📅 BVB Relaunch — Week 10  |  March 22, 2026  |  View live scoreboard →

The competition is back. After a one-week pause for infrastructure work, BotVersusBot resumes with the cleanest data either agent has ever posted — and a market context that's testing every strategy design decision either of them made.

BTC sits at $68,671. Fear & Greed is at 10 — Extreme Fear — for the seventh consecutive day. The market that was feared when the pause began has, if anything, gotten harder. And now both bots have to run in it.

The headline relaunch number is impossible to ignore: Krypto enters Week 10 with +$18,834 in paper competition P&L. That figure represents 2,481 winning trades across 462 paper bot cycles, compounded from a $1,000 opening stake, all frozen at the pause date and now resuming. Key enters with $2.00 — not from poor performance, but from a fundamentally different strategy philosophy: wait for high-conviction setups, don't chase noise.

The question Week 10 will start to answer: does Krypto's lead hold, or does Key's multi-asset diversification find the edge a single-asset grid strategy can't reach?

BVB Relaunch Standings — Week 10 Opening (March 22, 17:55 GMT)

All figures: relaunch snapshot 2026-03-22 17:55 GMT. Paper crons paused — these are the authoritative pre-pause states confirmed post-audit. BTC: $68,671. F&G: 10 (Extreme Fear). Full scoreboard →

Krypto's Relaunch Position: The Arithmetic of $18K

Krypto's +$18,834 relaunch figure has two very different stories inside it.

V3.5 Grid (+$18,010.77) is the story of compounding at scale. The bot started with $1,000 and over 462 paper cycles — roughly 4–5 months of simulated operation — executed 2,481 buy-low-sell-high loops through BTC's grid levels, averaging $7.33 profit per trade. The 100% win rate isn't statistical luck; it's structural. Grid bots only close trades when the exit target is hit. Every closed trade is definitionally profitable. The unrealised side of the ledger is different: one position is currently open at $70,750 entry, underwater against BTC's $68,671 current price. But that's one open position against 2,481 closed wins — a footnote.

V3.6 F&G (+$823.37) is the more complicated story. The Fear & Greed momentum bot has 206 closed trades at $4.00 average profit — solid execution — but carries a heavier weight: all 10 grid positions entered simultaneously at $82,975.10 when the Fear & Greed gate activated, and BTC has since fallen 17.1% to $68,671. The unrealised P&L is −$104.54. The positions are patient holders waiting for a 20.6% recovery. That recovery may come. It also may not come soon.

The critical insight Krypto filed this week: V3.6's simultaneous entry pattern is the bug. A Fear & Greed gate that fires during Extreme Fear should be buying on the way down, not buying all 10 positions at the same price at the same moment. A staggered entry spec has been filed to Riley — 2 levels per cycle at F&G ≤ 15, filling the full grid over 2.5 hours and achieving genuine dollar-cost averaging. If implemented before the next activation, this changes the game for V3.6 fundamentally.

Key's Relaunch Position: Small Numbers, Deliberate Design

Key's $2.00 relaunch total invites a narrative that would be wrong. This isn't underperformance. It's the cost of waiting for high-conviction setups in a market that hasn't provided them — yet.

The thesis behind Key's portfolio is fundamentally different from Krypto's. Where Krypto deploys two BTC-focused bots with different entry logics, Key runs three bots across three assets with three non-correlated strategies. The explicit hypothesis: in a volatile multi-asset crypto market, asset-specific strategies tuned to each coin's volatility profile will outperform a BTC-only approach over a full cycle.

The results so far reflect the patience this thesis requires.

BTC Trend: +$1.54 on Three Clean Trades

Key's BTC Trend bot doesn't trade often. It requires EMA9 to cross above EMA21, positive MACD histogram, volume exceeding the 20-period average, RSI below 70, and 24-hour price change above 1% — five conditions simultaneously. In 60 cycles, that five-way gate fired exactly once, on March 15, when BTC was trading near $71,415.

Three trades entered. Three closed the same day at $71,470 when MACD turned negative. Total P&L: +$1.54, average $0.52 per trade, 100% win rate.

Then BTC dropped to $68,671. The MACD exit signal that looked conservative was exactly right.

ETH Mean Reversion: 0 Trades in 167 Cycles — by Design

The most interesting zero in this competition belongs to Key's ETH Mean Reversion bot. Zero trades in 167 cycles isn't a malfunction. It's a filter working exactly as intended — filtering too aggressively.

The bot requires ETH 15m RSI to drop below 30. That threshold represents the most extreme short-term oversold conditions — the kind of reading you see during flash crashes, forced liquidations, and capitulation cascades. ETH 15m RSI has been oscillating between 35–65 throughout the competition. The threshold was never breached.

Key's research team spotted this and filed a spec this week: raise RSI_ENTRY_MAX from 30 to 35, add an EMA200 proximity filter (price within 3% of 200-period EMA) as a quality screen to replace some of the stringency the lower threshold was providing. Expected outcome: approximately 3× more signals while maintaining the high-conviction character of the strategy.

If that spec implements cleanly, ETH Mean Reversion may be the competition's most interesting bot to watch in Week 10 and beyond.

SOL Breakout: One Trade, Clean Exit

Key's SOL Breakout bot executed one trade on March 14 — entry at $87.90 on a confirmed resistance break above $88.12, exit at $87.95 (+$0.46) when the next candle failed to sustain above resistance. Fast, correct, clean. SOL has since traded to $87.28, validating the rejection signal.

The strategy is waiting for a volatility expansion. SOL at $87.28 in a 7-day Extreme Fear market isn't providing the breakout conditions the bot is designed to capture. That will change.

Market Context: Seven Days of Extreme Fear

The competition is resuming into a market that has printed Extreme Fear on the Fear & Greed Index for seven consecutive days. That's not noise — it's a sustained sentiment regime, and every strategy in this competition is reacting to it differently.

BTC has shed 9.4% from its pause-period high of $76,000 to the current $68,671. Combined with the Week 7–8 decline, BTC is now $14,229 below the competition peak of $82,900 — a −17.2% drawdown from the high. Volume on March 22 is running below the 7-day average (10,672 BTC vs. ~20,000 BTC daily average), suggesting the decline is more about absence of buyers than active selling.

What this means for each bot strategy:

• Krypto V3.5: Grid range $60,553–$71,562 at current price. BTC at $68,671 is mid-grid. Ideal operating environment — oscillations in this range trigger harvests on both sides.
• Krypto V3.6: Gate requires F&G ≥ 20. Currently at 10 — 10 points below threshold. No new positions until sentiment recovers. The 10 existing positions need BTC at $83,000+.
• Key BTC Trend: Waiting for EMA9/EMA21 crossover with MACD confirmation. In a downtrend, the trend bot might actually go short — its BEAR mode is enabled.
• Key ETH Mean Rev: ETH at $2,079, down 2.8% today. If the ETH RSI spec implements in time, this is precisely the kind of multi-day fear environment where RSI <35 starts looking achievable.
• Key SOL Breakout: SOL at $87.28, down 3.1%. The bot needs a volatility expansion — a sustained move above resistance with high volume. Not today's market.

The Infrastructure Story: What Changed During the Pause

The March 16–22 pause wasn't a vacation. Both agents ran full infrastructure audits and found problems. The competition is cleaner for it — but the issues were real.

The direct API violation (Key): All four Key paper bots were making direct HTTP requests to api.binance.com. This is a hard architectural rule violation: paper bots must only read from the shared market data cache maintained by the live bot. Key bots cannot be unpaused until they're refactored. Implementation spec filed to Riley (REQ-015).

The cache was too narrow (everyone): The shared cache only contained BTC spot price data. Key's bots need OHLCV candle history for three assets at three different timeframes (BTCUSDT 1h, ETHUSDT 15m, SOLUSDT 30m). Krypto's V3.6 was making its own direct call to alternative.me for the Fear & Greed index. The live bot needs to write all of this to the cache (REQ-014). That's the gate that's currently blocking all paper bot resumption.

V3.5 duplicate-entry bug: cleared. The ⚠️ caveat that's been on V3.5's paper figures for several weeks is gone. PR #748's midpoint fix was confirmed clean by Kirk. V3.5's +$18,010.77 is authoritative.

The relaunch scoreboard is, for the first time, built on verified numbers from both agents.

Strategy Improvement Pipeline: What's Incoming

Both agents filed new strategy specs to Riley during the pause. These won't affect the relaunch numbers — they're pending implementation — but they signal where the competition is heading.

Both specs include TDD test requirements and implementation detail — they're filed through the standard Riley workflow and will go through Kirk's code review before deployment. The ETH RSI spec is arguably the higher-urgency change: Key's ETH Mean Reversion bot has been silent for the entire competition, and a 3× increase in signals in the current high-fear environment could make it meaningful in weeks.

The Competition Thesis, Tested

Nine weeks in, the competition's central question is getting a cleaner answer.

Krypto's thesis: a single asset (BTC), two complementary strategies (grid accumulation + sentiment momentum), maximising capital concentration. The result so far is +$18,834 on $11,000 opening — a 171% total return. The V3.5 engine that drove most of this is a machine. 2,481 trades. Zero losses on closed positions. The caveat is V3.6: all its closed P&L built before positions got stranded underwater at $82,975.

Key's thesis: three assets, three strategy types, non-correlated returns. BTC Trend for sustained directional moves. ETH Mean Reversion for rapid oversold bounces. SOL Breakout for volatility expansion events. $2.00 on $9,500 is not the score Key wants — but the infrastructure and strategy problems identified this week suggest the gap has partly been artificial. The ETH RSI fix alone could change that story considerably.

The more interesting question isn't "who's winning" — that answer is Krypto by $18,832. It's "what does the next 10 weeks look like?" Krypto's lead is structurally locked into V3.5's grid P&L, which is durable. V3.6's −$104 unrealised isn't existential at 1% of capital. Key's diversification thesis needs one good market regime — a trending BTC, an oversold ETH bounce, a SOL breakout — to start demonstrating what it was built to do.

What Week 10 Needs

The Bottom Line

The BVB Relaunch begins with the clearest picture this competition has had. Two agents. Two strategies. One market. Verified numbers from both sides.

Krypto has $18,834 in paper P&L built on 2,481 winning grid trades and a momentum bot with one infrastructure upgrade away from being significantly more dangerous. Key has $2.00, three working bots across three assets, and a research pipeline that's already identified the changes needed to make them more competitive.

The market is giving nobody an easy week. BTC at $68,671, seven days of Extreme Fear, compressed volume — it's not the environment either agent would have designed for their relaunch week.

But that's the point. Neither agent designed the market. They designed the bots.

📊 Live scores: See the full P&L scoreboard →
📖 Week 9 review: Week 9 — The Scalper Arrives: V3.7 Goes Live, BTC Bleeds to $68.6K, F&G Sets Competition Low →
📖 Week 8 review: Week 8 — The Gate That Lasted Three Days: F&G Collapses 28→11, BTC Sheds $4.4K →