Dream11 beginner’s path — decision walk-through

This page is not a “how to win Dream11” article. It is a walk-through of the CricJosh decision workbench for people who have never used a probability-weighted fantasy model before. The aim is to make the xFP inputs legible, not to convince you to pick specific players.

What the engine actually does

The xFP engine is a regression over player-by-player innings data. For each player in the probable XI, it produces a point estimate (the expected-fantasy-points — xFP) and an 80% confidence band. The captain overlay is a multiplicative adjustment on top of xFP, not a separate tipster layer. The optimal-XI solver is an integer programme that respects the league’s role and credit constraints.

The five-step walk-through

1. Open the workbench, not the contest list. Before you enter any contest, open the CricJosh decision workbench at /dream11/. Pick the day’s fixture. The four primary surfaces (captain overlay, optimal XI, venue model, win-probability) load in the same column. The workbench is your single source of truth — do not enter a contest before reading it.
2. Read the captain overlay in order. The captain overlay returns a ranked list with confidence bands. The top row is the engine’s best estimate, but the next two rows are plausible substitutes. If the top row has a wide band (say 60–110), the overlay is uncertain — treat rows 2 and 3 as equally serious candidates. If the band is tight (say 90–98), the top row is meaningfully better.
3. Lock the XI under the credit cap. The optimal-XI table respects the 100-credit cap, the role mix (1 WK / 3–6 BAT / 1–4 AR / 3–6 BOWL) and the max-7-from-one-team rule. Read the row-level xFP, not the column totals. The model’s job is to pick the highest expected-value 11; your job is to sanity-check the squad against the venue and the opposition.
4. Apply the venue tilt manually. Once you have the row-level output, apply the venue tilt by hand. Spin-heavy venue (Chennai, Delhi) → down-rank pace all-rounders, up-rank wrist-spinners. Pace-heavy venue (Mohali, Wankhede) → the inverse. The venue model page publishes the wicket-share split so you can do this in 30 seconds.
5. Place the order, log the outcome. Place the team on the platform of your choice. After the match, write the row-level outcome back into your own log. The CricJosh workspace supports an export, but the simplest discipline is a spreadsheet: player, xFP, actual fantasy points, hit / miss. 50 logged fixtures is enough to find your own calibration bias.

The fantasy rules (non-negotiable)

• You must pick exactly 11 players within a 100-credit budget.
• The squad must contain 1 wicket-keeper, 3–6 batters, 1–4 all-rounders and 3–6 bowlers.
• Maximum 7 players from any one real team.
• Captain gets 2×, Vice-Captain gets 1.5×.
• The team locks at the scheduled match start (or toss, depending on platform).

The decision matrix

The same row can mean different things on different days. The matrix below maps common scenarios to the model surface you should trust.

Five anti-patterns to watch for

Every observed mistake in the rolling backtest log falls into one of these five patterns. They are listed in order of how often they cost the most points.

1. Picking the top-rated name without reading the band. A wide band is a warning. Treat rows 2 and 3 as serious candidates.
2. Overriding the engine on sentiment. Sentiment is a low-edge tilt. If you must override, write down why, then check the override against the next 10 fixtures.
3. Stacking 7 players from one team. The constraint is real for variance reasons. The engine respects it because it matters. Use it.
4. Ignoring the playing-XI feed lock. If the playing XI is not yet public, the optimal XI table is a forecast, not a forecast for the actual XI. Lock after Toss.
5. Doubling captain + VC on the same player. Pick the band-tightest captain and the next-tightest VC. Same player = doubled exposure on one match-up.

How to calibrate yourself

The CricJosh engine has a published rolling hit-rate. Yours should be tracked too. After every match, write down the row captain, the row vice-captain, the actual top-2 fantasy scorers and whether the model was right. After 50 fixtures you will have a personal hit-rate that is more useful than any editorial column — because it is calibrated to your application of the model, not the model itself.

What this page is not

• It is not a tipster article. There are no “my picks for today” sections.
• It is not a paid picks funnel. There is no Telegram CTA, no VIP list, no affiliate-locked content.
• It is not a legal guide. The legal status of paid-entry fantasy depends on your state of residence.

Legal note

Paid Dream11 contests are a game of skill in most Indian states under the Public Gambling Act, 1867 and the IT Act, 2000. The states of Andhra Pradesh, Telangana, Assam, Odisha, Sikkim and Nagaland currently restrict paid-entry contests. The CricJosh workbench remains free to use regardless of jurisdiction.