Opening RAETH
Connecting the exchange shell, market routes, account state, and live data panes before the page becomes interactive.
Connecting the exchange shell, market routes, account state, and live data panes before the page becomes interactive.
Late in a 60-second window, one side of the binary usually trades in the 90s — the market is "sure." Selling the OTHER side at 5–10¢ is selling the tail: you collect small premiums that pay off almost every window, and pay out the full dollar on the rare violent reversal. A classic short-volatility book in miniature.
With 6 seconds left and BTC $80 above the open, YES might trade 93¢ / NO at 7¢. For NO to win, BTC must drop $80 in those last seconds — possible, but the 7¢ price often overstates it because late buyers of cheap tails (lottery tickets) push the price above fair. Selling that 7¢ (equivalently: buying YES at 93¢) harvests the overpricing. Done every window, you win ~7¢ × N windows and occasionally lose 93¢ — the strategy is profitable exactly when the tail is overpriced relative to its true frequency, which late-window lottery flow tends to produce.
GET /v1/series/btc-up-down-1m/context: the spot-vs-open move and seconds to close.MIN_FAV (its side is "decided").tail_price + 1¢ — let lottery buyers cross to you (zero fee + rebate beats taking).BAIL_FAV), buy the tail back immediately at market — pay 20–30¢ to avoid the 95¢ loss.ENTER_WINDOW = 10 # only act in the last N seconds
MIN_FAV = 90 # favorite must trade >= 90c
BAIL_FAV = 75 # if favorite drops below this, buy the tail back
SIZE = 50 # contracts (max loss ~ SIZE * 95c = $47.50)
every second on the active window:
ctx = series_context()
if ctx.seconds_to_close > ENTER_WINDOW: continue
fav_px = max(best_bid_yes, 100 - best_ask_yes) # favorite's price
tail = "NO" if yes_is_favorite else "YES"
if flat and fav_px >= MIN_FAV and ctx.seconds_to_close > 3:
# SELL the tail as a maker: on the YES book, selling NO == buying YES,
# so rest a YES bid at (100 - tail_ask - 1) ... or directly:
rest SELL on tail side at tail_ask, qty SIZE (POST_ONLY)
if short_tail and fav_px < BAIL_FAV:
# reversal — pay up to cap the loss NOW
buy tail back with IOC at best ask
stand down for this window
at settlement: position expires; log premium kept or loss taken
# Expectancy sketch at 7c premium, SIZE=50:
# win (~93% of armed windows): +$3.50
# bail (~5%): -$7.50 (buy back ~22c)
# full loss (~2%): -$46.50
# EV ≈ +0.32c/contract/window — thin, so costs and the bail
# discipline are the whole strategy.| Parameter | Starting value | Why |
|---|---|---|
| MIN_FAV | 90¢ | Below 90 the window isn't 'decided' — you're not selling a tail, you're picking a side. |
| ENTER_WINDOW | 10s | Earlier entries collect more premium but face far more reversal risk per cent collected. |
| BAIL_FAV | 75¢ | The line where 'noise' becomes 'reversal'. Bailing at 75 costs ~3 wins; not bailing costs ~13. |
| SIZE | 50 | Max loss ≈ SIZE × 95¢. Set it from the loss you can shrug at, not from the win you want. |
Related: Settlement & Resolution