# Live vs Online Hold'em > Live tells vs online timing. Eight-handed deep vs 6-max zoom. Rake is a tax, straddles change SPR, and a HUD is a rumor until the sample is real. Canonical: https://unmatchedpoker.com/strategy/live-vs-online Publisher: Unmatched Poker — https://unmatchedpoker.com/ Last reviewed: August 31, 2026 Note: Educational Hold'em magazine. Not a poker room, not financial advice, not a signup funnel. 18+/21+. --- Strategy · Formats # Same cards, different tax, different tells Live poker is eight-handed, slow, and full of people who will tell you their life story while they overcall a river. Online poker is six-max or zoom, timed, and full of people who have a HUD where a face would be. The cards are the same. The economics are not. If you import your zoom continuation-bet frequency into a 2/5 straddle pot with 280 big blinds, you are the tourist. If you import your live "I can fold river because he sighed" into a 6-max pool, you are the tourist the other way. Last reviewed August 31, 2026 Hold'em · Live · Online This is not a ranking. Live is not "real poker." Online is not "the future" in the way a 2005 banner meant it. They are two games that share a deck and split on speed, rake, stack depth, information, and how many people see a flop. You should be able to beat both or have the humility to specialize. Pretending they are the same because Hold'em is Hold'em is how online crushers donate a buy-in an hour to a live nit who only bets a river with it, and how live regs get run over in zoom because they cannot fold top pair to a delayed barrel. The floor is still a [GTO baseline plus a sample-size check](/strategy/gto-vs-exploitative). The exploit just uses different evidence. Live: timing, posture, who looks at the board twice, who already started stacking chips for a call they should not make. Online: timestamps, HUD counts, tabling patterns, the player who tanks every river they are going to fold. Neither oracle is clean. Both beat "I had a feeling" if you treat them as data instead of mysticism. ## The tax: rake as economics, not a coupon Rake is the house's cut. It is a tax on pots that go to showdown and, in some rooms, on pots that do not. It is not a loyalty program. We are not going to sell you a way around it. We are going to price it, because if you do not, you will think you are a 6 bb/100 winner in a game that is charging you 12. Live cash typically takes a drop: a fixed amount from the pot, sometimes plus a jackpot drop, often capped. Small pots get murdered. A 1/2 pot that limps around and ends on a flop bet of 15 can still lose 4–6 plus jackpot. That is not 5%. That is a mugging. Big pots get a better percentage because the cap already hit. This is why live grinders care about playing pots that can actually get large, and why "I win a lot of small pots" is not the flex it sounds like. You can win a lot of small pots and still be the drop's favorite customer. Online cash typically takes a percentage with a cap, dealt or contributed, at a frequency you will feel in your graph. Six-max zoom at small stakes is a rake furnace: many hands, many capped pots, many showdowns. The same 4 bb/100 skill edge that prints in a juicy live 2/5 can be break-even or worse in a zoom pool once the tax is paid. That is not a conspiracy. That is arithmetic. If the game is too expensive for the edge you actually have, you do not wait for a promotion. You change stakes, change rooms, change to a table with more mistakes, or you do not play. Table selection is how you pay less tax per mistake you extract. Two taxes, same "winner" Player true skill: +8 bb/100 before rake in a recreational lineup. That is a real winner. Live 1/2, 30 hands/hour, drop that works out to about 3 bb/100 in the games they choose (deep, capped, not a limpfest every hand). After tax: +5 bb/100. At 30 hands/hour that is +1.5 bb/hour. Times 2-dollar blinds: a modest hourly. Fine, if the lineup stays recreational and they do not spew the straddle pots. Online 50nl 6-max zoom, 80 hands/100 on one table, effective rake+jackpot in the 10–14 bb/100 neighborhood depending on room and cap. After tax: 8 − 12 = −4 bb/100. The "winner" is now a donor. They did not get worse at Hold'em. They brought an 8 bb/100 edge to a 12 bb/100 toll booth. Same human, opposite graph. If your response is "I'll make it up in volume," you just volunteered to pay the toll more often. Volume multiplies the after-rake number. It does not bless the before-rake number. Pick games where the after-rake number is green. That is the whole economics lesson. No cashier required. Dealt rake versus contributed rake changes who pays. Dealt: every seat pays for existing. Contributed: the people in the pot pay. Neither is a moral system. If you are table-selecting online, a dealt-rake room punishes sitting in a dead game. Get up. If you are live, the drop punishes tiny pots. Stop celebrating your 18-dollar wins as if they were a career. ## Eight-handed live versus 6-max zoom [Six-max table with UTG, hijack, cutoff, button, small blind, big blind] A 6-max online table has UTG, hijack, cutoff, button, small blind, big blind. Live "full ring" in 2026 is often 8 or 9, sometimes 7 when the room is lazy, sometimes 6 when they are spreading a shorter game. Extra seats mean extra people who can wake up with a real hand behind an under-the-gun open. Your UTG range live is not your UTG range in 6-max zoom. If you are opening King-Jack offsuit under the gun at a nine-handed 1/3, you are the reason the cutoff is smiling. Multiway pots are the live default. Limps happen. Overlimps happen. A "single-raised pot, heads-up, in position" file from your trainer is a guest, not the mayor. You need a limp-isolation size that actually isolates, a limp-behind range that is not "any suited trash," and a willingness to check flop as the preflop raiser when three people came along. Online 6-max, especially zoom, is still mostly heads-up after the flop. Continuation-bet frequencies can be high. Live, a 70% c-bet into three recs on 9♠ 8♣ 4♦ is how you build a pot to lose. Zoom specifically deletes table dynamics. You do not have four orbits to figure out that seat 4 overfolds the river. You have a HUD and a population. Population reads are real. They are also not the guy who just sat. Play the pool until the player gives you a reason. Live is the opposite: the pool is wide, the sample on this human might be two hours, and the two hours can be worth more than 400 zoom hands of HUD colors if you are actually paying attention. ## Stack depth is the hidden format Online cash is often 100bb effective, sometimes 200bb at the cap, sometimes 40bb in a short-table or a rec who just sat. Live cash is routinely 150–400bb because people buy in deep, win, and do not leave. SPR (stack-to-pot ratio) after a single raise can be 12–20 live and 8–12 online. That extra SPR is why [suited connectors](/strategy/suited-connectors) and speculative pairs change value. It is also why a live top-pair-top-kicker stack-off on a wet board is a bigger crime. You had room to get away. You did not use it. Deep live: more implied odds, more reverse implied odds, more room for a fourth street of credit. You can peel a flop with 76s in position against a single raiser who will pay a set. You cannot peel three limpers out of the big blind with 76s and 90bb because someone will raise behind and you will have created a mess for a hand that needed position and a customer. Short online: push-fold appears earlier, 3-bet pots get shallower, and "I'll see a flop cheap" is a lie you tell before you call a 3.5x open out of the big blind with 54s and 70bb. Different article, same arrogance. ## The straddle is not spice. It is a new big blind. A live UTG straddle turns 2/5 into 2/5/10 with the last live action on the straddler. SPR drops. The button is no longer the button in the same way. Blind-on-blind wars get replaced by button-versus-straddle wars. If three people straddle because the table is "gambling," you are now in a 2/5/10/20 game with a 200bb buy-in that is actually 50bb effective. Your 100bb trainer file has left the building. Playable: a single optional straddle in an otherwise deep game, if you widen the penalty on the straddler's leftover range and you still have SPR to work with. Not playable: mandatory straddles in a room already taking a fat drop, unless the table is a clown car of mistakes large enough to pay both taxes. Do the after-rake math. "It keeps the game good" is sometimes true and sometimes the recs saying it are the only reason you stayed. If the recs leave and the straddles stay, you leave too. ## Tells: live bodies, online clocks Live tells are real and overrated in the same sentence. A shaking hand is more often adrenaline than a bluff, especially in a 1/3 player who just flopped a set for the first time this year. A long stare at the board can be a draw doing math, or a set pretending to do math, or a person who cannot remember if the flush got there. You need a baseline for this human, not a TV rule. Useful live tells, when you have a baseline: - The player who usually chats and goes quiet. Silence is a deviation. Ask what they deviate toward — for some people it is the nuts, for some it is a bluff they are trying not to leak. - Chip handling: the rec who already has calling chips in a pyramid before you finish betting is not folding. Stop bluffing. Start value-betting thinner. That is the same station exploit as online with a better camera angle. - The snap-check versus the tank-check on a brick river. Snap-check is often air or a give-up pair. Tank-check is often a medium hand trying to look dangerous. Population-dependent. Take notes in English, not in mysticism. Online timing is the cousin. Instant river check is not the same as a 12-second river check. Instant fold to a 3-bet is not the same as a 8-second fold. Instant jam is often the poles — the nuts or the air they already decided on flop. A tank-then-small-bet is often a polarized attempt to look like a value size. Use it. Do not marry it. Bots and clients and lag exist. A HUD that says "fold to flop cbet 71% (14)" is still a rumor. A HUD that says the same at (380) is a person. The leak is searching for a tell so you can skip range construction. Range first, tell as a tie-break. If your live game is 80% "he looked left," you are going to call off Queens versus the nit's river raise because he looked left when he had it, too. If your online game is 80% HUD color, you are going to 4-bet a 28/22 over 60 hands and run into Kings. Sample size is the religion we actually allow. See [GTO versus exploitative](/strategy/gto-vs-exploitative) for the sermon, and [pot odds](/tools/pot-odds) for the arithmetic when the tell is cute and the price is not. ## Street by street: the same hand, two rooms Take one holding to both environments. You have Jacks in the cutoff. Live 2/5, eight-handed, effective 280bb because the big blind is a winning rec who buys deep. Online 100nl 6-max zoom, 100bb. Same cards. Different tree. [Hero's hole cards] Live preflop — 2/5, 8-handed, 280bb, UTG straddled to 10 UTG straddles. Two limps. You raise to 45 with Jacks, which is a size aimed at the straddler and the limpers, not a 2.5x you copied from zoom. Button folds. Big blind — deep rec — calls. Straddler calls. Limpers fold. Three ways, 160 in the middle, 1,350 behind. SPR about 8.5. Playable. Not a 100bb SPR 13 single-raised pot. Do not pretend it is. Live flop [Th 7c 2s] Checked to you. You bet 90. BB calls, a little too fast, chips already in his hand. Straddler folds. Fast call from this rec on a dry ten is often a ten, a seven, or a pair of nines they cannot dump. It is rarely a draw they are disguising, because this human does not disguise. You have a tell. You also have an overpair that is not the nuts deep. You do not need to stack off on a later street to prove a point. Live turn [Th 7c 2s Td] He checks after ten seconds, looks at your stack, looks at the board again. That extra look is not nothing — the second ten changed the hand he had, or the hand he is representing. You bet 150, a size that keeps 7x and 99 and underpairs, and does not look like a panicked third barrel in waiting. He calls, slower. You are done firing a river unless the river is a Jack. If he check-raises this turn, you have a very sad fold more often than a call. Deep live, Jacks are allowed to lose to Tens without a speech. Online version — 100nl zoom, 100bb, no straddle You open 2.2bb in CO. BB calls. Heads-up, pot 5bb, 97 behind. Flop T72 rainbow, he checks, you c-bet 1.6bb because this is a high-frequency c-bet texture in a heads-up SRP and the pool folds enough. He calls. Turn Td. He checks. You can barrel a chunk of your range, including Jacks as thin value, because his zoom calling range is wider and weaker than the live rec's, and the SPR is about to force a river decision anyway. River brick. He checks. You can value-bet small. You cannot overfold this river to a jam in zoom the way you can overfold live to the nit's check-raise. The pool jams missed clubs and stubborn sevens. The live nit check-raises Tens. Same Jacks. Different tax, different opponent, different river. If you played the live river like zoom, you get stacked by Tens and you will say you ran bad. If you played the zoom river like live, you check back Jacks and you will say the pool is unbluffable. You just used the wrong file. ## Leak: importing the wrong format's courage Online player live: c-bets every flop, including multiway, then bluffs a river because the HUD in their head says fold-to-river 54%. There is no HUD. There is a rec who did not come to fold. Live player online: calls three streets with second pair because "he would have bet bigger with it," ignoring that the pool's small river bet is value-heavy at this cap. Related leak: sitting in a zoom game you do not beat because you beat a live 1/3, or sitting in a live straddle game you do not beat because you beat 50nl. Edge is format-specific. The [bankroll page](/strategy/bankroll-management) will not save you from a tax you refused to price. ## Deviate: overfold the live nit's river, overcall the zoom nit's river? No. Look. Live nits who bet big on rivers after two checks are a folding station for your bluffs and a folding station for your hero calls. Give them the pot with second pair. Pay them with the top of your range. Zoom nits are different: they under-bluff some rivers and under-value-bet others. Population reports at your stake tell you which. A useful live-only deviation: isolate limpers larger and in position more than GTO 6-max would, because the leftover ranges are trash and they will overcall flop. A useful online-only deviation: delayed barrels against the pool's overfold on turn after they peeled flop "to see." Write the deviation in one line. If the line starts with "I can just tell," it is not a deviation. It is a memoir. ## Dealer and button etiquette, without the scold Protect your cards. Act in turn. Do not touch another person's chips. String-bets will be called back; if you want to raise, say raise or put out the raise. The dealer is not your employee and not your enemy. If you need the pot counted, ask before you tank-call. If you muck face-down, you muck. None of this makes you a winner. All of it keeps you from being the delay, the argument, or the person who voided a winning hand because it was not protected when the deck ate it. Online equivalent: do not stall every river to manufacture a tell you do not have. Do not table-talk in chat about a hand in progress if the room forbids it. Do not sit out in a way that griefs a cash table you are losing to. The goal is not to be liked. The goal is to not add extra ways to lose, extra ways to get a floor called, or extra reasons for a good rec to leave. Recs leaving is a rake problem for you. Behave like someone who understands the tax. One live habit worth importing online: look at the people. Online that means looking at the HUD count, the timing, the showdown they went to two tables ago. One online habit worth importing live: think in ranges even when a face is doing theater. Theater is optional. Ranges are not. ## Table selection, both rooms Live: walk. If the 2/5 has four regs, a dealer you have never seen because the table just opened, and a rec who is already stretching, wait for a better seat. If the 1/3 has two empty seats and a loud table of mistakes, take it even if the blinds are "smaller than your game." Hourly lives in mistakes per orbit, not in the sign. Online: the lobby is a market. A 6-max table with two weak VPIP/PFR combinations and a fish icon you actually trust is worth more than a full zoom pool at the same stake. If the software makes table selection impossible (zoom), the selection is the stake and the room's tax, not the seat. Play a lower stake with a worse tax than a higher stake with a better pool? Sometimes. Do the after-rake math, not the ego math. Bankroll still applies. A great live game at a stake that is 8 buy-ins of your roll is a great way to go home early. Two bullets as a written shot, then back. See [bankroll management](/strategy/bankroll-management). If you cannot leave a good table because leaving would mean feeling the night, that is not table selection anymore. That is the other page. ## What to study for each Online: solver sessions on heads-up SRP and 3-bet pots at 100bb, plus a HUD discipline that refuses to believe 80-hand samples. Live: multiway c-bet lines, isolate sizes, and a notebook that records people in sentences. Both: [pot odds](/tools/pot-odds) until the break-even percent is automatic, so a tell does not talk you into a 22% call you needed 31% for. Play the room you are in. Price the tax. Use the information that room actually gives you. That is the conversion course. Everything else is people trying to make their A-game portable without doing the work. Last reviewed August 31, 2026.