170
edits
Polymarket: Difference between revisions
no edit summary
No edit summary |
No edit summary |
||
| Line 18: | Line 18: | ||
== Trading strategies == | == Trading strategies == | ||
* For large orders in markets with decent volume, place limit orders at or near the current price and wait for fills to trickle in if you don't expect the price to move quickly on its own | * For large orders in markets with decent volume, place limit orders at or near the current price and wait for fills to trickle in if you don't expect the price to move quickly on its own | ||
* have ways of knowing orthogonal to the market and bet when multiple of those point in the same direction | * have ways of knowing orthogonal to the market and bet when multiple of those point in the same direction | ||
| Line 31: | Line 28: | ||
* if you think a probability is approximately correct, place yes and no orders on both sides and then merge them together into USDC later | * if you think a probability is approximately correct, place yes and no orders on both sides and then merge them together into USDC later | ||
* For beginners, making money is more about avoiding dumb bets and bet sizing than betting smart. | * For beginners, making money is more about avoiding dumb bets and bet sizing than betting smart. | ||
* [https://www.lesswrong.com/posts/TNWnK9g2EeRnQA8Dg/never-go-full-kelly | * If you're worried an opportunity will disappear before you have a chance to bet on it, increase your bet size as you gain confidence by researching the opportunity, instead of betting all at once | ||
* If you have a smaller amount of money to bet (a smaller '''bankroll'''), bet on lower liquidity markets that aren't worthwhile for larger traders | |||
=== Fractional Kelly betting === | |||
The '''Kelly criterion''' or '''formula''' describes the theoretical optimal maximum size to bet as a portion of the participant's bankroll, assuming the bettor has perfect knowledge and zero risk aversion. '''Fractional Kelly''' (often referred to as just "'''Kelly betting'''") is a strategy that some real-world market participants use for bet sizing, which multiplies the Kelly criterion by a percent. For example, "25% Kelly" means to bet a quarter of what the Kelly criterion would suggest. Colloquially, this fraction is known as the '''Kelly factor''', and is used to describe the maximum or present bet amount relative to the bankroll. | |||
On its own, the Kelly criterion only takes into account the market probability and your probability. Fractional Kelly takes into account your confidence level. Neither take into account market spreads nor other possible competing uses of your time and money. | |||
* Beginner traders should cap their bets at 25% Kelly | |||
** This smooths out losses and cushions their emotional impact | |||
* Skilled traders should cap their bets at 50% Kelly | |||
** Traders should be prepared by having the emotional fortitude to handle losing money. Otherwise, the first time you get burned, you'll reduce your Kelly factor | |||
** Roughly, for each $1 of upside risk you get beyond 50% Kelly, you get $3 of downside risk | |||
The [https://claude.ai/public/artifacts/cc354d9e-0510-4334-a777-644449667439 Kelly Bet Calculator v1.2] ampdot made combines most of this information into an interactive tool. | |||
You can (and should!) read more about Kelly betting and the Kelly formula: | |||
* [https://www.lesswrong.com/posts/TNWnK9g2EeRnQA8Dg/never-go-full-kelly Never Go Full Kelly - LessWrong] | |||
=== Pitfalls === | |||
* Kalshi has a track record of misresolving markets or silently changing rules | |||
* Many prediction markets (e.g. Coinbase Predict) are actually frontends to Kalshi | |||
* Some markets are largely noise unless you have a detailed model of what's going on, including, in descending order of randomness: Earnings, sports, crypto, app store rankings, and mentions. Avoid betting on them unless you know what you're doing! | |||
* Avoid markets with ambiguity or misresolution risk. Markets linked to an organization taking action in response to the event may resolve based on when the organization announces it, not when the event happens. | |||
* Ask yourself if there is potential for semantic ambiguity before trading on an event contract and avoid betting if so | |||
* The user aenews is known to win markets by manipulating the oracle into misresolving markets. Check if they're betting on a market and be aware of elevated risk if they are | |||
== Trading tools == | == Trading tools == | ||
* | * | ||
* [https://betmoar.fun Market analytics that allows seeing which side has more smart vs dumb money] | * [https://betmoar.fun Market analytics that allows seeing which side has more smart vs dumb money] | ||