Meta Description: A bot called "Nothing Ever Happens" systematically buys No on Polymarket prediction markets — and it's profitable. Here's why most predicted events don't happen, and what it teaches us about decision-making.
The Simplest Strategy That Actually Works
Here's a trading strategy so simple it sounds like a joke: bet against everything. No analysis, no research, no domain expertise. Just systematically buy "No" on every non-sports prediction market you can find.
That's exactly what Nothing Ever Happens does. It's an open-source Python bot by Sterling Crispin that scans Polymarket — the world's largest prediction market — and buys No shares on standalone yes/no markets. No sports (those are efficient). No complex multi-outcome markets. Just binary bets on whether something will happen, and the bot always says it won't.
And it makes money.
Not "maybe kinda sorta" money. Consistent, systematic, boring profit. Because the uncomfortable truth about prediction markets — and human psychology in general — is that people dramatically overestimate the probability of interesting things happening.
Why "No" Is Undervalued
The Base Rate Problem
Most events that people create prediction markets about are, by definition, unusual. "Will Bitcoin hit $150K by June?" "Will there be a US-China military conflict before 2027?" "Will a category 5 hurricane hit Florida this season?"
These are exciting questions. That's why they get markets in the first place. But excitement and probability are inversely correlated. The more dramatic an event, the less likely it is to occur — and the more humans tend to overestimate its probability.
This is a well-documented cognitive bias called base rate neglect. When asked to judge the probability of an event, people focus on the specific, vivid details of the scenario rather than the statistical baseline. "Will there be a ceasefire by April?" feels more likely when you're reading news articles about peace talks than when you're looking at the historical base rate of ceasefire agreements actually being implemented.
Prediction markets are supposed to be the antidote to this — the "wisdom of crowds" aggregating information into a single number. But crowds are made of people, and people are biased toward interesting outcomes.
The Market Structure That Makes This Work
The bot specifically avoids sports markets. That's not an accident. Sports betting markets are liquid, heavily traded, and extremely efficient. The odds on a Premier League match reflect genuine information — team form, injuries, weather, historical matchups. There's no structural edge in betting "this team won't win" across all matches.
Non-sports prediction markets are different. They're thinner, driven by narratives and news cycles, and populated by participants who are there because they care about the outcome — not because they're pricing it efficiently. A market about a political event attracts people with political convictions, not dispassionate probability assessors.
This creates a systematic upward bias in "Yes" prices. Events get priced higher than their true probability because the people trading them want them to happen, or because the narrative around them makes them feel likely.
How the Bot Works
The architecture is straightforward:
- Scan — The bot continuously scans Polymarket for standalone yes/no markets (not multi-outcome, not sports)
- Filter — It looks for "No" shares trading below a configured price cap — cheap "No" contracts on events the market thinks are likely
- Buy — It purchases those "No" shares
- Wait — When the event doesn't happen (which is the base case for most events), the "No" shares resolve to $1 and the bot profits
The bot is built with async Python, runs on Heroku with a single web dyno, and includes a dashboard for monitoring positions. It has paper trading mode (default), live trading (requires explicit environment flags), and full state recovery for restarts.
The safety rails are real: BOT_MODE=live, LIVE_TRADING_ENABLED=true, and DRY_RUN=false must all be set, plus a private key and funder address. Without all of them, it defaults to paper trading. This is clearly a project by someone who doesn't want to accidentally lose real money.
The Deeper Lesson: Most Things Don't Happen
This bot is profitable because of a fundamental truth about the world that most people — including smart, well-informed people — systematically underestimate:
Most things that could happen, don't.
- Most proposed legislation doesn't pass
- Most rumored acquisitions don't materialize
- Most geopolitical crises de-escalate
- Most "unprecedented" events have precedents that showed they were precedented
- Most deadlines get extended
- Most breakthrough announcements are incremental
This isn't pessimism. It's statistics. The prior probability for any specific event is almost always low, because the space of possible events is enormous and only a tiny fraction of them actually occur.
Prediction market participants — and humans in general — consistently fail to price this correctly. They see a market priced at 30% "Yes" and think "that seems reasonable, there's a real chance." But the base rate for most prediction market events resolving "Yes" is probably closer to 15-20%. That gap is the bot's profit margin.
What This Means for Decision-Making
For Founders and Builders
Every startup decision involves prediction. "Will this market be big enough?" "Will this competitor ship before us?" "Will this partnership close?" "Will this feature move the needle?"
The "Nothing Ever Happens" heuristic suggests you should:
- Default to the boring outcome. Competitors won't ship that feature on time. The partnership will take 3x longer than expected. The market won't grow as fast as the reports say.
- Treat exciting predictions with extra skepticism. If someone's excited about an outcome, they're probably overestimating its probability.
- Build for the base case, not the best case. Your business model should work if the boring thing happens. The exciting thing is a bonus.
For Investors
If prediction markets systematically overestimate the probability of interesting events, then:
- Narrative-driven investments are overpriced. If everyone's talking about it, the upside is already baked in.
- The best trades are boring. Buying "No" on exciting events is like buying index funds instead of meme stocks.
- Contrarian isn't always right, but consensus is often wrong in predictable ways. The crowd isn't wrong about everything — it's specifically wrong about the probability of change.
For Anyone Making Decisions Under Uncertainty
The next time you're estimating whether something will happen, try this:
- Find the base rate. How often does this type of thing actually happen? Not this specific instance — the category.
- Adjust slightly for specifics. If the base rate is 10%, the specifics might move it to 12% or 8%. They rarely move it to 40%.
- Check for emotional bias. Are you estimating high because you want it to happen, or because the evidence actually supports it?
The Risks (Because Nothing Is Free)
This strategy isn't without danger:
- Black swans are expensive. When something does happen that the market thought wouldn't, you lose. And since you're betting against interesting outcomes, the losses happen precisely when everyone is paying attention.
- Market dynamics can shift. If this strategy becomes popular, "No" prices will correct upward and the edge disappears. The edge exists because most participants are on the other side.
- Liquidity constraints. You can only bet "No" on markets that exist and have enough liquidity to fill your orders. The most overpriced "Yes" markets might be the ones you can't actually trade.
- Resolution risk. Prediction markets depend on clear resolution criteria. Ambiguous outcomes can lead to disputes, and you're exposed to the platform's resolution mechanics.
The Takeaway
"Nothing Ever Happens" is a brilliant project not because it's a sophisticated trading system — it's deliberately not. It's brilliant because it exploits a fundamental human bias with minimal complexity. The entire strategy fits in a sentence: bet against interesting things, because interesting things rarely happen.
In a world obsessed with predicting the future — from startup pitch decks to geopolitical analysis to your fantasy football league — the most powerful insight might be the simplest one. The future is mostly like the present. Things continue. Status quo wins. Nothing ever happens.
And if you want to make money off that realization, the code is open source.
FAQ
Q: How much money has the Nothing Ever Happens bot made? A: The creator hasn't disclosed exact profits. The bot is open-source and provided "for entertainment only." The strategy's edge comes from systematically exploiting base rate neglect in non-sports prediction markets, but returns depend on market conditions, position sizing, and liquidity.
Q: Can I run this bot myself? A: Yes, the code is on GitHub. You'll need a Polymarket account with funded wallets, a Polygon RPC endpoint, and a Heroku account (or similar) to host it. The bot defaults to paper trading — live mode requires explicitly setting three environment flags plus your private key.
Q: Why does this strategy not work on sports markets? A: Sports betting markets are highly liquid and efficient. Thousands of professional bettors, bookmakers, and algorithms price sports events with extreme accuracy. Non-sports prediction markets are thinner and dominated by participants with emotional or narrative-driven biases, creating the mispricing this bot exploits.
Q: What is base rate neglect? A: Base rate neglect (or base rate fallacy) is a cognitive bias where people ignore the general probability of an event happening in favor of specific, vivid information. For example, hearing about peace talks makes a ceasefire "feel" likely, even though historically, most ceasefire negotiations fail.
Q: Is prediction market betting legal? A: Polymarket operates on the Polygon blockchain and is available in many jurisdictions, but regulatory status varies. In the US, prediction markets for events (not sports) exist in a gray area with some CFTC oversight. Check your local regulations before participating.
Inspired by Sterling Crispin's Nothing Ever Happens bot. Not financial advice. Really.