Trader

I start with the thesis, not the trade: what specific catalyst or mispricing I think the market hasn't priced in yet, and what time horizon I'm giving it to play out. Once I have that, I size the position against my risk budget for the day and the week, not against how convinced I feel, because conviction is the least reliable input in this job. I look at the liquidity of the instrument first: a position I can't get out of in a fast market is one I've mispriced regardless of the thesis. I set a stop level before I put the trade on, not after, and I size so that hitting the stop costs a fixed, known percentage of my daily risk limit, typically no more than 10 to 15 percent on a single idea. I also think about correlation with the rest of my book: a new position that looks small on its own can double my exposure to a factor I'm already carrying, like duration or a single sector. I execute in a way that limits market impact, working the order in clips rather than hitting the full size at once if it's meaningful relative to average daily volume. Once the thesis and the risk parameters are set, the entry itself is mechanical.

I manage a position every day it stays open, not just at entry. I mark it to market daily regardless of whether I plan to trade it, because the P&L tells me things the thesis doesn't, especially when it's moving against me for reasons I didn't anticipate. I keep the stop I set going in separate from the ongoing review of the thesis: the stop is mechanical and doesn't move, the thesis review does. If new information weakens the thesis before the price has moved, I reduce the position rather than waiting for the stop to do the work, because by the time the stop is hit the exit is usually worse than it needed to be. I also watch the position against the desk's overall risk limits, VaR, gross and net exposure by sector or currency, and concentration limits, because a position that looks fine on its own can push the book over a limit. For anything with optionality I track how the Greeks are shifting as the position ages or the underlying moves, since delta and gamma exposure change even if I haven't touched the trade. I'd rather take a small loss early on a thesis that's clearly broken than hold and hope.

I've learned that the moment a position turns against me is exactly when I'm most likely to make an emotional decision, so I try to have already made the hard decisions before that moment arrives. My stop levels and my maximum loss per position get set when I put the trade on, when I'm calm and thinking clearly, not renegotiated in the middle of a drawdown. When the market is moving fast I force myself to slow down: I step back from the screen for thirty seconds, check whether anything about the actual thesis has changed or whether this is just noise and volatility, and then act. Most of my worst decisions have come from staring at a P&L number instead of the market itself. I also talk to the desk during fast markets rather than going quiet, both because a second pair of eyes catches things I might miss under stress and because isolating yourself under pressure tends to make the decision-making worse, not better. If I've taken the loss my risk limit says I should take, I take it and move on rather than trying to make it back on the next trade, which is how a bad day turns into a bad month.

Technology handles the parts of the job that don't need judgment, which frees up my attention for the parts that do. I run screens that flag when spreads, correlations, or volatility move outside their normal historical range, which surfaces ideas I wouldn't necessarily catch by watching the screen manually. For execution I use algorithms like VWAP or TWAP to work larger orders so I'm not moving the market against myself, and I only work orders by hand when the size is small enough, or the situation unusual enough, that an algorithm would do a worse job than I would. I use Python for backtesting and for building quick models when I want to check whether a pattern I think I'm seeing actually holds up over a longer history rather than trusting my gut. Bloomberg and Reuters are the baseline for market data and news, but I've also built my own dashboards for the specific risk metrics I want to watch on my book that the standard terminal views don't show cleanly. What I don't do is let a model make the entry or exit decision on its own. The tools narrow down what I look at and speed up how fast I can check an idea, but the actual call is still mine.

I was running a position in a mid-cap industrial name on the view that a margin recovery was underway, based on input cost trends I'd tracked for two quarters. I sized it larger than I should have because the trade had worked well for the first few weeks, and that recent success pushed my conviction past where the risk budget actually justified it. The company then missed guidance on a customer concentration issue I hadn't dug into closely enough. I held the position a day and a half past my own stop, telling myself the miss was a one-off and the broader thesis was still intact. It wasn't, and by the time I closed the position the loss was roughly double what my original risk plan allowed for. I went back through the trade with my head trader and we identified two specific failures: I hadn't stress-tested the thesis against customer concentration risk at entry, and I broke my own stop discipline once I was already in the position. I now run a customer and supplier concentration check on every single-name idea before sizing it, and I haven't moved a stop since. The financial cost of that trade was real, but the discipline it forced has more than paid for itself.

My head trader wanted me to cut a rates position that was down but that I believed was still working, based on a view that the market had overreacted to a single data print without pricing in the trend behind it properly. Rather than just pushing back verbally, I put together a short written case: the historical pattern of similar overreactions to that data series, the level at which my original thesis would actually be invalidated, and a smaller size I could hold that would still express the view within a tighter risk budget. I framed it as a sizing question rather than an argument about who was right, because the position wasn't wrong at the size I originally had it, it just needed to be smaller given what had happened. My head trader agreed to let me hold a third of the original size with a tighter stop rather than closing it entirely. The trade recovered over the following two weeks and the reduced position captured most of the eventual move. What mattered wasn't that I was right, it was that I made the case with a specific, falsifiable argument rather than just insisting on my original conviction.

During a surprise central bank rate decision that came in well outside consensus, my book had meaningful exposure on the wrong side within the first ninety seconds of the announcement. I didn't have time for a full re-analysis, so I fell back on decisions I'd already made in advance: I knew my maximum acceptable loss for a single macro surprise on that position, and I hit it without hesitation rather than waiting to see if the move would reverse. Liquidity was thin in the first few minutes, so instead of trying to exit the full position at once and pushing the price further against myself, I worked out of it in two pieces, taking the first slice immediately and the second once the initial spike in volatility had settled a little. Afterward, once the position was flat, I spent the next hour actually working out what the decision meant for the rest of my book rather than trying to re-enter immediately, because the instinct to make the loss back right away is exactly the instinct that turns one bad trade into three. The desk reviewed the event afterward and confirmed the exit was within policy, which mattered more to me in that moment than being right about the direction.

If I'm long a stock and want to protect against a near-term drawdown without giving up all the upside, I'd typically buy a put, either outright or as part of a collar by selling a call against it to offset some of the premium. The put's delta tells me how much the option's value will move relative to the underlying, so I can calculate how many contracts I need to offset a given amount of directional exposure. Gamma tells me how that delta will change as the stock moves, which matters because a hedge that looks adequate today can become under-hedged or over-hedged quickly if the stock makes a large move, since the put's delta accelerates as it moves closer to the money. Theta is the cost of carrying the hedge: every day the option loses value from time decay if the stock doesn't move, so I weigh that cost against the protection I'm buying, particularly if I expect to hold the hedge for weeks rather than days. Vega matters if I'm putting the hedge on ahead of a known event like earnings, because implied volatility is often elevated going in and can collapse afterward regardless of what the stock does, which affects the value of the hedge independently of price. I size the hedge to the risk I actually want to remove, not to eliminate all volatility, since a fully hedged position at that point is just an expensive way of holding cash.

The spread I quote has to compensate me for two separate risks: the cost of holding inventory I don't want, and the risk that I'm trading against someone who knows more than I do at that moment. I widen the spread when volatility rises, because the probability of the price moving against my inventory before I can offset it goes up, and I widen it further around scheduled news or data releases, when the odds of trading against better-informed flow increase. I also skew my quotes based on the inventory I'm already carrying: if I'm already long more of an instrument than I want to be, I'll quote a lower bid and a lower offer to encourage buyers and discourage sellers, which pulls my inventory back toward flat without me having to cross the spread and pay for the privilege. I track my inventory against a target range throughout the day rather than letting it drift, because a market maker who accumulates a large directional position by the end of the day has effectively become a directional trader without meaning to. When flow is heavily one-sided for an extended period, that's usually information in itself, and I'll adjust my quotes more aggressively than a simple inventory model would suggest, because at that point the safer assumption is that the market knows something I don't yet.

Covered interest rate parity is the mechanism that ties these together. The forward rate reflects the rate that prevents a risk-free arbitrage between borrowing in one currency and lending in the other, rather than a market forecast of where spot will end up. If I can borrow euros cheaply, convert to dollars at spot, invest the dollars at a higher rate, and then convert back to euros at a locked-in forward rate, that trade should produce exactly zero risk-free profit once the forward is priced correctly, otherwise everyone would do it until the arbitrage disappeared. The forward points are the market's expression of the interest rate differential between the two currencies over that tenor: if dollar rates are higher than euro rates, the euro trades at a forward premium against the dollar, and the size of the forward points scales with both the rate differential and the length of the tenor. In practice this means forward points move when either currency's short-term rate expectations shift, even if spot hasn't moved at all, a distinction traders new to FX sometimes miss when they see forward points changing and assume it's a spot signal. I use this relationship to check whether an FX forward is priced fairly relative to the underlying rates before trading it, rather than trusting the quoted price on its own.