In this article, you’ll learn why win rate alone tells you little about trading profitability, how my post-earnings strategy maintained positive expectancy with only 45% positive outcomes, and why an average +8.74% winner versus -4.74% loser makes losses manageable. I’ll compare those results with $12,963.50 in Uber earnings across 594.15 hours and explore the common lesson: good risk management isn’t about eliminating bad outcomes—it’s about making the winners large enough to compensate for them.

---

I spend a surprising amount of my week doing two things that seem completely unrelated — studying post-earnings stock setups and delivering food for Uber.

One involves charts, earnings reports, stop losses and trading capital. The other involves kilometres, restaurants, delivery requests and waiting for somebody to order a Big Mac.

But after tracking both closely, I’ve noticed something strange: both activities depend heavily on averages.

My Uber earnings can vary dramatically from one shift to the next, just as individual trading outcomes can swing between winners and losers. But over a large enough sample, those individual results begin to matter much less than the average outcome they produce together.

In this article, I’m going to use my own Uber driving and post-earnings trading data to explore why losses aren’t necessarily a problem, why averages matter more than individual outcomes, and how risk management can create positive expectancy even when plenty of trades don’t work out.

We’ll also take a quick look at Uber’s fundamentals and payout methodology, because I believe the company itself demonstrates the same principle:

Individual order earnings don’t necessarily need to produce identical economics when the math is applied across a large enough sample.

After all, Uber is no longer merely a high-growth gig-economy company—it is now a business producing nearly $10 billion in annual free cash flow.

Ultimately, whether I’m staring at a losing stock position, sitting through a lousy $17-per-hour Uber shift, or looking at how Uber manages billions of trips, the underlying lesson is surprisingly similar:

• Profitable trading doesn’t require every trade to win. Decent Uber earnings don’t require every shift to pay well. And Uber doesn’t need every transaction to produce identical economics. What matters is whether the good outcomes are large enough to compensate for the bad ones.

---

---

Quick Answer: Why Does Average Win Size Matter More Than Win Rate?

Average win size matters more than win rate because traders don’t need to win most of their trades to maintain positive expectancy—they need their winners to be large enough to pay for their losers. In my post-earnings data, only about 45% of modeled outcomes were positive, but the average winner was approximately +8.74%, compared with an average loss of just -4.74%. That means the average winner was roughly 1.84× larger than the average loser, allowing the strategy to maintain positive expectancy despite losing more often than it won. Whether trading stocks, investing, or delivering food, the lesson is the same: you don’t need every outcome—or even most outcomes—to be good when your stronger results are large enough to compensate for the weaker ones.

---

Key Statistics – Day Trading, Uber & Risk Management

My own trading and Uber data tell remarkably similar stories: the average matters much more than any individual outcome. My post-earnings strategy produced an estimated positive expectancy despite fewer than half of the modeled outcomes being profitable, while my Uber earnings were dragged down by plenty of mediocre shifts but lifted by much stronger ones.

Here are some of the numbers behind that comparison:

• 45.0% — Approximate positive-outcome rate across my completed post-earnings setups using my modeled exit rules.
• +8.74% vs. -4.74% — My average winning outcome compared with my average losing outcome.
• 1.84:1 — Ratio between my average winner and average loser. One average winner could therefore offset approximately 1.84 average losses.
• +1.38% — Estimated average expectancy per post-earnings setup under the modeled exit system.
• +9% / -5% — The profit-target and stop-loss framework used to model those outcomes.
• 10.89% vs. 5.67% — Average next-day maximum favourable excursion (MFE) compared with average absolute maximum adverse excursion (MAE) in my post-earnings dataset.
• 130,628 traders — A 2024 study of individual futures traders found that profitability was linked to whether traders’ realized gains were large enough to cover their realized losses. Traders exhibiting a stronger disposition effect—holding losers while realizing winners—experienced greater losses.
• 193 professional traders — A separate experimental study illustrates why risk management can’t be reduced to a universal “always cut immediately” rule: whether holding or realizing a position was rational depended partly on the behaviour of the underlying market.
• $12,963.50 — Total earnings across the valid periods in my Uber-driving dataset.
• 594.15 hours — Total recorded online time, producing approximately $21.82 per online hour.
• 53 of 97 shifts — More than half of my individually timed Uber shifts earned less than $20 per hour, yet the broader average remained above $21.
• $17.44 vs. $28.24/hour — Approximate weighted earnings for my sub-$20 shifts versus shifts earning $20 or more. The strong periods compensated for the weak ones.
• $27.61 vs. $21.08/hour — Average active-time versus online-time earnings across the Uber periods where I recorded both measures.
• $193.5 billion — Uber Technologies generated approximately this much in Gross Bookings during 2025, across 13.57 billion trips.
• $52.0 billion — Uber’s 2025 revenue, up 18% from $44.0 billion in 2024.
• $5.57 billion — Uber’s 2025 GAAP income from operations, nearly double the $2.80 billion recorded in 2024.
• $9.76 billion — Uber’s 2025 free cash flow, up 42% year over year, showing just how profitable the platform has become at enormous scale.

---

The comparison isn’t that my trading strategy, my delivery shifts and Uber Technologies operate identically. They obviously don’t. The connection is that profitability is ultimately a distribution problem.

A trader can afford losing trades when winners are disproportionately larger.

I can afford lousy Uber shifts when better periods lift my long-term hourly average.

And Uber itself operates billions of individual transactions while ultimately being judged by what those transactions produce in aggregate.

*My trading and driving statistics are based on my own records and modeled trading rules. They should not be interpreted as guaranteed trading returns or representative earnings for other Uber drivers.

---

Losing Trades Are Fine When Winners Are Bigger

One of the easiest ways to misunderstand risk management is to assume that a profitable trading strategy should produce more winners than losers.

It doesn’t necessarily have to.

My post-earnings data is a good example. Using my modeled +9% profit target and -5% stop-loss exit framework, only about 45% of outcomes were positive.

On the surface, a system that loses more often than it wins doesn’t sound particularly attractive. But the size of those wins and losses changes the picture.

My average winning outcome was approximately +8.74%, compared with an average losing outcome of -4.74%. That gives the system an average reward-to-risk ratio of roughly 1.84:1.

With a fixed $1,000 position, that’s about +$87 for an average winner versus -$47 for an average loser. One average winner can therefore pay for approximately 1.84 average losers.

Using my results, a relatively unimpressive win rate can still produce positive expectancy because the payoff is asymmetric.

If I consistently risk roughly 5% for the opportunity to capture approximately 9%, I don’t have to correctly predict every earnings move. I just need to control what happens when I’m wrong and make enough when I’m right.

It’s also true that my results aren’t unique in showing why the relationship between gains and losses matters.

A 2024 study examining 130,628 individual futures traders found that profitability was closely related to whether realized gains were large enough to cover realized losses. Traders displaying a stronger disposition effect—the tendency to realize winners while continuing to hold losers—experienced greater losses.

That’s risk management in its simplest form: don’t eliminate losses—keep them disproportionately smaller than your gains.

---

My Uber Earnings Follow a Similar Risk-to-Reward Pattern

Driving for Uber obviously doesn’t have a literal stop loss or profit target, but tracking my earnings has made the logic behind risk/reward much easier for me to visualize.

Across 97 individually timed Uber shifts in 2026, 53 produced less than $20 per online hour, while only 44 reached $20 or more.

In other words, more than half of those shifts fell below a pretty modest $20/hour benchmark.

But again, the average outcome matters, not any one individual shift or hour within a shift.

My sub-$20 shifts averaged approximately $17.44/hour, while the stronger $20+ shifts averaged about $28.24/hour. Across the broader periods in my tracker, my weighted earnings still came to approximately $21.82 per online hour.

The stronger shifts compensated for the weaker ones, just like how in trading your average gains should compensate for your average losses.

Obviously, my driver earnings are not exactly the same as a trading risk/reward ratio—a $17 Uber hour is still positive income rather than a financial loss—but the averaging principle is remarkably similar.

There’s also an important lesson about measuring the correct denominator.

Uber (UBER) drivers, as well as drivers on similar platforms like Instacart or Maplebear (CART), Lyft (LYFT), or DoorDash (DASH), can track either the time they spend online OR their active time, which is the amount of time they spend actively delivering people, food or groceries.

During the periods where I tracked both measures (online and active time), I earned approximately $27.61 per active hour but only $21.08 per online hour.

If I only counted my active time, I’d make the economics of driving look substantially better than they really are. My online time captures the hours spent waiting for orders too—and drops that $27.61/hour figure to a much more realistic $21.08/hour.

Traders can make the same mistake by focusing on winning trades while mentally discounting losers, commissions, slippage or periods when capital isn’t productive.

Risk/reward only tells the truth when you count the entire distribution.

---

Uber (UBER) Plays the Averages Too

All of this gets even more interesting when I stop looking at myself as the driver and look at Uber as the company paying me.

I don’t have access to Uber’s internal economics or fare-setting algorithm for individual deliveries, so I can’t determine whether Uber makes or loses money on any particular order—or confirm exactly how factors such as tips, distance, demand, driver availability and order batching influence what Uber ultimately pays the driver.

But I can observe how some driver payouts behave.

In my market, I generally see about a $3 minimum fare for a standalone delivery. But when Uber stacks a second order onto a delivery I’m already completing, I’ve sometimes seen that additional order add only about $2 to my fare.

That’s just a $1 difference, but across enough deliveries, small savings can add up and those small differences could translate into meaningful savings for Uber at scale.

I’ve also noticed what may be a similar pattern with large customer tips.

A 20-km delivery without a tip might require Uber to contribute $10 to make the trip worthwhile.

But if a customer leaves a $10 tip, I’ve sometimes seen comparable trips where Uber’s contribution is closer to $5—meaning the driver still earns a better $15 total payout, while Uber contributes $5 less.

I can’t confirm that the tip itself causes Uber to reduce its fare, but I’ve repeatedly noticed that large tips can coincide with surprisingly low Uber-funded payouts.

And that’s really the point:

Small differences in the economics of individual deliveries can become meaningful when repeated across a massive sample.

Risk management isn’t always about making one enormous decision. It can mean repeatedly improving the relationship between cost and reward across a very large sample.

Uber operates at an almost absurd scale.

In 2025, the company reported approximately 13.57 billion trips, $193.5 billion in Gross Bookings and $52.0 billion in revenue. It also generated approximately $9.76 billion in free cash flow.

At that scale, marginal economics matter.

If two deliveries can efficiently share some of the same driver time and kilometres, Uber may be able to reduce the incremental amount required to get the second order delivered.

Other deliveries may require higher fares, incentives or supplements to attract a driver.

They win some, they lose some, but profitability ultimately depends on what those economics produce in aggregate.

That’s not evidence that Uber earns a specific amount from any single driver trip or customer order.

But it does demonstrate why unit economics must be considered across a distribution rather than judged from any one single transaction.

It’s another version of the lesson from my trading data: small differences between risk and reward become extremely meaningful when the process is repeated enough times.

---

Conclusion — Risk Management Is About Surviving Long Enough for the Math to Work

Ultimately, this brings everything back to trading.

My post-earnings strategy doesn’t need every setup to work. My Uber driving doesn’t need every shift to pay $25 per hour. And Uber doesn’t need every trip to produce identical economics.

Everything depends on the distribution of outcomes—not any single result.

Consider a strategy that wins just 45% of the time, making $90 on winners and losing $50 on losers. Across 100 trades, 45 winners generate $4,050, while 55 losers cost $2,750.

That’s still a theoretical +$1,300 despite losing more often than winning.

The key takeaway is that losses aren’t evidence that risk management failed. They’re exactly what risk management is designed to survive.

Keep losses small.

Give winners room to matter.

Repeat the process.

In the end, good risk management keeps you in the game long enough for the averages to work.

---

If you want to go deeper:

• Explore the Trading Statistics Hub to understand how different sectors behave across market cycles
• Study real setups inside the Trade Reviews section
• Learn the framework behind high-probability setups in the Post-Earnings Momentum Strategy

This is how you turn raw market data into repeatable trading edge.

---

Frequently Asked Questions

What is risk-to-reward ratio in trading?

A risk-to-reward ratio compares how much a trader is willing to lose with how much they expect to gain. For example, risking 5% to target a 9% gain represents roughly a 1:1.8 risk-to-reward ratio. Risk/reward becomes especially useful when combined with win rate because the two determine whether a strategy has positive or negative expectancy.

Can a trading strategy be profitable with a win rate below 50%?

Yes. A trader can lose more often than they win and still be profitable if their average winners are sufficiently larger than their average losers. In my post-earnings data, only about 45% of modeled outcomes were positive, but my average winner was approximately +8.74%, compared with an average loss of -4.74%—a ratio of roughly 1.84:1.

What is positive expectancy in trading?

Positive expectancy means that the average expected outcome of a trading strategy is profitable across a sufficiently large sample. It is commonly calculated as:

Expectancy = (Win Rate × Average Win) − (Loss Rate × Average Loss)

A high win rate doesn’t automatically produce positive expectancy; the size of the average winner and average loser matters too.

Why can average win size matter more than win rate?

Because one sufficiently large winner can pay for multiple smaller losses. In my dataset, an average winner of +8.74% was approximately 1.84 times larger than my average loss of -4.74%. With a $1,000 position, that’s approximately +$87 versus -$47, meaning one average winner could compensate for nearly two average losers.

What is the purpose of a stop loss?

A stop loss helps define and control the amount of capital exposed when a trade moves against you. It doesn’t prevent losing trades—it prevents an ordinary losing trade from becoming disproportionately large. This becomes particularly important when a trading strategy depends on average winners being larger than average losers.

How are Uber earnings similar to trading expectancy?

The comparison isn’t exact because a slow Uber shift still generates income, while a losing trade actually loses capital. However, my data shows a similar averaging effect. 53 of 97 individually timed Uber shifts earned less than $20 per online hour, while my sub-$20 shifts averaged $17.44/hour and stronger $20+ shifts averaged $28.24/hour. Those stronger periods helped lift my broader weighted average to approximately $21.82 per online hour.

Why does sample size matter when measuring a trading strategy?

A few trades can be heavily influenced by randomness. Expectancy describes an average across many observations, not what will happen on the next trade, so a larger sample provides more useful evidence about whether the observed win rate and average win/loss relationship are persistent. Even then, historical positive expectancy does not guarantee future results because market conditions can change.

Is active hourly pay or online hourly pay more useful for Uber drivers?

Both measure different things, but online hourly earnings provide a better picture of what the driver’s total time generated because they include time spent waiting for orders. In the periods where I tracked both, I averaged approximately $27.61 per active hour versus $21.08 per online hour. Looking only at active time would therefore make my driving earnings appear substantially better.

Does Uber pay drivers less when customers leave large tips?

I can’t confirm that. I don’t have access to Uber’s fare-setting algorithm, and factors such as distance, demand, driver availability and batching can affect individual payouts. I have personally observed large tips sometimes coinciding with unusually low Uber-funded fares, but that is an observation from my own deliveries—not evidence that tips directly cause Uber to reduce driver pay.

What is the biggest risk-management lesson from comparing trading and Uber driving?

Don’t judge a system by one outcome. A losing trade doesn’t necessarily make a trading strategy unprofitable, just as one $17/hour Uber shift doesn’t establish what driving pays over hundreds of hours. What matters is the entire distribution: how frequently good and bad outcomes occur, how large they are, and whether the favourable outcomes are large enough to compensate for the unfavourable ones.

References

Ben-David, I., & Hirshleifer, D. (2012). Are investors really reluctant to realize their losses? Trading responses to past returns and the disposition effect. The Review of Financial Studies, 25(8), 2485–2532. https://doi.org/10.1093/rfs/hhs077

Odean, T. (1998). Are investors reluctant to realize their losses? The Journal of Finance, 53(5), 1775–1798. https://doi.org/10.1111/0022-1082.00072

Uber Technologies, Inc. (2026). Uber announces results for fourth quarter and full year 2025. Uber Investor Relations. https://investor.uber.com/news-events/news/press-release-details/2026/Uber-Announces-Results-for-Fourth-Quarter-and-Full-Year-2025/default.aspx

Uber Technologies, Inc. (2026). Uber announces results for second quarter 2026. Uber Investor Relations. https://investor.uber.com/

Original Data

Laforest, J. (2026). Post-earnings momentum tracker 2026 [Unpublished dataset].

Laforest, J. (2026). Uber driver earnings tracker 2026 [Unpublished dataset].