MACD 86% claim (BTC)
Paper bot
Viral claim: 86% win rate.
6 trades recorded · 3 settled
Simulated paper trading only — orders are not sent to Hyperliquid.
MACD 86% claim (BTC) returned +2.1% from its first trade on 28 Oct. Buying the S&P 500 returned −2.3%, measured from its 27 Oct base close. +4.4 pts vs the S&P 500.
All recorded dates · 28 Oct – 16 Nov
1M and 3M unavailable: only 18 days recorded so far.
- +2.1%Lifetime · from $1,000.00
- +4.4 ptsLifetime · S&P base close 27 OctS&P −2.3% from that close
- $1,021.16Paper account value
- 63 settled
- 2.7%Lifetime · worst fall from a peak
- 23Since the paper record began
Claim check
- Claimed
- 86%
- win rate
- Our 2-year backtest
- 38%
- win rate (136 trades, −24% return)
- Live (paper)
- 2/3
- won (67% win rate)
Break-even with these exits: about 40% wins before fees (take profit +3.39%, stop −2.27%).
Realized P&L +$21.06
Too few trades to judge yet; the backtest averaged about 6 trades a month.
In plain words
TL;DR
Viral claim: 86% win rate. Our 2-year rebuild won 38% and lost 24%. Forward test of the exact rule.
How it bets
A viral YouTube video (4.2M views) says one chart rule wins 86% of the time: buy when the MACD line crosses up through its signal line below zero while the price is above its 200-hour average; sell short on the mirror image. Every hour, just after the hourly Bitcoin candle closes, the bot checks the rule on Hyperliquid's BTC chart. On a signal it puts $500 of paper money into the trade at the live order-book price, then waits for +3.39% (take profit) or −2.27% (stop loss). One trade at a time, at most 4 a day. With those exits the rule only needs to win about 40% of trades (plus fees) to break even. Our rebuild on two years of BTC hourly data won 38% of 136 trades and lost 24%. This page counts the live win rate against the 86% claim. Credit: the viral video (the claim) and Golden Algo Trader's rebuild on Medium, which found about 36% on BTC. Hyperliquid funding is booked every hour at the current rate (a long pays a positive rate, a short receives it); our backtest rebuild does not include funding.
Example
BTC is above its 200-hour average and MACD, below zero, crosses up through its signal line → buy $500 at about $84,000; take profit at about $86,850, stop at about $82,090; whichever the price touches first closes the trade.
Why it might work
Trend filters plus momentum crosses are a real, widely used idea, and a 1.5:1 payoff only needs about 40% winners.
Why it might fail
Two independent rebuilds found 36–38% winners, just under break-even, and the 86% figure was never reproduced. Fees and slippage eat the thin margin.
Exact rules
Paper bot on the Hyperliquid BTC perpetual, testing the "MACD + 200 EMA, 86% win rate" rule from a viral YouTube video (4.2M views). Every hour, 5 minutes after the hourly candle closes, it reads Hyperliquid's 1-hour BTC candles (closed candles only) and computes the 200-period EMA of the closes, MACD (12-period EMA minus 26-period EMA) and its signal line (9-period EMA of MACD). Long when MACD crosses above the signal line while both are below zero and the close is above the 200 EMA. Short when MACD crosses below the signal line while both are above zero and the close is below the 200 EMA. It enters at the live Hyperliquid order book (walked best-first, taker fee 4.5 bps): $500 per trade, no leverage, never more than the cash allows, one position at a time, at most 4 entries per UTC day. Take profit at +3.39% and stop loss at −2.27% from the entry price; each hour it checks every closed candle since entry, and a candle that touched a level exits there (the stop first if a candle touched both), with the taker fee and 1 bp slippage. No time exit. The claim: 86% of trades win. With these exits a trade breaks even at about a 40% win rate (2.27 ÷ 5.66) plus fees. Our rebuild on Binance BTCUSDT 1h candles from 2024-10-01 to 2026-09-30 (0.04% fee per side + 1 bp slippage): 136 trades, 38.2% won, −24% compounded, never near 86% in either year (ETH: 38–40%, −16% to −26%). An independent rebuild by Golden Algo Trader (Medium) found about 36% on BTC. This is a forward test of the exact rule, not a recommendation. Hyperliquid funding is booked every hour at the current rate (a long pays a positive rate, a short receives it); our backtest rebuild does not include funding.
Source: medium.com ↗
Bets
Signed amounts show gains and losses. Open amounts are unrealized; settled results retain their recorded status.
3 settled and open bets on this record.
Closed +$16.46
@ $73669.00 · $500.00 + $0.44 fee · exit @ $71178.74
Closed −$11.86
@ $77931.00 · $500.00 + $0.46 fee · exit @ $79708.00
Closed +$16.46
@ $80720.00 · $500.00 + $0.44 fee · exit @ $77991.39
Show raw fills · 6 recordedHide raw fills · 6 recorded
All 6 recorded fills available on this page are shown (up to 100).
HL BTC Close short (take)
$483.10
notional
short at 71178.738 · 0.006787115340238092 shares · $0.22 fee
HL BTC short
$500.00
notional
short at 73669.000 · 0.006787115340238092 shares · $0.23 fee
HL BTC Close short (stop)
$511.40
notional
short at 79708.004 · 0.006415932042447807 shares · $0.23 fee
HL BTC short
$500.00
notional
short at 77931.000 · 0.006415932042447807 shares · $0.23 fee
HL BTC Close short (take)
$483.10
notional
short at 77991.390 · 0.0061942517343904855 shares · $0.22 fee
HL BTC short
$500.00
notional
short at 80720.000 · 0.0061942517343904855 shares · $0.23 fee
Last hourly run
Considered — · Trades 0 · Skipped —
Skips are normal when the market fails the bot's published filter, the price is outside its band, the book is empty, or the full stake cannot fill.