is a result that's 98.8% unlikely to be chance
Why can it beat a 100% return rate?
The difference between "predicting every race" and "curating"
Ryujin Radar's standard predictions cover every single race. Naturally, some races are easier to predict than others. Averaged across everything, the return rate falls short of 100%.
"Today's Picks" takes a different approach:
Thoroughly mine past data to find statistically proven conditions where the return rate exceeds 100%, then extract only the races that meet those conditions.
In other words:
- Standard predictions → participate in every race and land on an average score
- Today's Picks → only participate in the races you can win
That "courage to pass" is the source of the over-100% return.
The "distortions" data mining uncovers
Boat racing odds are shaped by the collective betting behavior of a huge number of fans. But human judgment has biases.
- Patterns where, under specific conditions at a specific venue, the odds run higher than the racer's actual ability warrants (i.e., undervalued by the market)
- Patterns where, for specific race numbers or time slots, a statistically favorable bet exists
These "odds distortions" only become visible through mechanical analysis of thousands of races — patterns that are hard to catch through human intuition or experience alone. They're patterns only data can reveal.
What we should be honest about
104% isn't "a fortune"
Against a total wager of 2,577 races × ¥100 = ¥257,700, the profit is ¥11,470. Honestly, that's not enough to live on by itself.
But what matters is the fact that it maintains a net-positive return. Many prediction services are considered "excellent" at a 80–90% return rate — simply exceeding 100% is rare in itself.
Not every rule wins consistently
"Today's Picks" is built from multiple patterns (rules). Among them:
- Rules that have performed well consistently since launch
- Rules currently going through a temporary rough patch
both exist side by side. The overall return rate holds at 104%, but individual rules have their ups and downs.
Past performance doesn't guarantee the future
Even statistically valid patterns can shift over time due to changes in the boat racing environment — racer turnover, venue renovations, seasonal variation, and more.
Ryujin Radar periodically re-validates each rule's effectiveness and revises any rule whose performance has deteriorated.
How to use "Today's Picks"
Basic usage
- Go to the Ryujin Radar's latest performance page
- Check today's recommended races
- Tap a race you're interested in to see the details
- Check the bet type (Win, Place, Trio, Trifecta) before deciding
Tips for using it effectively
If you're a beginner: Start with the Place recommendations. They have a high hit rate and give you a stable, low-stakes way to get a feel for the feature.
Bankroll management matters: Fixing your bet size per race and betting evenly across all recommended races is, according to the data, the most reproducible approach. Letting emotion drive your bet size up or down erodes the statistical edge you're relying on.
Summary
| Point | Detail |
|---|---|
| Return rate | 104% (based on 2,577 races) |
| Hit rate | 30% (772 hits) |
| Mechanism | Automatic matching against high-return patterns discovered through data mining |
| Strength | Turns the 25% deduction rate around into a net-positive return |
| Caveat | Past performance, not a guarantee of future results |
Ryujin Radar's "Today's Picks" is a feature that selects only "the races with the highest win probability" out of the 100+ races run each day.
Daily results are updated on the page in real time. We don't hide the bad numbers any more than the good ones.
If this piqued your interest, check out Ryujin Radar's latest performance.
Tags: #RyujinRadar #TodaysPicks #104PercentReturn #DataMining #BoatRacing #TrackRecord #2577Races