Honest comparison
More is not value. Follow the money.
A longer feature list, a slicker app, a free tier — none of it tells you whether a platform is on your side. One question does: what does your success, your losses, and your doing nothing each do to their revenue? Here's that lens run across the industry — and, at the bottom, across us.
The one question
Incentive gradient beats feature count
Every product optimises for its own revenue — that's not cynicism, it's just how businesses survive. So the useful question isn't whether a company wantsto help you; it's whether helping you is how it gets paid. When the two come apart, the incentive gradient wins over years, no matter how good anyone's intentions are. Three revenue shapes recur, and each bends the product a specific way:
Order flow, spread, per-trade fees. The product is pulled toward making you trade more and more impulsively— your inactivity is the thing it's built to prevent.
Subscriptions for engagement. Pulled toward keeping you feelinglike you're winning — which is why flattering backtests and streaks survive and honest costs don't.
Alpha harvesting, IP on their servers. Pulled toward owning what you make — the value flows from you to the house, dressed as a platform.
The follow-the-money matrix
What each model earns when you win, lose, or walk away
Read the last three columns as: does their revenue move with your success, indifferent to it, or against your interest?
| Model | Paid by | You win | You lose | You go idle |
|---|---|---|---|---|
Gamified zero-commission brokerages | Payment for order flow (PFOF), margin lending, securities lending, options volume | indifferent — they earn on the transaction, not your outcome | still paid, as long as you keep trading | you stop trading → their revenue stops |
Exchanges | Trading fees (% of volume), spread, listings, custody, staking cut | the fee is the same whether you win or lose | the fee is the same whether you win or lose | no volume → no fee; leverage & frequency are the ask |
Charting subscriptions | Subscription — you are the paying customer | you pay them directly; better alignment than PFOF | fine while you stay subscribed | honest-looking bad backtests → churn → lost sub |
Quant clouds & crowdsourced alpha | Metered cloud compute + data fees; or a fund built on your predictions | billed for compute/data regardless of your returns | still billed; or your prediction just underperforms | lost compute/data revenue, or lost data-labor |
Copy-trading & signal sellers | Spread on copied volume; or a subscription for the signals themselves | they earn on the spread / the sub either way | the signal seller is paid whether or not it worked | no follows / no subs → no revenue |
AI black-box bots | Subscription, regardless of performance | paid the same whether it works | paid the same whether it works | revenue depends on marketing, not results |
Mederos | Paid depth & progression on top of a free core; native app. No order flow, no spread, no cut of your volume, no lock-in, no claim on your alpha. | a tool that genuinely helps is one you keep — our revenue rises with it being good | if it doesn't help, you leave and we lose. That's the incentive we want | your dissatisfaction costs us — not your inactivity, so we push usefulness, not frequency |
Note the shape of the last row: for the volume models, your inactivity is what costs them — so they push you to act. For us, your dissatisfactionis what costs us — so we're pushed to be worth keeping.
Category by category
The honest version of each
For each: what they're genuinely good at, the structural catch their revenue creates, and who they actually serve best. We're attacking incentive structures, not anyone's integrity — good people inside a volume-priced business still ship a volume-priced product.
Gamified zero-commission brokerages
the $0-commission app archetypePaid by: Payment for order flow (PFOF), margin lending, securities lending, options volume
Genuinely frictionless. Beautiful onboarding, $0 commissions, fractional shares — they lowered the access barrier for a whole generation, and that was real.
The paying customer is the market maker buying your order flow — not you. Revenue scales with how often and how impulsively you trade, which is why the product is engineered for engagement: streaks, nudges, confetti, options prompts, push notifications. A tool that helped you trade less but better would cannibalise the revenue line, so there is a structural ceiling on how good the analytics will ever be allowed to get.
Exchanges
spot & derivatives venuesPaid by: Trading fees (% of volume), spread, listings, custody, staking cut
Deep liquidity, custody, a real matching engine, regulatory rails. Genuinely hard infrastructure that is genuinely valuable when you actually need to execute.
The 'pro' charting and tooling are retention loss-leaders: good enough that you don't leave for a dedicated tool, and capped right there — because anything that made you trade less works against a fee model priced on volume. It's an execution venue wearing a tool's clothes.
Charting subscriptions
the dominant charting-social platformPaid by: Subscription — you are the paying customer
Best-in-class charting UX, an enormous community, a vast script library, broad real-time data. We are not trying to beat their chart or their social network.
Two structural pulls. Their scripting language is a walled garden — your strategy can't leave, run natively/offline, or be truly owned. And repainting and lookahead are widely-documented backtest-inflation patterns the platform tolerates, because honest costs and no-lookahead make strategies look worse → sadder users → churn. Retention quietly rewards backtests that flatter.
Quant clouds & crowdsourced alpha
hosted-backtest clouds; prediction tournamentsPaid by: Metered cloud compute + data fees; or a fund built on your predictions
Serious, institutional-grade backtesting, real data catalogs, and — for the tournament model — a genuinely novel meta-model. Real engineering, no question.
Your IP and your backtests live on their infrastructure, and compute is metered — you pay to search your own ideas, through a data moat that sits between you and your own research. The tournament variant is starker: you are the data laborer, your edge feeds their fund, and you're paid in a lottery, not ownership. (The cautionary tale of the community that got shut out after its strategies were harvested is this model's failure mode.)
Copy-trading & signal sellers
copy-trade leaderboards; paid signal groupsPaid by: Spread on copied volume; or a subscription for the signals themselves
Low-friction social discovery. For a curious beginner, seeing what others do is a real and human on-ramp.
Copy-trade leaderboards are survivorship-selected, and 'popular investors' are incentivised to gain followers — volume, spread — not to be right. And selling signals is adverse selection: if an edge were real and capacity-limited, why sell it instead of trade it? The seller's edge is selling the signal, not the signal.
AI black-box bots
'we'll trade for you' subscription botsPaid by: Subscription, regardless of performance
The appeal of hands-off. Convenience is a real want, and they sell it well.
No pre-registration, no reproducibility, survivorship-marketed. You can't read the rules, so you can't separate skill from luck from fabrication. 'Trust us' is the entire product — and opacity is the feature, because a verifiable claim could be falsified.
Now turn it on us
The same lens, pointed at Mederos
A comparison page that only indicts everyone else is just marketing with extra steps. So here's us in the same matrix, limitations first.
Mederos
the desk you ownPaid by: Paid depth & progression on top of a free core; native app. No order flow, no spread, no cut of your volume, no lock-in, no claim on your alpha.
We're pre-launch and small, the flagship is one pre-registered strategy rather than a hundred, and we're asking you to install software and link your own accounts. None of that is frictionless.
Turn the lens on us and the honest risk is the game/economy layer — we could engineer engagement like anyone else. The guardrails are structural: we never touch your order flow, so we have no reason to juice your trade frequency; the scheduler only suggests and a human approves every order; and we publish NO-GOs, which an engagement-maximiser would bury.
We never route your orders
Execution is a human-gated hand-off to your own linked accounts. No order flow to sell means no reason, ever, to want you trading more.
We don't want your alpha
Strategies run on your machine and compile to code that leaves with you. We sell tooling and rigor — a business that survives only if the tool is genuinely good.
We publish the NO-GOs
Negative results are logged with the same weight as positive ones. Honesty is the one thing a volume- or engagement-maximiser structurally can't offer — so it's the moat.
Anti-fit
Who Mederos is not for
The fastest way to respect your time is to tell you when to leave.
- If you want confetti, streaks, and a dopamine loop — that's a different product category, and it's very good at being that.
- If you want a done-for-you black box you never have to understand — we can't help, because understanding it is the whole point here.
- If you want someone to blame when a trade goes wrong — we hand you the controls and the honest costs; the decisions stay yours.
- If you're not willing to install software and link your own accounts — the local-first, keep-your-own-edge model needs a little setup that a hosted app doesn't.
Check every word of this yourself
Read the architecture, run the engine in your browser, and hold our numbers to their pre-registration. We'd rather you verify than believe.
This page is our own analysis and opinion of publicly-documented business models and incentive structures, offered as fair comment and comparison. Company and product names, where referenced, are trademarks of their respective owners; Mederos is not affiliated with, endorsed by, or partnered with any of them. Business models evolve — verify current specifics yourself before relying on them. Nothing here is investment advice, and backtested or hypothetical results are not indicative of future performance.