What a big-blind check tells the button
Published · Reviewed
A big-blind check changes the range the button faces only through the hands that check. In a heads-up 100bb single-raised pot, that conditioning step comes before any claim about how often the button should bet.
Fix the situation before reading the check
Consider online six-max No-Limit Hold’em: everyone folds to the button, the button opens to 2.5bb, the small blind folds, and the big blind calls. With no ante or rake deduction, the flop pot is 5.5bb and each active player has 97.5bb remaining. These are accounting assumptions, not a claim that a particular site uses no rake.
Use K♠ 7♦ 2♣ as an illustrative board. The decision is the button’s response after the big blind checks. A solver analysis of that decision still needs both complete preflop ranges, each hand’s weight, and the permitted later actions. The board and positions alone do not determine a betting strategy.
To isolate the meaning of the check, the example below uses a deliberately tiny hypothetical range. It is not Turnstone’s configured preflop range, a complete cash-game model, or an equilibrium recommendation. No solve was run for it.
A suit-pattern average is not the big blind’s defence
SolvedHand’s flop c-bet study, published October 21, 2026, averages the button’s betting after a big-blind check over its fixed-range 100bb game. It covers 1,755 suit-isomorphic board patterns and compares rainbow, two-tone, and monotone groups. Those groups differ in rank and pair composition as well as suits; the study separately compares matched unpaired ranks.
The study gives A♣ T♦ 7♥ at 35.2% button betting and A♣ T♣ 7♣ at 42.0%, a counterexample to treating its aggregate suit trend as a universal rule. Its source frequencies are rounded before aggregation, and its reported numerical residuals concern convergence in that tree, not a confidence interval around every frequency.
These are frequencies for the checked-to button range, not the big blind’s response to a bet. The public study does not supply full per-hand big-blind defence. It therefore cannot tell you which hand to call or raise after your check. The toy ranges below explain this missing conditioning step; they do not reconstruct the study’s fixed preflop ranges or extend its numerical findings.
A small range makes the conditioning visible
Let the button’s toy starting range contain A♠ J♥, A♦ J♣, Q♠ T♠, and Q♥ T♥. Let the big blind’s toy starting range contain K♥ Q♦, 8♠ 8♥, and 6♠ 5♠, with equal weights. All these combinations are legal on the board, and no button combination conflicts with a big-blind combination in this example.
Now suppose the big blind checks every one of its hands. After observing a check, all three combinations remain in the checking range with equal weight. Top pair has not disappeared. The check carried no information about which of these hands the big blind holds because every hand took that action.
Change only the hypothetical checking strategy: K♥ Q♦ always checks, 8♠ 8♥ checks half the time, and 6♠ 5♠ never checks. Before normalising, their checking weights are one, one half, and zero. The remaining range shares are two thirds for K♥ Q♦, one third for 8♠ 8♥, and zero for 6♠ 5♠. In this constructed example, a check increases the share of top pair.
These invented strategies demonstrate conditional weighting. They say nothing about whether those hands should take those actions. The general relationship between starting ranges and subsequent action probabilities is described in PioSOLVER’s range and strategy FAQ. In a complete range, opponent-card removal also matters; the disjoint toy combinations keep that complication out of this illustration.
A check is not evidence of a capped range by itself
Calling a range “capped” asserts that sufficiently strong hands are absent. Observing a check cannot establish that without a reason to think those strong hands would have taken another action. In the first toy policy every strong hand available to the big blind checks; in the second, checking selects toward the stronger hand.
Inspect the combinations that remain after multiplying the starting weights by their checking probabilities. That range composition explains what the button is responding to; a board-level bet percentage does not.
A real analysis must also include the hands omitted from the toy ranges, the big blind’s other actions, and the button’s subsequent strategy. There is no recommendation here to bet every hand, to check every hand, or to choose a particular size. Those choices require a specified game and a calculation appropriate to it.
Keep the available bet sizes attached
If an analysis offers the button one size, its reported bet frequency refers to that option against checking. Adding another size changes the question and can move hands between actions. The first result is not a direct answer for the expanded menu.
The same qualification applies to later streets. An attractive flop action depends on what both players can do after a call or raise and after the next card. A restricted turn tree is part of the flop model even when you are viewing only the first decision. Our solver-interpretation article explains why a small convergence target cannot correct a mismatched tree.
Separate one decision from a board average
A study averaging the button’s bets over many boards answers a broader question. It does not identify the big blind’s checking range on this particular board, or prescribe a frequency for this toy game. SolvedHand’s original flop c-bet study reports a separate configured dataset and its weighting assumptions; this article does not reproduce or extend those results.
Turnstone and SolvedHand have the same owner. The link supplies related research context, not independent validation. For a specific hand, retain the exact ranges, action sequence, board, and menu with any conclusion. The big blind’s check is one event in that model, and its meaning comes from the policy that produced it.