SN88 Picks — Week 1 Research Brief (Sample)

⚠️ NOT FINANCIAL ADVICE — STANDARD DISCLAIMER

SubnetAIQ Picks is a data-driven signal and research tool, not investment advice. Bittensor subnet alpha tokens are highly speculative and can lose 100% of their value. Past performance (including our 76% backtest win rate across 88 historical picks) does not guarantee future results. We run this portfolio with our own money to demonstrate methodology — your risk tolerance, tax situation, and financial goals are different from ours.

Do not invest more than you can afford to lose completely. Consult a licensed financial advisor before making any investment decisions. We are not registered investment advisors, broker-dealers, or financial professionals. Cryptocurrency involves substantial risk including total loss of capital.

By reading this brief you acknowledge you understand these risks and take full responsibility for any decisions based on information provided.

Publishing this week's full brief as a free preview. Starting next Sunday these go to paid subscribers only — $9.99/mo at signals.subnetaiq.io. This is what you get.

Regime: FAVORABLE

Translation: SN1 Apex is flat, the broader alpha market is mildly negative. In that setup, capital rotates out of bluechips looking for yield — exactly the environment where small-subnet accumulation strategies work. If SN1 broke above +5% 30d, we'd move to cash. It didn't.

The 8-pick portfolio is live. Equal-weighted at 11.88% each (≤ 5% dust float).

The picks

SN31 — Halftime (14d price: +11.9%)

Signal: tao_in 1,100 (bottom 25% threshold), price 0.00485τ (bottom 75% threshold), emission 162 (top 30%), 14d price change +11.9% (under the +30% ceiling).

What it is: A subnet in the sports/media attention space per public registry metadata. "Halftime" framing suggests real-time sports analytics or fan-engagement inference.

Why it qualified this week: tao_in rank is bottom quartile meaning capital hasn't flooded in yet. Emission rank is top-third meaning the subnet IS actively paying miners — not dying. Price has moved +11.9% in 14d but stayed under the overheating threshold. This is the "already warming, not yet pumping" profile the algorithm is tuned for.

What kills the thesis: If the 14d price change crosses +30%, we exit (rule-based, no discretion). Also if Halftime's narrative cools — sports attention-related subnets move in cycles, and if Big Sports Subnet gets outflows, SN31 follows.

Exit watch: 14d change > +30% → out. Emission drop below median → out (narrative breakdown).

SN76 — Byzantium (14d price: −16.3%)

Signal: tao_in 732 (deep bottom), price 0.00583τ, emission 117 (median+), 14d price change −16.3% (reverting down).

What it is: Byzantium in crypto-context typically refers to consensus or Byzantine fault tolerance research — likely a subnet around distributed consensus mechanics or BFT-related model benchmarking.

Why it qualified this week: This is a classic deep-value pick. tao_in at 732 is bottom 10% of the universe. The −16.3% 14d drop means capital's been sellings, which usually signals either distress or forgotten. Emission at 117 (above median) says the validator distribution is still working — so it's not dead, it's discounted. The algorithm loves this setup: real activity meets a depressed price.

What kills the thesis: If emission drops below 50 (true death), exit. If the −16.3% turns into a −30%+ drawdown over the next week, reassess — this could be the start of a slide, not a bottom.

Highest conviction pick of the 8 based on the alpha-buying-at-a-discount framing — but also the most speculative.

SN87 — Luminar Network (14d price: −5.9%)

Signal: tao_in 883, price 0.00393τ (cheapest in the portfolio), emission 158 (top-third), 14d price −5.9% (mildly down).

What it is: "Luminar Network" hints at computer vision / lidar-style sensor subnet — vehicle telematics, autonomous driving data, or geospatial inference. Real commercial use cases.

Why it qualified this week: Cheapest alpha in the portfolio at 0.00393τ/α. Small tao_in means any capital inflow moves price meaningfully. Mild 14d drop (−5.9%) means it's not panicking, just drifting. Emission above the 150 mark says validators are actively rewarding miners, so there's real work being scored.

What kills the thesis: Vision-space subnets have strong competitors (Zeus, others). If the big vision play consolidates, Luminar could lose validator weight. Watch for emission collapse.

Exit watch: emission < 100 for 2 consecutive days → rotate out.

SN94 — Bitsota (14d price: −14.6%)

Signal: tao_in 1,496, price 0.00363τ, emission 167 (top-third), 14d price −14.6%.

What it is: A newer subnet without strong public narrative — that's actually a feature, not a bug. Obscurity is part of the signal — means other rotation strategies haven't found it yet.

Why it qualified this week: Combo of small tao_in + high emission + depressed price. This is the "unknown but productive" slot. The algorithm doesn't care about social consensus; it cares about the four filters. Bitsota hits all four cleanly.

What kills the thesis: Newer subnets have higher variance. If the team behind SN94 disappears or the validator reduces emission weight, this drops fast. Expect ±20% weekly volatility.

Exit watch: If tao_in drops below 500 (liquidity collapse) OR emission drops below 75 → exit.

SN108 — TalkHead (14d price: −5.8%)

Signal: tao_in 1,661, price 0.00410τ, emission 176 (top-20%), 14d price −5.8%.

What it is: Name suggests speech synthesis / voice cloning / talking-head video generation. Real product category — competing with commercial AI voice services.

Why it qualified this week: Highest emission rank in the portfolio at 176. That's validators actively paying — real work getting scored. Combined with mid-tier tao_in (not totally unknown, not crowded) and a mild price drop, this is the "quietly productive" profile.

What kills the thesis: Voice synthesis has real centralized competitors (11Labs, OpenAI Voice). If validators shift weight to more defensible compute subnets, TalkHead takes a hit. Also: emission at 176 is high — if it stays this high AND price doesn't respond upward over 2-3 weeks, something's wrong with the inflow side.

Exit watch: 21 days at high emission without price response → exit (thesis broken).

SN109 — Academia (14d price: −5.0%)

Signal: tao_in 871, price 0.00428τ, emission 153, 14d price −5.0%.

What it is: Academic research / paper summarization / citation inference — likely a niche academic-workflow subnet.

Why it qualified this week: Small tao_in (871, bottom quartile), stable emission (153 = median+), barely-moving price (−5%). Classic "forgotten but alive" profile. Academia has narrow customers but sticky ones — if they're using it, they keep using it.

What kills the thesis: Narrow-customer subnets have lower ceilings. The upside on Academia is moderate (+30-60%) rather than moonshot. If emission drops below 100, it means customers left.

Exit watch: emission < 100 → rotate.

SN113 — TensorUSD (14d price: +5.3%)

Signal: tao_in 1,796, price 0.00506τ, emission 118 (median), 14d price +5.3% (mildly trending up).

What it is: "TensorUSD" suggests a stablecoin or synthetic-dollar inference subnet — possibly a price-oracle or derivatives-related subnet.

Why it qualified this week: Mid-tier on all metrics with a mildly positive 14d (+5.3%) — this is the "early-trend" profile. Not yet overheated (+30% threshold far away), but capital has started leaning in. Emission exactly at median means it's not a runaway winner yet — still cheap relative to future potential.

What kills the thesis: Financial-infrastructure subnets (stablecoin, derivatives) compete with real TradFi infrastructure. If adoption doesn't materialize, emission stays flat and price fades. Also: a +5% 14d can become a +30% 14d fast, which would trigger the upper exit rule.

Exit watch: 14d > +30% → exit (overheated). OR: if 14d drops back below −10%, we rotate.

SN118 — HODL (14d price: −3.8%)

Signal: tao_in 2,464 (top of the bottom-25% range), price 0.00324τ (2nd cheapest), emission 183 (highest in portfolio), 14d price −3.8%.

What it is: Portfolio / holding / accumulation strategy subnet. The name is playful but the signal is serious.

Why it qualified this week: Highest emission in the portfolio (183) combined with the cheapest alpha price after SN94. This is where the algorithm is effectively saying "a lot of work is being scored here, and the alpha isn't priced accordingly." If the thesis ever plays out hard, SN118 is the one most likely to rip.

What kills the thesis: HODL-themed subnets are vulnerable to narrative rotation — if a "trading subnet" narrative takes over, HODL underperforms. Also the high emission at 183 is unusual; if that reverts to the median (normal behavior), the alpha price drops unless buyers step in.

Exit watch: emission drop of >40% in a week → exit immediately.

How we trade this portfolio

Entry: equal-weight (11.88% each) at Monday 00:00 UTC. Fills may take 30 minutes depending on liquidity — all pools are small.

Hold period: 30 days unless any individual exit rule triggers.

Individual exit rules (applied per pick): 1. 14d price change exceeds +30% → exit, lock in the gain. 2. Emission drops below 50 OR drops >40% week-over-week → exit, thesis broken. 3. tao_in drops below 500 (liquidity crisis) → exit.

Portfolio-level exit rule: - Regime flips (SN1 30d > +5% OR universe agg 30d > +10%) → exit ALL picks, hold TAO.

Rebalance: weekly. Picks that still qualify stay. Picks that fail rules or fall off the filter get replaced.

Why this is different from what other signals offer

Most Bittensor signals are either (a) momentum-chasing (which gets whipsawed), (b) bluechip-only (which can't 10x), or (c) vibes-based. This is none of those.

This is a rules-based, backtested filter with a regime gate and clear exit conditions. The backtest on 12 months of chain data produced 67 wins in 88 picks (76%). That doesn't guarantee the next 88 picks work — but it means the methodology has survived at least one market regime.

We run it with our own money. All 8 picks are live on Bittensor SN88 UID 180 with 257 alpha staked. Same subnets, same weights you see here.

Public track record — every past pick, every outcome, at signals.subnetaiq.io/signals/track-record.

The big caveats

1. One year of data is not ten. The filter has survived the split we tested it on, but Bittensor dTao is young. Market structure changed visibly since late 2025.

2. Small-subnet picks are volatile. Expect ±20-40% weekly swings on individual names. The portfolio smooths this — individual picks don't.

3. This is not financial advice. It's a signal. We run it with our own money. Do what makes sense for yours.

4. Dedup risk. If every SN88 miner runs similar logic, scores collapse via SN88's similarity penalty. We watch this weekly.

What you'll get every Sunday

• This full brief, ~2,500 words, 8 picks, named reasoning
• Regime gauge explaining why we're in or out
• Previous week's pick outcomes (what worked, what didn't, honest)
• Weekly commentary in the owner's voice (not generated)
• Live track record updates

Starting next Sunday this is subscribers-only. Join at signals.subnetaiq.io — $9.99/mo.

— SubnetAIQ