Build My Plan 因材施教
Two honest inputs. Move the sliders and everything below updates — your time allocation, your week, and how long the next few hundred points will actually take.
How the projection works
No black box. Here is the entire model, so you can judge it yourself.
Base rates
Expected annual rating gain by band, from ChessGoals' descriptive dataset (n≈384 self-tracked improvers, median 7 hours per week):
| Rating band | Expected gain / year at 7 hrs/wk |
|---|---|
| Under 800 | +369 |
| 800–1099 | +215 |
| 1100–1699 | +121 |
| 1700–1999 | +37 |
| 2000+ | +18 extrapolated beyond the data |
Scaling for your hours
Returns to extra hours are assumed sub-linear — doubling your hours does not double your rate.
The model uses (your hours ÷ 7)0.7, clamped between 0.3× and 2.0×. That exponent is a
reasoned choice, not a measured constant; nobody has published the real curve.
The spread
The "faster" and "slower quarter" columns are 1.9× and 0.4× the typical rate, derived from the observed distribution of annual change in the same dataset (25th percentile +25, median +110, 75th percentile +218 points per year).
Four things the model cannot know
- The sample is self-selected. These are motivated improvers who volunteered to track themselves for years. Their medians are closer to a ceiling for a typical hobbyist than to an average.
- Individual variation is enormous. The documented range of hours needed to reach master level is 3,016 to 23,608 — an eightfold spread, in a study that controlled for practice hours. Your personal exchange rate is unknowable in advance.
- Starting age matters, and it is not in the model. The same study found the probability of reaching IM was .24 for players who started by age 12 versus .018 for later starters — roughly thirteenfold, after controlling for practice hours. This matters for titles, not for the 1800–2000 range, but it is real and it is why nobody in the data made IM as an adult beginner.
- Quality is assumed, not measured. The model takes hours as input. It cannot tell whether those hours are deliberate practice or comfortable puzzle-clicking, and the difference between those two is larger than the difference between five hours and ten.
What the plan deliberately leaves out
Opening study, below 1200
Set to zero, on purpose. This is the one point of near-universal agreement across every source consulted. One improver studied openings for six months and reported it had "literally no impact on my rating" — his prepared lines occurred in under 0.5% of his games past move eight. Build a repertoire once, then leave it alone.
Watching chess videos
Not a category in the allocation, because it is not training. The rule: if a video makes you feel good about chess, it is entertainment. If it makes you pause and calculate before the answer is revealed, it is training. Watch one structured series once, then stop.
The one adjustment worth making yourself
If your week collapses, cut in this order: passive input first (reading, video), then the extra long game, then puzzle volume — but never to zero. Never cut the post-game review. Playing games you do not review is the single most common way adults spend years without improving. If you only have time for one of the two, review an old game instead of playing a new one.