BARDOWNSKI HOCKEY / THE PLAYER LAB
Build your line. Find the chemistry.
Choose a tool. Make your next move.
One workspace. Your call.
Opening the line builder…
01 / THE LINE BUILDER
Pick your players. See their grades. Find a line that clicks.
Pick your last line of defense
0 / 3 skaters selected2026–2027
The best-connected 3s lines, each with its best goalie in net.
69%CONNECTION
New group · built from pair history · fit −1
68%CONNECTION
New group · built from pair history · fit −1
59%CONNECTION
New group · built from pair history · fit −1
Connection is a Bardownski index from 0 to 100, not a win probability. The club's own season average shows as 60. Results are scored as half win share (draws count half) and half goal share, and every point above or below the club's result is worth 2 connection points. Each player starts from their form: the club's results in games they played in that role (skater or goalie), pulled toward the club average over their first five games. Their rating (OVR, or a points and plus-minus rating when OVR is missing) nudges that up or down against the roster average. Every pair in your lineup, including goalie–skater pairs, starts at the average of the two players. Each game they played together moves the pair toward their real results; after 4 shared games, real results and the projection count equally. The lineup's connection is the average of all its pairs. If the whole unit (with the same goalie in net) has played together, its own record is blended in the same way. Each skater slotted away from the positions they actually play loses up to 12 points, scaled by how rarely they play there. Positions are read for the same number of our skaters (a lone skater usually plays center, even if he's a winger beside a teammate); a goalie playing as a skater loses a further 10. Grades: A ≥80, B ≥65, C ≥45, D ≥30, F below 30. Team results depend on opponents and teammates too, so small samples move quickly.
Any two players can be rated, including a goalie and one skater, and the number updates with every player you add, swap or remove. Results are team outcomes, not isolated on-ice stats. Suggestions never use a goalie as a skater, keep the players you have already picked, and place each player in the slot that suits them best.
Each season stands on its own. Your saved line stays on this device and never publishes a team lineup.
OVR is the supplied season member rating, not a reconstructed or necessarily skater-only rating. Its local letter bands are A ≥90, B ≥80, C ≥70, D ≥60, F <60; missing/zero-filled OVR is unrated, not F. When current-season OVR is unavailable, Lab performance uses skater stats only: production = PPG / (PPG + 2); goal balance = clamp(0.5 + plus-minus per game / 4, 0, 1). Raw score = 100 × (0.6 × production + 0.4 × goal balance); Lab rating = round(50 + (raw score − 50) × skater GP / (skater GP + 5)). Lab bands: A ≥80, B ≥65, C ≥45, D ≥30, F <30. These local weights and reference points are not league-calibrated: this is an offense-leaning description, not an all-around skill or goalie grade. Small samples are dampened, not made reliable. Zero skater GP or invalid totals means no Lab grade; no archived stats, missing playstyles or optional zero-filled categories supply a fallback.
Compatibility is a local shared-performance index for one goalie and one skater, not a win probability or a claim of isolated influence. Each pairing uses games where that skater and goalie played together; it does not require the other selected skaters. Raw = 100 × [0.5 × (wins + draws / 2) / games + 0.5 × club goals for / (club goals for + against)]. The percentage is round(50 + (raw − 50) × games / (games + 5)); a goalless sample uses 50% goal share. Grades: A ≥80, B ≥65, C ≥45, D ≥30, otherwise F. Pairings with the same shared results can have the same score. No shared games means not rated, never zero compatibility. Save % is total saves / shots, and GA/game uses recorded goalie goals allowed. Pair samples can overlap and are never summed into a unit record. Every pairing stays within the selected season; other teammates and opponents also affect results.
Every row is one goalie + one skater. Early pairings have fewer than five shared games. Open a pairing’s details beside the goalie for its source results.
Selected-season full regular/finals match records only; no partial archive, private games, forfeits without recorded play or conflicting scores. Exactly 2, 3 or 5 distinct skaters must coappear with one verified OUR goalie in the same selected whole record, with no goalie/skater identity overlap. Conflicting equally selected goalie identities are excluded, not merged. Shared wins/losses/draws and the local getLineRating index describe club outcomes, not goalie influence or a win probability. No shared games means no rating or projection. Save percentage is 100 × total saves / total shots, never a mean of percentages. Missing or inconsistent counts stay unknown; aggregate fields require complete evidence across the shared sample. Recorded goalie goals against per game is NOT GAA: ice time is unavailable and club goals against can include empty-net goals. Profile totals and valid GAA come only from supplied season members; impossible zero GAA is unknown. Observed-only profiles show recorded appearances with unknown season totals. Callers must supply members and matches from one selected season.
Save % uses total saves ÷ shots when available, otherwise the reported season percentage. GA / game is goals allowed per shared game, not time-adjusted GAA. Season GAA is the reported average.
Win % is wins ÷ games. A dash means unavailable, not zero. Only games with every selected skater and this goalie count in ‘With this line’; no shared games means no rating or projection.
Small samples change quickly. List order is not a recommendation, and team results do not isolate a goalie’s contribution.
