39_nsba4_intro_intel.md

39 — NSBA4 #introductions intel (live-pool self-reports)

Goal. Parse the #introductions Discord export (intros.txt, captured 2026-06-02 — the first real signal on the live NSBA4 pool) into a structured table, reconcile every handle to canonical identity / combine row / game tape, and map the self-reported sandbag/fade claims onto the game-backed buy-low/fade board (finding 19). Begin an eligibility/captain list (Open Q14/Q17/Q18).

Env: .venv/bin/python. Script: scripts/A39_parse_intros.py.
Inputs: intros.txt; canonical_players.csv, combine_player_overall.csv, discord_player_signals.csv, outputs/buylow_board.csv.
Output: data/processed/nsba4_intros.csv (42 rows, one per intro post).

Read this as cheap talk. Per D7/D12 we trust game tape over the combine, and over self-report most of all. An intro is a self-serving signal: everyone wants to be drafted and several are openly joking. Claims are logged, not believed. Where game tape exists it overrides the intro; where it doesn't, an intro is a weak prior only.

Pool size & composition (claim: High confidence — it's a roster, not a model)

42 distinct posters parsed. Of these, 2 are staff/non-draftable who posted (a.new.rag=Anurag, auride0=Mingle "im staff") — leaving ~40 self-declared draftable players who have publicly committed to the pool.
34 of 42 carry an nsba4 combine row (matched via discord_tag, incl. two aliases: Achuth V's combine tag is Giannis, Sriman D's is Sr¡m@η!). 12 have prior NSBA game history (a canonical_id with games on the buy-low board).
2 are genuinely new (no combine row, no canonical match): Dennis Ji (sellmeperson) — physics main, USAPHO distinction, wrote half of nssb physics; and Charith N (dragongam3r23) — physics-only, 1st-year, positive ppg at nats. Both read as real prospects who must be priced off the combine once it posts (no tape).
Grade skew: mostly rising seniors/seniors (12th in 26-27); a handful of returning college players (solvior, dannyridel44, draoethar, suminyy, so.what.if., oof7373) and one rising freshman (didman_5/Sriman D — youngest, growth upside, no tape). Many grades are jokes ("13th/14th/15th grade", "elderly", "grade 13") and are preserved verbatim in grade_joke; normalized best-guess in grade_26_27.
This confirms the Q17 "thin pool" worry is real-but-better-than-feared: the marquee game-backed names (Vishnu, Kian, Rohan G, Akhil, Edwin, Nihar, Roshan, Varyan, Advai, Shannon, Chris W) all introduced themselves → they are in the pool. The historical value is not mostly locked in no-shows after all.

Category self-reporting vs combine (claim: Med — patterns, not measurements)

CS is claimed loudly but thinly, consistent with F5/D21 (the field doesn't study CS). Self-declared CS-capable: Riyan N, Uddip K, Daniel L, Harry G, Daniel Y, Dennis Ji, Justin K, Alex L, Travis D, Eric L (secondary), Aryan B ("ML, 1/4 cs"), Vishnu M ("apparently cs"). Most list CS as a secondary tag behind math/phys — matching F15 (CS is not a standalone archetype; high-CS players cluster with broad elites). No pure-CS specialist with game tape introduced. A credible CS body is still a draft-day fill, not a lever (do not reach — D8/F5).
ESS is the most over-claimed category relative to combine: varnite, abacus14, travis_842, monoionic, caravan._. all main ESS, and three of them attach a "don't-look-at-my-combine" qualifier. ESS specialists are exactly the cluster F17 flags as collapsing under playoff difficulty — treat ESS-only self-mains as regular-season coverage, discount for playoff equity.
Several self-reports invert the combine (the "I'm actually an X main" trope): Sumin Y ("despite my math combine I'm the math goat" — then admits mathcounts alternate), Monish S/kingrel124 ("actually a math main"), Riyan N (math despite combine). These are the cheapest claims on the board; weight near zero without tape.

The sandbag-claim landscape, mapped onto the buy-low/fade board

10 players make an explicit sandbag claim (combine understates me); 21 make a self-fade (downplay / "don't draft me"). Cross-referenced against finding 19:

CONFIRM an existing game-backed buy-low (claim: Med — intel + tape agree)

Chris W (.orca.whale.) — game-backed buy-low (#5, +0.61, 4 games, conf 0.40) and he explicitly sandbag-claims: "better at math and physics than combined shows (I got fried on both ... played with evan and kian)". Tape and self-report point the same way. The low-PPG-because-of-teammate-usage story is plausible but unverified (4 games). Modest, directionally-confirmed buy-low. (He is also Blair SB captain + WADF secretary — see eligibility.)
Rohan G (moeba4) — the cleanest game-backed buy-low on the whole board (#4, +1.00, 16 games, 0.56 career PPTF, conf 0.79) and he self-fades ("im washed but back in my prime I did bio ess"). Self-fade here is bullish: the field will believe "washed," the tape says top-tier producer. Highest-confidence steal in the intro set.

CONFIRM a game-backed FADE via self-fade (claim: Med-High where tape is deep)

Vishnu M (oof7373) — the marquee fade (−1.20, 27 games, conf 1.00) literally tells you not to draft him: "don't waste first round on me im gonna be mad washed." The self-fade agrees with the deepest tape on the board. Don't pay the #4 combine sticker.
Edwin H (draoethar) — modest game-backed fade (−0.38, 11 games, conf 0.66) and his intro is one long self-fade ("my combine is basically a lie ... can't do physics/ess/math"). This corroborates the fade and, usefully, tells you his combine #5 is inflated by bio/chem he can't repeat under difficulty. Still plausibly a fine bio/chem body — treat as "don't pay the top-5 price," not "bad."
Self-fades that agree with smaller game-backed fades: Roshan A (−0.93, "some physics when I don't neg" + hellcat4024 warns "don't draft roshan without a neg-heavy player"), Nihar B (−0.15, "biology and very early chem"), Advai S (−0.66, "idt nsba is gonna be ≥regs DE"), Shannon I (−1.10, + a real availability flag, camp 5/31–6/13).

CONTRADICT a game-backed fade — the one to watch (claim: Med — trust tape, D7)

Varyan J (varnite) is the lone direct conflict: he sandbag-claims ESS ("ESS — don't look at ts combine", + USESO/NAC/4x AIME résumé) but his game tape is a −1.11 fade (6 games, 0.08 PPTF, conf 0.49). Per D7 the tape wins, but conf is only 0.49 (6 games) and his external résumé is real, so this is "don't believe the sandbag hype, but don't hard-fade either" — a genuine ambiguity to resolve with David's read.

Brand-new intel-only sandbag claims (claim: Low — no tape, cheap talk)

No game tape; combine row only or nothing. Log, don't price up: - Harry G (meduh.) — inverse sandbag: combine #2 (36) but says "do not look at combine, I got lucky on bio/chem; ZERO bio chem, I'm math/phys/cs." This is a self-FADE on his inflated categories, i.e. he's telling you his #2 combine is partly biked-luck. Treat the 36 as a soft ceiling, not a floor. - Nithin N (redst.ne) — "better at bio & cs ... somehow negged them"; combine raw −2 (no rank). New, no tape; classic low-info sandbag. - Alex L (abacus14), Travis D (travis_842) — ESS sandbag claims, school SB captains, scioly résumés, bottom-of-pool combines (ranks 48 / 15). No NSBA tape. These are the new "Riyan/Ziang-type" intel buys — motivated (low combine + real external science résumé) but unconfirmed (conf ~0.40 equivalent). - Sriman D (didman_5) — rising freshman, "Chem (bombed combine) > Phys > ...". Youth + growth upside, no tape. Lowest floor, highest variance. - Sumin Y, Monish S — weak/jokey math sandbags (see above); near-zero weight.

The two pre-existing intel buys, now with intro context (claim: Low — still no tape)

Riyan N (nocombomomento) — board #1 buy-low (+2.02, intel-only). Intro reinforces the prior: detailed, credible math/CS résumé (AIME, t10 mmaths/cmimc/smt/pumac, USACO, Codeforces) and an explicit sandbag frame ("despite what the combine says ... I didn't get killed by the combine"). Still no NSBA game tape — conf stays 0.40, but this is the most résumé-backed of the intel buys.
Ziang Z (deasert_willow) — board #2 buy-low (+1.90, intel-only). Intro does NOT re-claim sandbag — he leads with staff/captain status and a solid-but-not-elite résumé (3x AIME, USESO camp, ~36 ppg nats but "didn't make it out of RR"). Two cautions: (1) he is a 3rd-year captain → likely a captain, can't be "bought low" (D18 says captaincy doesn't change value, but he still may be drafting, not drafted — confirm); (2) the self-report is more "good returning officer" than "hidden star," so the +1.90 magnitude remains a placed prior, not evidence.

Eligibility / captain notes (addresses Open Q14 / Q17)

Confirmed in the draftable pool by introducing themselves: all 40 non-staff posters (see CSV in_nsba4_combine + handles). This is the first hard eligibility signal.

Staff / non-draftable (self-identified here): - auride0 (Mingle) — "im staff" (matches the exclude list). - a.new.rag (Anurag) — posted the intro template; head organizer (exclude list). - dannyridel44 (Daniel L) — "helped run nsba3 hehe." Ambiguous: staff and listed as a buy-low (+0.59, intel). Confirm whether he is draftable or organizer-only. - .orca.whale. (Chris W) — "WADF secretary, I will ban opponents when they bike" and deasert_willow (Ziang) — "WADF treasurer, will ban opponents biking." These are officer roles, not necessarily draft-ineligible — both also have combine rows and ask to be drafted. Confirm WADF-officer ≠ non-draftable.

Captain / co-GM hints (self-reported): - Akhil Batchu (abat48) — "I was an NSBA captain last yr ... if drafted first round would like to help make other draft picks." Strong captain/co-GM hint. (D18: doesn't change his value, but matters for whether he's a draft target or a drafter.) - Ziang Z — 3rd-year school captain + WADF treasurer → likely captain (and the marquee intel buy-low — the same conflict flagged in finding 19 caveat #5). - Varyan J — "Captain of BASIS Chandler." Nihar B — cocaptain BASIS Chandler (twin of Kian). Chris W — Blair SB captain. Travis D, Alex L, Nithin N — school SB leaders/captains. School captaincy ≠ NSBA captaincy, but it's the candidate pool.

Availability flags surfaced in intros: - Shannon I (shamp0) — off Discord 5/31–6/13 (camp). Real near-term flag (also a game-backed fade). Jake R — "ran mitosisphere bio bowl ON SCHEDULE" (positive reliability tell). Justin K — "I rarely neg / can't pedal" (honest non-biker, low variance). Praneel A — nocturnal per discord signals (late availability).

Punch-list — what David should confirm

Eligibility edge cases: Is Daniel L (dannyridel44) draftable or staff-only? Are WADF officers (Chris W, Ziang) draft-eligible despite the "I'll ban bikers" role?
Captains: Confirm who is actually captaining (drafting) vs draftable. Candidates: Akhil B (explicit), Ziang Z (3rd-yr capt), Varyan J, Nihar B, Chris W. A captain can't be a buy-low target — this directly affects the board's #2 (Ziang) and #9 (Akhil).
The one conflict — Varyan J: his ESS sandbag claim vs a −1.11 game fade (conf 0.49). Believe the tape, the résumé, or split? Your read settles it.
Two new entrants — price them: Dennis Ji (physics, USAPHO distinction) and Charith N (physics, 1st-yr positive nats ppg) have no tape; confirm their combine rows post (or hold board slots — Q10/D25).
Harry G inverse-sandbag: he says his #2 combine (36) is inflated (lucky bio/chem, real strengths math/phys/cs). Trust his self-fade enough to discount the #2 sticker?
Intel buys still tape-less: Riyan N, Ziang, + new Alex L / Travis D / Nithin N. All rest on résumé + low combine. Confirm any have NSBA (not external) playing history we can pull, else they stay conf ~0.40.

Limitations / caveats

Self-report is the weakest signal we have. Everything here is downstream of game tape (D7) and even the combine. The CSV's sandbag_claim/self_fade_claim are what they said, not what's true.
No new measurement. This finding adds eligibility/identity confirmation and a cheap-talk layer over finding 19 — it does not re-rank the board on its own. The only board-relevant moves are: (a) Rohan G's self-fade makes him a more confident steal, (b) Vishnu/Edwin self-fades corroborate their fades, (c) Varyan's sandbag is the one to adjudicate, (d) two new tape-less prospects to slot.
Handle aliasing is manual (2 confirmed aliases); 2 players remain unmatched (likely truly new). Grades are jokes for many — grade_26_27 is best-guess, grade_joke verbatim.
Combine ranks shown here come from combine_player_overall.csv and may differ by 1–2 from finding 19's board (tie handling); read ordinally, not exactly (per KEY_FINDINGS).

NSBA Draft Analytics · embargoed until after the SSB draft · ← hub