Neil's Numbers: The Mental MWC Toolkit
The full Match Equity Table is roughly 600 cells of precise numerical equity, indexed by both players' away-scores. Holding all 600 values in working memory while also tracking the position on the board, the dice probability distribution, the gammon prices, and the cube context is not how any human plays โ not even the world's strongest.
Neil's Numbers is the mental compression of that table into a few dozen anchor values and a small number of derivation rules. Developed by Neil Kazaross โ the same Kazaross of the Rockwell-Kazaross MET โ and refined over years of his published lectures and clinics, the system gives tournament-level players a way to compute MWC to within 1โ2% in their head, in seconds, at the score combinations that matter most.
This page covers the anchor values, the derivation rules, two worked examples, and the limitations.
For the full table the heuristics approximate, see /mathematics/match-equity/. For the cube-equity formulas that consume MWC values, see the same page.
1. Why a Mental System Is Needed
A tournament match is played under a clock. The standard match clock allocates each player roughly 12 seconds per move plus 2 minutes of reserve time per point of match length. A 9-point match gives each player about 18 minutes of reserve total. Meaningful cube decisions โ the ones where a percentage point of MWC matters โ must therefore be made in single-digit seconds, against full positional analysis, full gammon-and-backgammon-probability estimation, and current-game equity assessment.
The Janowski formula, covered on the match equity page, gives a closed-form approximation:
(the pre-Crawford form โ see the match equity page for the Crawford-specific variant). The formula is useful, but the arithmetic (a division by , then a multiplication by 0.87, then an addition) is not instant under clock pressure. Memorised anchor values โ Neil's Numbers โ are.
2. The Three Canonical Anchors
Kazaross's published "starting set" consists of three away-score combinations whose MWC values are deliberately spaced across the equity surface, so they can be used as reference points for interpolation:
| Score (player / opponent) | Player's MWC | Notes |
|---|---|---|
| 5-away / 1-away | 15% | Maximum-asymmetry pre-Crawford score for short matches. |
| 4-away / 2-away | 33% | The "trailing-by-two-with-cube" benchmark. |
| 2-away / 3-away | 60% | The "leading-by-one-mid-match" benchmark. |
Memorising just these three values is sufficient to anchor most of the in-match equity landscape for matches up to 9 points. The system extends with additional anchors for longer matches.
3. The Extended Anchor Set
Strong players memorise an extended set of approximately fifteen to twenty anchors covering the asymmetric scores from 1- to 5-away. The extended set:
| Score (a / b) | Player a's MWC |
|---|---|
| 1 / 1 (Double Match Point) | 50% |
| 1 / 2 | 70% |
| 1 / 3 | 75% |
| 1 / 4 | 83% |
| 1 / 5 | 85% |
| 2 / 2 | 50% |
| 2 / 3 | 60% |
| 2 / 4 | 68% |
| 2 / 5 | 75% |
| 3 / 3 | 50% |
| 3 / 4 | 58% |
| 3 / 5 | 64% |
| 4 / 4 | 50% |
| 4 / 5 | 57% |
| 4 / 6 | 63% |
| 5 / 5 | 50% |
| 5 / 6 | 56% |
| 5 / 7 | 62% |
The mirror cells (e.g., 2-away/1-away) are simply minus the player's MWC at the reversed score.
These eighteen values together cover essentially every reachable score combination in matches up to 7 points, which is the most-played match length in serious tournament play.
4. The Derivation Rules
For away-scores not in the memorised set, Kazaross's system uses three structural rules to derive intermediate values:
Rule 1: The Symmetry Rule
If the player and opponent are at the same away-score, MWC is 50%. This is a direct consequence of the conservation law derived on the match equity page.
Rule 2: The Trailing Decay
For a fixed leader-away , the trailer's MWC at -away decays approximately geometrically as grows:
The decay coefficient is closer to 0.7 in the early-trailing region (small ) and closer to 0.85 in the deep-trailing region (large ). Memorise the 1-away column values and the rest of the trailing column is reconstructable.
Rule 3: The Cube-Turn Reduction
If the cube has already been turned to 2, the effective post-cube score is shifted by the cube-value. For practical mid-game cube decisions:
\text{Post-double MWC} \approx M(a, b - 2) \text{ if leader takes and loses}$$ This is the rule that connects raw MWC values to actual cube actions, and it is the rule under most stress when a player's "take or drop" decision under clock pressure depends on getting $a - 2$ or $b - 2$ correct. --- ## 5. Worked Example 1: A 7-Point Match Cube Decision **Position:** 7-point match. Current score: 5-3 (you leading). You are 2-away; opponent is 4-away. Opponent doubles you to 2. Your current-game winning probability, estimated from the position, is 28%. **Step 1.** Identify the relevant MET cells. - Take and win: your new score is 0-away vs. 4-away (you win the match outright via the 2-cube). MWC: 100%. - Take and lose: your new score is 2-away vs. 2-away (opponent gains 2 points, going from 4-away to 2-away). MWC: 50%. - Drop: your new score is 2-away vs. 3-away. From Neil's Numbers, $M(2, 3) = 60\%$. **Step 2.** Compute take equity. Expected MWC if you take: $$0.28 \times 100\% + 0.72 \times 50\% = 28 + 36 = 64\%$$ **Step 3.** Compare to drop equity. Drop gives 60%. Take gives 64%. **Take.** The take is correct by approximately 4% match equity. Total mental compute time, with Neil's Numbers memorised: ~5 seconds. Without them, requiring a Janowski calculation or a full MET cell lookup: ~30 seconds or unrealizable. --- ## 6. Worked Example 2: Post-Crawford Free Drop **Position:** 7-point match. Current score: 6-2 (you leading). You are 1-away (Crawford passed); opponent is 5-away. Opponent doubles to 2 on the first roll of the post-Crawford game. **Step 1.** Identify the relevant MET cells. - Take and win: your new score is 1-away vs. 7-away... but the match ends at 7, so you simply win. MWC: 100%. - Take and lose: opponent gains 2 โ opponent at 3-away. Your new score is 1-away vs. 3-away. From Neil's Numbers, $M(1, 3) = 75\%$. - Drop: opponent gains 1 โ opponent at 4-away. Your new score is 1-away vs. 4-away. From Neil's Numbers, $M(1, 4) = 83\%$. **Step 2.** Compute take equity, assuming your current-game winning probability is 50% (first-roll, no specific advantage): $$0.50 \times 100\% + 0.50 \times 75\% = 50 + 37.5 = 87.5\%$$ **Step 3.** Compare to drop equity. Drop gives 83%. Take gives 87.5%. **Take** โ the take wins by ~4.5%. Now consider the same situation but at 1-away vs. **2-away**. The relevant cells: - Take and win: 100% (match over). - Take and lose: 1-away vs. 0-away โ match over, opponent wins. MWC: 0%. - Drop: 1-away vs. 1-away (DMP). MWC: 50%. Take equity (50% game winning probability): $0.50 \times 100\% + 0.50 \times 0\% = 50\%$. Drop equity: 50%. The take and drop are **exactly equal** at first-roll equity. This is the canonical **free drop** window in post-Crawford play. If the position is even slightly worse than 50% for the leader (e.g., the leader rolls a poor opening like 5-1 and the position drops to 47% game equity), the drop is correct. This is the score situation at which strong players watch for free-drop opportunities. --- ## 7. Limitations Neil's Numbers approximate the MET; they do not replace it. Three categories of decision where the heuristics fail: 1. **Gammon-rich positions.** The Neil's Numbers system assumes "normal" gammon prices. In positions with extreme gammon distortion (e.g., late-game blitz positions with 60%+ gammon probability), MWC must be computed with the gammon multiplier built in, and direct MET lookup beats heuristic interpolation. 2. **Long matches.** The memorised anchor set covers matches up to 7 points. For 11-, 13-, 15-point matches the Janowski formula or the MET itself is more reliable. 3. **Cube ownership asymmetry.** When the cube is owned by one player at value 4 or higher, the doubling possibilities open up secondary recube decisions that the Neil's Numbers anchor values don't directly capture. Modern engine analysis (XG, Wildbg) is the only reliable resolution at these positions. For the matches and score combinations that comprise the vast majority of competitive play, however โ 5- and 7-point matches at common score asymmetries โ Neil's Numbers is the standard mental toolkit, and is taught explicitly in most strong-player coaching programmes. --- ## See Also - [Match Equity Tables](/mathematics/match-equity/) โ the full Rockwell-Kazaross MET and Janowski formula. - [Crawford Rule](/rules/crawford/) โ the post-Crawford context for the free-drop examples. - [Mathematics hub](/mathematics/) โ pip counts, race equity, take points. - [Glossary](/glossary/) โ formal definitions for *MWC*, *DMP*, *free drop*, *cube vig*. --- ### Footnotes [^1]: Kazaross, N. *Neil's Numbers โ Match Equity Mental Shortcuts.* Published lectures and clinic notes, US Open and World Championship analysis sessions, 2000sโ2010s. [^2]: Rockwell, T., and Kazaross, N. โ Rockwell-Kazaross MET. Reference table reproduced on [the match equity page](/mathematics/match-equity/). [^3]: Robertie, B. *Advanced Backgammon, Volume II: Match Play* (Gammon Press, 1991), Chapter 3. [^4]: Woolsey, K. *How to Play Tournament Backgammon* (Gammon Press, 1993).