Not Your Keys Applies to Your Casino Balance Too, and Almost Nobody Sizes It That Way

Anyone who has held crypto for more than a year can recite the custody ladder without thinking about it. Hardware wallet at the bottom. Hot wallet above that. Exchange above that. And somewhere near the top sits every balance parked inside a service you signed into with an email address.

The ladder is well understood for trading. It gets quietly forgotten the moment a balance is meant for entertainment instead of investment.

A gambling balance is an exchange balance in different clothes. You send coins, a number appears on a screen, and the keys that control the underlying asset belong to somebody else. That is not a scandal. It is how any custodial service works, and it is the same trade you make when you leave USDT on an exchange overnight to catch a move. What is odd is the behavior gap. People who would never leave their entire stack on an exchange will leave a four figure balance in a casino account for six weeks, because withdrawing feels like admitting the session is finished.

So here are four rules for sizing that balance. All four are arithmetic. None of them is about whether you should gamble.

Rule one: fund the session, not the bankroll

Work out what the session actually costs before deciding what to send.

Say you plan two hundred spins at one dollar on a game published at 96 percent return to player. Turnover is two hundred dollars. The house edge is four percent of turnover. Expected cost of that session: eight dollars.

Eight. The variance around that number is enormous, and you might drop sixty or finish ninety up, but the figure the math is aimed at is eight dollars.

Now look at what people actually deposit. Five hundred is a common opening number. Against an eight dollar expected cost, five hundred is sixty two times more custody exposure than the session needs. You have handed a third party sixty two sessions worth of coins so you can play one.

Nobody does this with an exchange. Nobody moves a year of trading capital onto a hot venue to place one order.

The same arithmetic runs on any game once you know the published figure. A table game at 99 percent turns two hundred dollars of turnover into a two dollar expected cost. A game at 94 percent turns it into twelve. Those numbers are small enough that the deposit size stops being a function of the game at all, and becomes a function of how much you are willing to have sitting somewhere else.

Rule two: withdraw on a schedule, not on a feeling

The reason balances sit is that the decision to withdraw is emotional. You are up, so you might keep going. You are down, so withdrawing the remainder feels like closing the book on a bad night. Either way the coins stay where they are.

Replace the feeling with a rule. Pick a trigger and never argue with it. End of session. End of week. Whenever the balance crosses some multiple of your usual buy in. It does not matter which one you pick, only that the trigger is mechanical.

This is worth testing before you need it. Send a small amount into an account, play nothing, and withdraw it immediately. You will learn the actual round trip time, the actual fee, and whether the payout is automated or waits for a human. Jacks Club publishes its supported networks and processes withdrawals without a manual step for ordinary amounts, which is the behavior you want to confirm for yourself rather than read about. Do the same test anywhere else you hold a balance. The cost of the experiment is one network fee.

Rule three: the fee ratio picks the chain, not the ticker

Withdrawal fees are quoted in absolute terms and should be read as percentages.

A three dollar network fee on a sixty dollar withdrawal is five percent. The same three dollars on six hundred is half a percent. Same fee, two very different transactions.

Put that next to rule one. The expected cost of the session was four percent of turnover. If moving the money off the platform costs five percent of the amount moved, the rail was more expensive than the game. That is a genuinely absurd outcome and it happens constantly, because people withdraw small amounts on expensive chains out of habit.

Two ways out. Move larger amounts less often, or move on a chain where the fee is rounding error. Most services that take deposits on multiple networks let you choose, and the choice is usually worth more than any promotion attached to the deposit.

Rule four: denominate the balance in something that does not move

Here is the one that costs people the most and gets discussed the least.

Leave five hundred dollars of BTC in an account for three weeks. An eight percent move in either direction is forty dollars. The expected cost of the session you funded that balance for was eight. So the price exposure was five times the size of the house edge, and you never chose to take it.

You placed a bet on the game and a much larger bet on the market, and only one of those was deliberate. If the price runs your way this feels like a bonus. It is not a bonus. It is an unhedged position you did not size, did not time, and cannot control.

Stablecoins solve this in the least interesting way possible, which is the point. A balance held in USDC is worth the same tomorrow, so the only variable left is the game. The mechanics differ by network, and the same stablecoin on different chains carries different fees and confirmation times, so paying in USDC is worth reading up on before you assume one deposit is the same as another. If you want the price exposure, take it in your own wallet where you control the exit.

A worked week, all four rules running at once

Monday. You want three sessions this week, two hundred dollars of turnover each. Expected cost at 96 percent is eight dollars a session, so twenty four dollars for the week. You send eighty in a stablecoin on a cheap network, which is roughly three times the expected cost and leaves room for an ordinary bad run.

Wednesday. Down thirty one. Nothing happens, because the trigger is Sunday, not a mood.

Friday. Up nineteen on the week. Still nothing happens.

Sunday. You withdraw the balance, whatever it is. The fee is well under one percent of the amount because you moved it once instead of four times, and the balance was denominated in something that did not drift while it sat there.

Total custody exposure across the week: eighty dollars for six days. Total price exposure: zero. Compare that with five hundred dollars of BTC left in an account for a month, which is the default behavior and which nobody ever writes down as a position.

The test that settles all four

Withdraw something today. Not everything, and not because anything is wrong. Just move a small amount out of every custodial balance you are holding and watch what happens.

You will find out three things. How long it takes, what it costs as a percentage, and whether you flinched at the idea of doing it at all. The third one is the useful answer. If pulling money out of an account feels like an event rather than an errand, the balance is too big.

Custody discipline is not a crypto ideology. It is a habit of asking, for every number on every screen, what would happen if that screen went dark tomorrow. Traders ask it constantly. Apply the same question to the account you play in, and most of the sizing decisions make themselves.