Body: Most session-budget advice you see for casual-sessions assumes you are playing heads-up against one opponent with no house cut. That model breaks the moment you sit down in a room where the rake comes out every hand, the game runs short-handed, and two or three regulars are simply better than you on that particular night. This thread is my attempt to write down what I actually do, not what a solver would tell you. It is aimed at people who play live or on club apps where rake is taken from each pot and where session swings are messier than the neat 10–20 buy-in rules suggest. I keep my session-budget separate from daily money. That sounds obvious, but the useful part is that I do not measure it in buy-ins alone. I measure it in “rake-adjusted buy-ins.” If the game is 1/2 with a 5% rake capped at 10, and I expect to win roughly 8–12 big blinds per hour before rake in a good lineup, the rake can remove a meaningful chunk of that. In a slower, passive game, the same cap hurts more because fewer big pots are built. So my first adjustment is to treat a 1/2 game with heavy rake as closer to a 1/3 or 1/4 game for session-budget purposes, even though the blinds have not changed. The second thing I track is not my total profit but my “rake-paid” total. Most apps and community-hub cards show this. If I have paid 300 in rake over a month and I am only up 500, I know my actual edge before rake is thinner than it looks. That changes how aggressively I move up. A player who ignores rake might think they are beating the game comfortably and take shots too early. A player who tracks it might stay at the lower stake longer, which is usually the right call. For actual sizing, I use a floor rather than a single number. For a typical raked 1/2 or 1/3 game, I want at least 30 full buy-ins of 100 big blinds before I consider it my main game. That is 6,000 for 1/2 with a 200 cap. If the game plays deeper, with a 300 or 400 cap, I want 35–40 buy-ins because the variance per hand is larger. For 2/5, I want closer to 40 buy-ins of 100 big blinds, so 20,000, and I do not take a shot until I have 25,000 or more. Those numbers are not magic. They are based on watching my own downswing history: in raked live games, a 10-buy-in downswing over two or three weeks is common even when I am playing well, and a 15-buy-in downswing happens once or twice a year if I play regularly. If my session-budget cannot absorb that without forcing me to move down or play scared, I am under-rolled. A practical rule I use is to move down when my session-budget drops below 20 buy-ins for the stake I am playing. Not when I feel bad, not after one losing session, but when the number hits that line. Moving down early feels like admitting defeat, but it is cheaper than playing with scared money. I have made more money by dropping from 2/5 to 1/3 for a month than by grinding 2/5 with a shrinking roll and passing on thin value spots because I could not afford the loss. Rake also changes which hands I play. In a high-rake game, I fold more small and medium suited connectors from early position, especially when the table is short-handed and the rake cap is hit easily. The reason is simple: those hands make medium-strength hands that win small or medium pots, and the rake takes a bigger slice of exactly those pots. I would rather play fewer hands that can win big pots, like suited aces and pocket pairs, and avoid building pots with hands that rarely cooler someone. This is not a huge adjustment, but over a month it changes how much rake I pay. One thing I do not do is chase losses by moving up in stakes. That sounds obvious, but rake makes it more tempting. If you are stuck three buy-ins at 1/2 and you see a 2/5 game running with what looks like a soft lineup, you might tell yourself one good pot gets you even. The problem is that a 2/5 game with the same rake structure requires a bigger edge to beat, and if you are already tilted or under-rolled, you are more likely to make the exact mistakes that make rake variance worse. I treat any shot at a higher stake as a separate session-budget decision, not an emotional one. Another thing I track is session length. Rake is a per-hand tax, so the more hands you play, the more rake you pay, even if your win rate stays the same. There are days when I leave a good game after two hours because I have already locked up a solid win and the lineup is getting tougher. That is not some mystical discipline. It is just math: if my edge shrinks as the game gets tougher, and the rake stays constant, then at some point I am playing a break-even or losing game just to stay in the room. Leaving then is a session-budget decision, not a tilt decision. The biggest mistake I see in threads like this is treating session-budget as a fixed number you check once. It is not. It is a moving target that changes with the rake, the game size, the depth of stacks, and your own recent play. I re-check mine every two weeks, not every session. That is often enough to catch a slow leak without making me overreact to normal variance. If you are starting out, start lower than you think you need. Play a stake where a 10-buy-in downswing would annoy you but not change your life. Track rake separately from wins and losses. Move down on a rule, not on a feeling. And remember that beating a raked game means beating the rake first and the players second. The players only have to be worse than you by enough to cover the house cut. If they are not, no session-budget size will save you.