2025 didn’t go as planned.
I spent most of that year sitting on cash and biotech stocks, watching Nvidia rally while I convinced myself the AI bubble was about to pop. It didn’t. The S&P 500 did 16.32%. I did 3.98%.
That’s a 12.34% gap. For context, in 2024 I beat the market by 7.13%. So over the full 2+ years at Saxo Markets (October 2023 to December 2025), I’m up 65.44% versus the S&P’s 61.28%. A 4.16% lead.
Let me back up.
How I Got Here (The Short Version)
I started buying stocks in March 2009. Yes, literally the market bottom during the financial crisis. Anyone who started then and didn’t panic did well. So when people ask “how’d you do over 16 years,” I have to be honest: I don’t know.
UOB Kay Hian traded directly from my personal bank account—no separate brokerage account with deposit/withdrawal tracking. This makes cumulative return calculations impossible, even though I have the trade history. When I switched to TD Ameritrade in 2019, I lost access to their platform data.
I switched to TD for better US market access, lower fees, and superior research tools. When Charles Schwab acquired TD Ameritrade in 2020, they eventually renamed TD Ameritrade Singapore to Charles Schwab Singapore. In 2023, Schwab Singapore exited retail clients to focus exclusively on accredited investors. I had to switch. I liked the thinkorswim platform, actually. But I couldn’t go back to UOB Kay Hian—their fees were too high. So I picked Saxo Markets over Interactive Brokers.
Rookie mistake not keeping my own records. So this whole article is based on my Saxo period: October 2023 to December 2025. It’s the only time I have complete, verified data.
That’s important context because I’m showing you a 2+ year snapshot during a bull market, not a 16-year track record. If I’d started in 2007 instead of 2009, this would be a very different conversation.
What Actually Happened
2023 Q4: I switched to Saxo in October and immediately built positions in AI infrastructure and tech. Returned 21.77% while the S&P did 12.25%. Outperformed by 9.52% in 2.5 months.
Everything felt right. Momentum was there, conviction was high.
2024: This was the year I convinced myself I was actually good at this.
Palantir up 150% as the AI infrastructure narrative exploded. Meta up 75% as advertising recovered and Zuckerberg stopped lighting money on fire with the metaverse. PDD Holdings up 100% because I bought Chinese stocks when everyone else was running away.
My portfolio was concentrated—Palantir, Meta, Cloudflare, Microsoft dominated. I wasn’t diversified, I was focused on a single theme: AI/tech.
Returned 30.66% versus the S&P’s 25.02%. Beat the market by 7.13%.
2025: This is where it gets frustrating.
In early 2025, I looked at AI valuations and thought: this is insane. Everyone’s piling into the same seven stocks. Nvidia’s trading at whatever multiple. I’d seen this before—it never ends well.
So I positioned defensively. I expected a 50% drawdown in tech, used VIX signals anticipating 15-20% market decline. I bought put options, built up cash, rotated out of AI/tech into biotech (TScan Therapeutics, Ovid Therapeutics), energy (Schlumberger, Scorpio Tankers), fintech (SoFi).
The crash never came. Well, there was a black swan event in April that caused a brief selloff, but it wasn’t the sustained AI bubble collapse I expected. By mid-year, AI was rallying even harder.
I sat there watching Nvidia, AMD, semiconductors boom—the exact stocks driving the S&P’s 16.32% gain—while my biotech positions lost money and my options expired worthless.
Returned 3.98% versus 16.32%. Underperformed by 12.34%.
The painful irony? In 2024, I won by concentrating in AI/tech. In 2025, I became convinced that exact concentration would blow up, so I diversified. The market proved me wrong.
My Numbers
I don’t have cumulative returns from UOB Kay Hian (2009-2019)—they traded directly from my personal bank account with no deposit/withdrawal tracking, making performance calculations impossible. But I can compare trade-level metrics across both periods:
| Metric | UOB Kay Hian (2009-2019) | Saxo Markets (2023-2025) |
|---|---|---|
| Trades | 38 closed positions | 40+ trades |
| Win Rate | 76.3% | 75% |
| Average Winner | +40% | +84% |
| Average Loser | -21% | -37% |
| Win/Loss Ratio | 1.93x | 2.3x |
| Sharpe Ratio | N/A | 1.1 |
| Sortino Ratio | N/A | 1.2 |
That 2.3:1 ratio at Saxo is what matters. When I’m right, I’m very right. When I’m wrong, I cut before it spirals.
The comparison shows evolution in my approach: UOB period had more conservative position sizing (smaller wins, smaller losses). Saxo period shows bigger conviction bets with higher win/loss ratio despite similar win rate.
My Sharpe ratio of 1.1 is decent—hedge funds typically hit 0.5-1.0. But that’s measured during a 2+ year bull market. The S&P’s Sharpe during this period was probably similarly elevated, so I’m not sure it means I actually did better on a risk-adjusted basis.
The Concentration Risk Thing
Here’s what bothers me: I’m concerned about S&P 500 concentration, but the concentrated index beat me in 2025.
As of December 2025, the Magnificent 7 (Apple, Microsoft, Alphabet, Amazon, Meta, Nvidia, Tesla) are 34% of the S&P 500. When I diversified away from this concentration in 2025, it cost me 12.34%.
The market rewarded concentration. I fought it. I was wrong on timing.
But historically, such concentration eventually reverses. When it does, passive holders have no choice but to ride it down. Active investing lets me try to position ahead of that reversal—even if I’m early.
The question isn’t whether concentration will reverse. It’s when. And whether being early by one year, two years, or five years matters more than capturing the final surge.
The Thing I Keep Coming Back To
Market timing is brutally hard. I got 2024 right by staying concentrated in AI/tech. Then I got 2025 wrong by rotating out because I thought the bubble would pop. Same analytical framework, same conviction level, completely opposite outcomes.
Before I make any moves for 2026, I need to understand what actually happened. Not just the returns, but the decision-making process.
That means going through the winners—Palantir, Meta, PDD, NIO, Cloudflare. What did I get right? Was it thesis development, entry timing, position sizing? Nearly selling Palantir at +40% because the valuation looked insane. Talking myself into Chinese stocks when everyone was fleeing. The NIO position down 39% before I figured out what I’d gotten wrong about the thesis.
And it means going through TScan and Ovid. The averaging down because I “understood the sector.” The holding on because selling would mean admitting the mistake. How mental accounting made me treat biotech losses differently than my AI positions.
I need to evaluate my strengths and weaknesses from 2023-2025 before I decide what to do about 2026. Because I still think the AI bubble crash could come in H1 2026. The question is: do my 2024 winners and 2025 failures tell me to trust that conviction, or do they reveal blind spots I need to account for?
Once I’ve worked through that analysis, I’ll have a clearer picture on how to reallocate. Whether to stay defensive, cut biotech losses, rotate back into tech, or something else entirely.
That’s what the next pieces are about—evaluating the past to inform what comes next.