Spoiler: I don't have an answer. This is just a pile of comparison points.

When placing a bet, there are many reasons to want it to be well capitalized (aka low leverage); basically they amount to making sure the loser both can and will pay up. But bets are not the same as all other financial assets, and the answer isn't always obvious. We would do well to compare to other cases and reason carefully about what makes sense, rather than simply pattern matching or assuming.

Stocks are traditionally fully capitalized, as are loans, for good reason: the party receiving the loan or selling the stock actually needs the full amount now. That's the point; the goal isn't just risk exposure, it's the actual cash for purpose of paying a bill or an employee. You don't put the money in escrow or prove you have the ability to pay or pass a credit check, you actually pay the amount now. However, this is not the only technique available in the financial world.

People who want to bet on interest rates could take out or supply loans. But since the point is the financial risk exposure, they usually find other techniques. Interest rate derivatives and swaps provide a method to define the risk and take it on, but with far less capital than the exposure would imply if structured as a loan.

Futures trades are often lightly capitalized (high margin), as is forex trading. The assumption is that "large" swings are still fairly small, and it's set up as a commitment to pay the change in amount later, rather than purchasing the full amount directly (or taking out a loan to purchase it).

Sometimes extreme leverage is risky, and that is the point: the house uses it as a tool to structure deceptively high interest rates and win at gambler's ruin phenomena.

One way long-term bets are often conducted is with a pledge to pay in the future. No escrow is used, the bet is entirely uncapitalized. Leverage is functionally infinite in some sense, with reputation on the line and possibly legal backup. Sometimes the capital and escrow requirements are asymmetric, for reasons of asymmetric reputation or trust; in that case, it also has the effect of having the penalty from opportunity cost reduced for the party that would care most. (The interest on the $150k side of the bet matters far more than on the $1k side; you can reframe it as actually being at ~ 120:1 odds or whatever.)

If someone is routinely in the habit of making significant bets with friends, it might well be the case that if they lost all of them at once it would be financial challenge for them; it's entirely possible for this to be reasonable behavior, because the portfolio is diverse and risk is managed.

These scenarios differ on a number of features, including immediate access to the money, counterparty risk, time horizon, available risk management techniques, and impact of a default.

So what should a prediction market bet be capitalized at? What's the proper leverage ratio, in general? Should it be fully capitalized, always? Probably not!

Low capitalization helps make long-term betting attractive. Highly capitalized bets make sense for short-term betting, for risk management reasons. High capitalization makes more sense with low portfolio diversity. High capitalization makes more sense on markets that will definitely resolve to Yes or No vs. an intermediate resolution (PROB). (Relevant recent discord discussion.)