Understanding variance and downswings

Variance is the spread of results around your true win rate. If your true win rate is 4 bb per 100 hands, you will not win exactly 4 bb every 100 hands. Sometimes you win a lot. Sometimes you lose a lot. Over enough hands the average approaches your true EV, but the path is bumpy.

• Skill sets the average EV or win rate.
• Variance sets the swing the up and down around that average.
• Bankroll management exists to absorb that swing.

Standard deviation measures how wide your results can swing. In online NLHE cash games a common rough range is 70 to 110 bb per 100 hands. Games with more multiway pots, larger sizes, or looser opponents can swing even more.

• Standard error of win rate ≈ SD per 100 ÷ sqrt(sample in hundreds).
• Example SD 90 bb per 100 and 50k hands gives sqrt(500) ≈ 22.36. Standard error ≈ 90 ÷ 22.36 ≈ 4.0 bb per 100. Your reported win rate at 50k hands can easily be 4 bb above or below true skill just from noise.
• MTTs have much higher variance because payouts are top heavy. Downswings of hundreds of buy ins are possible even for winning players.

A downswing is a sustained period where results are below expectation. You can play well and still lose for long stretches due to card distribution, coolers, setups, and opponents hitting draws.

• Cash games a normal downswing can be 10 to 30 buy ins. Larger downswings happen in aggressive or high rake environments.
• MTTs standard is far larger. Good players can go 100 to 300 buy ins down and still be within normal variance, especially in large fields.
• PLO swings harder than NLHE. Expect bigger amplitude and plan a bigger bankroll.

Do not judge your whole strategy by one month. Zoom out to larger samples before making big changes.

All-in EV replaces the result of an all-in with the equity at the time the money went in. It helps separate luck from decision quality.

• Good use if your results are far below all-in EV, you are likely running cold in all-in pots.
• Limits all-in EV ignores non all-in variance such as missing flops, coolers without all-ins, or runouts after non all-in bets.
• Conclusion track all-in EV, but still review hands for leaks. EV lines can run hot or cold too.

• Cash games treat 50k hands as noisy and 150k to 300k hands as a stronger signal of true win rate.
• MTTs think in tens of thousands of entries for very stable ROI. Use ABI, ITM, deepest run percentile, and final tables to triangulate progress.
• Live volume is slower. You need hundreds of hours before drawing conclusions. Track in bb not money to reduce bias.

Risk of ruin is the chance your bankroll hits zero before your edge wins long term. You do not need complex math at the table. Use simple control levers.

• Use a bankroll sized for your game. Online NLHE cash often 50 to 80 buy ins. Live often 30 to 50. MTTs 300 to 500 for medium fields and more for large fields.
• Table select to reduce variance. One sticky caller is great. Four maniacs in one pot increases swing.
• Move down stakes quickly when you hit loss thresholds set in your bankroll plan.

Not every downswing is bad luck. Use data and review to separate variance from leaks.

• Compare to EV if result is far below all-in EV and your lines look solid, variance is a likely driver.
• Filter hands review losing pots by line. For example facing river bets of 75 to 100 percent pot. Check call frequency and blockers.
• Population checks if your pool under bluffs big rivers, tighten bluff catching at those sizes. That is not variance. That is an adjustment.
• Consistency if the same spot costs you repeatedly, design a drill and fix it regardless of short term luck.

• Stake guardrails move down when your bankroll drops below your plan threshold. Winning the ego battle is part of winning poker.
• Volume control shorten sessions and reduce table count until decision quality feels sharp again.
• Exploit focus lean value heavy versus callers and trim pure bluffs. Choose sizes that opponents will call with worse hands.
• Study ratio increase study to play ratio. For example 40 percent study and 60 percent play until the slide stabilizes.
• Game select harder prioritize soft tables and position on the most active player. Avoid reg wars while you recover.
• Process goals track only bb, decision quality, and adherence to plan. Hide the cashier for a while.

• Tilt alarms revenge calls, speeding up, ignoring notes, and chasing losses are stop signals. End the session or take a break.
• Pre session checklist energy, focus, and mood scores. If the score is low, study instead of playing.
• Post session reset log result in bb, write two things done well and one improvement for tomorrow.
• Life hygiene sleep, food, hydration, and exercise reduce emotional volatility as much as strategy work increases EV.

• Expect 10 to 30 buy in downswings in NLHE cash even if you play well.
• Expect 100 to 300 buy in downswings in MTTs and more in huge fields.
• Use small flop sizes on dry boards and avoid low equity bluffs in multiway pots to lower variance without losing EV.
• Track all-in EV and big river spots. Those two views explain a large share of perception of luck.
• Biggest wins come from value betting well and folding early when the story is strong against you.

• Myth: a bad month means you are a losing player. Reality: a bad month is a data point. Check larger samples.
• Myth: all-in EV tells the full truth. Reality: it helps but misses non all-in luck and card distribution.
• Myth: playing tighter always fixes variance. Reality: it can lower swing but may also cut win rate. Optimize, do not overreact.

• Variance is the swing around true EV. Bankroll exists to absorb it.
• Cash SD often 70 to 110 bb per 100. Expect noisy results at 50k hands.
• Downswings happen even to winners. Plan for them with stake guardrails.
• Use all-in EV to diagnose luck but still review hands for leaks.
• Increase study and value focus during downswings. Game select and move down fast when needed.
• Protect mindset with stop rules, breaks, and simple routines.

Trust the process. With a strong bankroll plan, disciplined session habits, and targeted study, variance becomes a storm you can sail through instead of a reason to quit.