Educational and non-advisory. Shishin publishes stock-signal research and is one of the two services compared here, we disclose that conflict openly and hold ourselves to the same standard. Nothing here is a recommendation to buy, sell, or subscribe to anything. Every Danelfin figure below is the company’s own published claim as of mid-2026, not an independently audited fact; verify it at the source before relying on it.
Shishin and Danelfin are close cousins: both are systematic, both call themselves research rather than advice, and both try to be more legible than a tip sheet. The difference that actually matters is the kind of proof each one offers. Danelfin publishes a self-reported AI score and headline numbers that are largely backtested. Shishin publishes a cryptographically timestamped record of every call it makes, winners and losers, that you can replay yourself. Here is the honest comparison.
The short version
If you want an AI score to fold into your own screening, Danelfin is a capable, affordable tool with real explainability. If what you care about is a track record you can independently verify, one where the picks were committed before the outcomes were known, that is the specific gap Shishin was built to fill: a public, timestamped attestation and a free board that shows the misses, not just the highlights. Neither is “advice,” and neither removes the risk.
At a glance
| Dimension | Danelfin | Shishin |
|---|---|---|
| Approach | AI Score 1 to 10 per stock/ETF | Daily ranked signals from a four-engine, regime-aware model |
| The signal | A machine-learning score across fundamental, technical, sentiment features | Rule-based ranks routed by a market-regime classifier |
| Track record | Self-reported; headline numbers largely backtested (live record shorter) | Five-year backtest plus a live, publicly paper-traded forward record |
| Independently verifiable? | Self-published “AI Audit,” no external timestamping | Yes, OpenTimestamps commit-reveal at /verify |
| Past scores | Scores update daily | Published signals are never purged (append-only board) |
| Method legibility | Feature contributions shown; model itself is a black box | Mechanical momentum/breakout rules, described openly |
| Pricing (mid-2026) | Free top-picks list; paid from roughly $25 to 29/mo | Free (delayed) board; paid from $20/mo |
| Advice or research? | Research (non-advisory) | Research (non-advisory) |
| Best for | An AI score as one screening input | A rules-based edge with a record you can audit |
What they actually share
It is worth being fair about the overlap, because it is real and most “versus” pages pretend it away.
- Both are systematic. Neither is a pundit picking stocks on a hunch. Danelfin scores from a model; Shishin ranks from rules. Removing the discretionary human is the point in both.
- Both are research, not advice. Neither tells you what to do with your money, and that is a feature, see what a signal can and cannot do.
- Both try to be legible. Danelfin shows which features moved a score; Shishin describes its engines and what each column on the board means. Both are more transparent than a sealed “buy this” alert.
- Both have a usable free tier. Danelfin publishes a free top-picks list; Shishin’s full board is free on a delay, with no email wall to see it.
Where they differ: the proof
The overlap ends at the one question that decides whether a track record means anything, can you check it yourself?
1. Backtested headline, or forward record?
Danelfin’s most-quoted outperformance figures are backtested, the result of running today’s model over yesterday’s data. The company’s own materials date its live, forward AI Score history to 2025, which is honest of them and also the whole point: a backtest is a hypothesis, not a track record. Curve-fitting, look-ahead, and survivorship bias make in-sample numbers look spectacular and prove nothing, that is exactly why backtests lie. Shishin publishes a five-year backtest and a live paper-traded forward record run on the same rules, and labels which is which. The backtest is the hypothesis; the forward log is the test.
2. Self-published audit, or external attestation?
Danelfin publishes an “AI Audit” on its own site. That is more than most competitors offer, but it is self-published: the company vouches for its own numbers, with nothing an outsider can use to prove the record was not adjusted after the fact. Shishin takes the opposite route. Every day’s signals and net-asset value are hashed and anchored to the Bitcoin blockchain through OpenTimestamps, a commit-reveal scheme, so anyone can confirm a given day’s calls existed, unchanged, at that date. The full log lives at /verify and the raw commitments are public. Self-reported and independently attested are not the same word.
3. Mutable scores, or an immutable board?
A Danelfin AI Score is a snapshot that updates daily; there is no public append-only ledger of what every past score was, so you cannot replay the history the way it was actually published. Shishin never purges a published signal. What the board said on a given date stays on the record, winners and losers alike, which is the only way a selective memory can’t quietly creep in.
4. A model you infer, or rules you can read?
Danelfin surfaces feature contributions, which genuinely helps, but the model underneath is a machine-learning black box, its win-rate claim (self-reported, above 60% for its top bucket) is not independently audited. Shishin’s method is mechanical: momentum and breakout rules, a regime classifier that routes between guardians, described in plain terms across this library. A method you can reason about is one you can keep trusting in a drawdown, when a black box is hardest to hold.
How to tell a backtest from an attested record
This is the transferable skill, and it applies to us as ruthlessly as to anyone. When any service shows you a number, ask:
- Was it committed before the outcome? A pick timestamped in public before the move is evidence. A number computed afterward is a hypothesis wearing a track record’s clothes.
- Can you reproduce it? Real transparency means every call, dated, including the losers, so you can re-run the tape. See how to vet a track record.
- Who is vouching? Self-published is a start; externally anchored (a blockchain timestamp, a third party) is stronger, because it removes the “trust me” step.
- Are the misses shown? A survivorship-scrubbed record that shows only the winners tells you almost nothing.
Both the SEC and FINRA warn that signal services routinely advertise vague, cherry-picked, unverified past returns, showing the winners and burying the losers. The defense is not to trust the prettier brand; it is to demand the proof, of us included.
Which one fits you
- Choose Danelfin if you want an accessible AI score as one input into your own research, value the per-stock explainability, and are comfortable that the headline numbers are largely backtested.
- Choose Shishin if the thing you cannot get elsewhere is a forward record you can independently verify, a free board that shows the misses, and a method you can read rather than infer.
- Our honest weakness: Shishin is new. We publish a five-year backtest plus a still-short livepaper-traded record, not a decade of real-money history. Weigh that exactly as skeptically as you should weigh a backtested headline. The difference is that ours is falsifiable, and dated in public.
There is no single winner here, only the better fit for what you are trying to trust. If you already believe a good-looking backtest, Danelfin is a fine tool. If you have learned to distrust one, that is the reason Shishin exists.
Sources & further reading
- Danelfin, product and “AI Audit” materials (AI Score methodology, performance, pricing), the company’s own published claims as of mid-2026. danelfin.com
- FINRA, “Know the Risks of Auto-Trading Services Offered by Unregistered Entities.” finra.org
- U.S. SEC, Office of Investor Education and Advocacy, Investor Alerts & Bulletins. investor.gov
- Shishin, the public attestation log. shishin.io/verify. See also the full field compared and why a verifiable record beats a famous face.