Data-Driven Sailing Home ICSA ISSA
For college coaches evaluating high school recruits

A recruiting guide to reading the ratings

The single hardest problem in sailing recruiting isn't finding fast high schoolers — it's knowing whether a fast sailor in one region is actually comparable to a fast sailor in another. This explains how the model answers that, and why you can trust an ISSA rating enough to compare it directly against a college sailor's.

The problem: college sailing is one graph, high school sailing is seven

ICSA schools race each other constantly and across conference lines — NEISA boats show up at MAISA invites, conference qualifiers feed into shared national championships, and a given season produces thousands of head-to-head links between sailors who'd otherwise never meet. That's what makes a single, unified skill scale possible: enough of the graph is connected, directly or through a few hops, that "who's faster" is a question with a real, comparable answer across the entire division.

High school sailing doesn't have that structure. ISSA is organized into seven regional conferences — MASSA, MISSA, NESSA, NWISA, PCISA, SAISA, SEISA — and the overwhelming majority of racing happens within a conference, not across them. A dominant sailor in PCISA and a dominant sailor in SEISA might never race the same regatta, the same opponent, or even the same boat class in a given season. Their win totals aren't on the same scale — they're not really on any shared scale at all.

This is why raw high school records are such a weak recruiting signal: "won conference" means something different in every conference, and there's no direct data connecting them.

The fix: every incoming freshman is a bridge

The site's ranking model is a single temporal Plackett–Luce fit across both leagues at once — ICSA and ISSA results feed the same model, not two separate ones. Every sailor gets one continuous skill trajectory, θ[sailor, week], that runs from their first tracked race through their most recent one — through high school and, if they continue, straight into college, with no seam between the two.

Sailor identity across that seam isn't guessed from results alone. A separate matching step links a sailor's ISSA and ICSA records into one identity based on name and graduation-year gap (typically 3–5 years between high school and college class year — wide enough to allow a gap year, narrow enough to avoid false matches), backed by a manual override list for edge cases like nicknames or misspellings.

5,340
Sailor identity clusters
3,490
Confirmed ISSA–ICSA bridges
7
ISSA conferences unified

Once a sailor is linked, the model's smoothness prior does the real work. It penalizes a sailor's skill jumping around unrealistically from one week to the next — which means a sailor's known, well-anchored college performance constrains what their high-school-era ratings must have looked like, working backward through their own trajectory. A recruit who becomes a 3.4-rated college sophomore couldn't plausibly have been a wildly different skill level as a high school senior; the model's own regularization enforces that continuity.

ISSA — 7 disconnected conferences NESSA MASSA SEISA PCISA SAISA +MISSA, NWISA a graduating senior's identity bridges their whole conference into the college graph ICSA — one connected network

Each ISSA conference (top) is internally connected but mostly isolated from the others. Individual sailors who continue into ICSA (dashed lines) act as bridges — once one is linked, the smoothness prior anchors their entire prior trajectory, and their former conference, onto the same scale as the college network.

What this means practically

  1. A rated ISSA sailor with a bridged conference is directly comparable to an ICSA sailor. If their conference has produced graduates the model has already linked into college results, their rating sits on the same numeric scale as every school on the ICSA Power Rankings page — not a regional-only approximation.
  2. Confidence varies by how well-bridged a sailor's conference is. A conference with many recent alumni racing in ICSA has more anchoring data pulling its ratings onto the shared scale than one with few. Two sailors with the same numeric rating aren't necessarily equally certain if their conferences differ in bridge density.
  3. The signal gets better every season. Every graduating class adds new bridges. A senior's rating today is better-anchored than the same sailor's rating was as a freshman, simply because more of the graph around them has since connected to college results.
  4. This isn't a replacement for watching sailors race. It's a way to make an otherwise incomparable set of regional stats into one number you can actually rank against, before you spend a travel budget confirming it.

Read the numbers yourself

ICSA Power Rankings ISSA Rankings Full model methodology