
FINX Analyst
48 posts







Retail Kospi investors bought everything foreign investors had to sell. But the retail party is ending: retail Margin Calls are soaring, hit 5% last Friday and on Monday will be far higher. Finally, retail brokerage deposits plunge by 30trn won to lowest level since feb 20



















this is going to be a critical weekly candle for $MAGS the Magnificent 7 ETF volume decreasing the entire way up to the right top and now increasing volume on the sell off












@alphaticaio So is this indicating near-term bearishness, or has a signal not yet fired?











$AMZN 🍌 NEWS: Amazon in talks to sell Nvidia rival chips to other companies 🍌 Opinion, this is very technical which is why I posted about switched up scale networks previously. The only other chip in the world that can do what they do is $NVDA. TPU’s don’t even come close


WEEKLY DARK POOL RECAP | June 8-12, 2026 "You can't classify direction in dark pools." We heard it (many times). Here are the receipts. Our direction classification algorithm produced 9 individual name signals this week across 1,809 off-exchange prints ≥$5M. We measured every signal against the actual return for the week ending June 12. Signal accuracy: 7 out of 9 correct direction (78%). Sell signal accuracy: 3 out of 3 (100%). Buy signal accuracy: 4 out of 6 (67%). ALPHATICA BUY SIGNALS vs ACTUAL RETURNS $INTC Signal: BUY Net: +$518M Return: +25.6% ✓ $MU Signal: BUY Net: +$230M Return: +13.6% ✓ $TSLA Signal: BUY Net: +$213M Return: +3.9% ✓ $NVDA Signal: BUY Net: +$712M Return: +0.04% ✓ $GLD Signal: BUY Net: +$634M Return: -2.4% ✗ $AAPL Signal: BUY Net: +$525M Return: -5.3% ✗ $INTC: our algorithm classified +$518M in cumulative dark pool flow as institutional buying. The stock returned +25.6%. By Friday, $INTC had +$743M in dark pool buys with zero sells and +$1.97B on the lit tape with $8M in sells. Both venues. Zero opposition. The algorithm identified the accumulation. The return confirmed it. $AAPL: our algorithm classified +$525M as institutional buying. The stock declined -5.3%. The signal missed. We show it. The reason for the selloff WWDC. We expect $AAPL to move higher from here. $GLD: +$634M classified as buying. A new position being built across two sessions. Gold declined -2.4%. Whether the dark pool was early or wrong is a question for this upcoming week. ALPHATICA SELL SIGNALS vs ACTUAL RETURNS $META Signal: SELL Net: -$700M Return: -4.4% ✓ $GOOGL Signal: SELL Net: -$424M Return: -2.4% ✓ $NFLX Signal: SELL Net: -$250M Return: -2.2% ✓ 3 out of 3. Every name our algorithm classified as institutional selling declined. $META: -$700M across three appearances, all classified as sell. Down -4.4%. $GOOGL: -$424M, mostly sell. Down -2.4%. $NFLX: -$250M in a single session with zero dark pool buys. Down -2.2%. The sell signals were perfect. When the algorithm flags persistent selling with zero or near-zero buyer interest, the name declined every time this week. INDEX SIGNALS: STRUCTURAL HEDGING $SPY Signal: HEDGE Net: -$2.76B Return: +0.57% $QQQ Signal: HEDGE Net: -$1.47B Return: +2.31% $IWM Signal: HEDGE Net: -$2.36B Return: +4.01% The algorithm detected institutional hedging at the index level. The dark pool sold all three indexes for the week while all three rose. This is not a miss. This is risk management. The institutions bought individual names ($INTC +25.6%, $MU +13.6%) and hedged broad market exposure off-exchange. The index selling funds the individual name buying. Different function. Different interpretation. WEEKLY SCORECARD Signal accuracy: 7/9 (78%) Sell signals: 3/3 (100%) Buy signals: 4/6 (67%) Biggest hit: $INTC +25.6% on buy signal Biggest miss: $AAPL -5.3% on buy signal We publish every signal in real time. We validate every signal against the actual return. We show the $INTC hit and the $AAPL miss. We show the $META sell that worked and the $GLD buy that didn't. "You can't classify direction in dark pools." 78% says otherwise. 100% on sells says otherwise. The receipts are on the timeline. Every day. Every week. The algorithm works. The tape doesn't have an opinion. It has a receipt.



