Spent 3 weeks digging through MELI's financials. It's the Amazon + PayPal of Latin America and the market prices it as a simple e-commerce player.Bull CaseFintech Mercado Pago: 52M active users, +45% YoY. The real growth engine. Consumer credit booming in Brazil and Mexico.Logistics: Own fulfillment in 6 countries. Next-day delivery in major Brazilian cities. Massive competitive moat.Advertising: Ads segment at $2.1B annualized, margins >70%. This is the next Retail Media play like Amazon.ValuationAt $1,850 — we're at 28x forward earnings for 35% revenue growth. US comps trade at 40-50x for less growth. My target: $2,600 in 18 months.Position: 45 shares @ $1,720 avg.
Q1 results dropped and AMD posted 80% growth in the data center segment. EPYC Turin is selling like hotcakes and hyperscalers are seriously diversifying away from Intel.What I'm watching:Server market share went from 18% to 24% in one yearMI300X is starting to find its place against the H100Gross margin expanding for the 6th consecutive quarterI'm holding 350 shares bought at $148, not moving. Valuation is still reasonable compared to NVDA at 35x forward.
Full transparency. $80K deployed in DeFi over 12 months. No shitcoins, no exotic farming.Allocation40%: Aave V3 on Arbitrum — lending ETH/USDC, ~8% APY net25%: Uniswap V3 LP ETH/USDC range 0.85-1.15 ratio — 22% APY with monthly rebalancing20%: Pendle — yield tokenization on stETH, locked 12 months, 15% fixed15%: GMX — LP on Arbitrum, 18% APY but more volatileTotal return net of gas fees: +$24,800 or 31%Risks taken: smart contract, IL on Uni V3, oracle manipulation on GMX. Nothing is risk-free.
Gas fees on Ethereum mainnet have never been this low. 3-5 gwei on average. Two possible reads:Bullish readActivity is migrating to L2s (Base, Arbitrum, Optimism) which settle on L1. ETH remains the security layer, L2s pay. The ecosystem is expanding.Bearish readLess mainnet activity = less EIP-1559 burn = ETH becomes inflationary again. Net supply +0.3%/year currently.I'm long ETH but starting to doubt. Are L2s capturing value at the expense of L1?
Sold Iron Condors on SPY every week for 14 months. $75K capital. Here are the raw numbers.Setup: Wings at 16-delta, 30 DTE, $5 wide. Managed at 50% of credit or 21 DTE.Results:58 trades closedWin rate: 71%Avg win: +$285, avg loss: -$520Total P&L: +$8,940 or 11.9% annualizedMax drawdown: -$3,200 (4-loss streak in October)It's not spectacular but it's consistent. The only problem is that strongly directional months (like the November rally) hurt me.
The $14.5B JEDI II Pentagon contract was just awarded to Amazon. Microsoft loses a structurally important deal.Sell volume up 2x this morning. The issue isn't just this contract, it's the signal: Azure is losing ground in the public sector.Anyone else worried about Azure growth slowing to 26%? We were at 35% four quarters ago.
I bought LEAPS (deep ITM calls) instead of shares for 2 years. Here's my takeaway:Pros:2-3x leverage with defined riskLess capital tied up — $15K controls the equivalent of $40K in sharesCan sell calls against them (poor man's covered call)Cons:Theta decay even at 70-80 delta, it nibblesBid/ask spread sometimes wide on deep ITM strikesNo dividendsNet result: +38% over 2 years vs +29% if I had held shares directly. The leverage worked in a bull market.
Glassnode data that nobody's looking at:Addresses >1000 BTC added 47,000 BTC in 6 weeksExchange reserves at their lowest since 2019MVRV ratio at 1.8 — historical accumulation zone, far from euphoria (>3.5)Hash rate at ATH despite the halving — miners aren't capitulatingPrice is stalling around $95K but smart money is buying. Distribution by retail, accumulation by institutions. This pattern preceded every major rally.
Quit my $72K/year accounting job in October to trade NQ futures full-time. Here's the unfiltered reality.Numbers:Starting account: $30KTotal P&L over 6 months: +$14,200 (47%)Green months: 4 out of 6Worst month: -$3,800 (December, choppy market)Best month: +$5,200 (January, nice trend)What nobody tells you:The loneliness is crushing. Former coworkers don't understand what I do anymore.Private health insurance costs $620/month in the US.Red months make you question everything.You actually work more hours than before (prep at night, post-close review).I'm continuing, but it's not the dream life YouTube sells.
USD/CHF at 0.8650, lowest since 2015. The market is voting with its feet.Why CHF is outperforming:The SNB has a much healthier balance sheet than the Fed and ECBSwiss inflation at 0.8%, lowest in the G10Safe haven flows from geopolitical tensions (Middle East, US elections)Short USD/CHF from 0.8780, stop 0.8850, target 0.8500. The trend is your friend here.
Hey everyone. Software developer for 5 years, finally have some savings to invest. I've read a lot of contradictory stuff and I'm afraid of making a mistake at the start.My questions:Is a broad market ETF like VTI or VOO really the best approach to start?Should I paper trade before putting real money in?Best broker? I keep hearing about Fidelity, Schwab, IBKR...How much time per week should I dedicate when starting out?I'm not looking to become a day trader. I want to learn to invest intelligently long-term and maybe try swing trading later.
1. UBER — Ascending triangle on the dailyResistance at $82, ascending support from $71. Breakout above $83 = target $94. Stop $76.2. CRM — Pullback to the 50 SMA after earningsPost-earnings gap up to $312 has been filled. Perfect support on the 50 SMA at $295. Bounce underway. Target $330, stop $288.3. PANW — Weekly bull flagTight flag after a 40% run. Entry >$385, target $440, stop $365.4. LLY — 5-month cup and handleCup bottom at $710, rim $920, handle at $870. Breakout >$925 = measured move to $1,130. Stop $845.
I've been trading for 3 years and tape reading (order book reading) was my main edge. But the last 6 months it's gotten much harder.Spoofing algos are everywhere. Massive orders appear and vanish in 50ms. Level 2 has become a funhouse mirror.Those who do tape reading: have you adapted your approach? What tools do you use to filter out algo noise?
The rate differential between the BoE (4.75%) and the BoJ (0.25%) makes GBP/JPY incredibly attractive for carry trade. You earn +$18/lot/day just by staying long.The problem: the BOJ is getting closer to tightening. Ueda signaled that ultra-accommodative policy "cannot last forever." If the BOJ hikes to 0.50%, the yen strengthens and the carry trade reverses violently.My position: long GBP/JPY at 50% of usual size. The carry is too good to ignore but the reversal risk is real.
Active trader for 4 years. Used each broker as my primary account for at least 6 months. Here's my honest ranking:1. IBKR Pro — 9/10+: Best execution, unbeatable commissions, global access, solid API. -: TWS interface is archaic, customer support is average.2. Tastytrade — 8.5/10+: Best options platform, built-in courses, community. -: Execution slightly below IBKR, no forex.3. Schwab/ThinkorSwim — 8/10+: Incredible research tools, TOS is powerful, integrated banking. -: A bit slow since the TD merger.4. Webull — 6.5/10+: Clean interface, good charting tools, fractional shares. -: PFOF, limited order types, no futures.5. Robinhood — 5/10+: Pretty interface, integrated crypto. -: Payment for order flow, mediocre execution, dangerous gamification.
Morning gaps are a swing trader's nightmare. Here's my system after 4 years:Position sizing: never more than 3% of the account on a swing. If I lose the max historical gap, that's -4.5% of the account = survivable.Gap-adjusted stop loss: I set my stop below support + the average gap margin of the stock (calculated over 90 days).No swings before earnings, FOMC, CPI: absolute rule. I close everything 2 days before.Hedging: if I have 3+ long positions, I buy a mini SPY put as insurance.Since applying these rules, adverse gaps cost me 0.5% of the account on average instead of 2-3%.
After 3 years of algo trading and hundreds of backtests, here's why 90% of backtests you see online are misleading:Survivorship bias: you test on stocks that still exist. The ones that went bankrupt are excluded. Inflates returns.Look-ahead bias: using data that wasn't available at the time of the trade. Even split adjustment can cause this.Overfitting: 20 parameters optimized over 10 years = perfect curve, catastrophic real performance.Slippage ignored: "I made 200%" — with perfect fills at mid-price that you'll never get in reality.Transaction costs: commission + spread + market impact. On a high-frequency algo, it changes everything.Data snooping: testing 50 strategies and keeping only the best one. Probability it works forward: close to zero.Regime change: what worked in 2015-2020 (infinite QE, zero rates) doesn't work in 2024-2026 (high rates, inflation).
Built a mean reversion bot on ETH/BTC and SOL/BTC pairs. Running for 8 months on Binance.LogicThe ratio between two cryptos tends to revert to its mean. When the 20-day z-score exceeds +2 or -2, I take the opposite position.Results (8 months, $20K)Total return: +18.4% ($3,680)Sharpe: 1.2Max drawdown: -6.8%Win rate: 63%Number of trades: 142Stack: Python, ccxt, PostgreSQL, Grafana for monitoring. Code is on my GitHub (link in comments).
I bought 50 shares of a small cap at $12.15 but the displayed price was $12.05. The $0.10 per share difference cost me $5 extra.Apparently it's the "bid-ask spread." Can someone explain it simply?Why is the spread sometimes wide and sometimes tight?How do I avoid getting caught?Does it change depending on the time of day?