When you invested in that stock, or added to that mutual fund — did you think about how much to put in?
Not what to invest in. Most of us think hard about that. We research, compare, ask around.
But how much — as a deliberate, reasoned decision — is a question most investors never seriously ask.
We invest what feels right. What feels comfortable. What we happened to have lying idle.
And in doing so, we make one of the most consequential decisions in our financial lives almost entirely on instinct.
There is a better way to think about this. And it comes from an unlikely place.
The Man Who Worked at Bell Labs
John Kelly was a physicist at Bell Labs in America in the 1950s — the same place where Claude Shannon invented information theory, which we explored in the Shannon's Demon post.
In 1956, Kelly asked a precise question: If you have an edge, How much should you bet to maximize the long-term compound growth of your wealth while avoiding eventual ruin?
The answer he arrived at is called the Kelly Criterion. And it starts with a simple gambling table.
The Formula — Walk Through It Once
Imagine you are offered a coin flip. The coin is slightly biased — it lands heads 60% of the time, tails 40%. Heads, you double whatever you bet. Tails, you lose whatever you bet.
You have ₹1,00,000 total. Before each flip, you decide how much of that total to put on the table. That amount is your bet size. The rest stays in your pocket, untouched.
Kelly's formula tells you exactly what percentage of your current total to bet each time:
Kelly % = P(win) − [ P(lose) ÷ Odds ]
Where P(win) is your probability of winning, P(lose) is your probability of losing, and Odds is how much you win per rupee risked. Here, every ₹1 you bet wins you ₹1 back — so Odds = 1.
Plugging in: 0.60 − (0.40 ÷ 1) = 0.20
Kelly says: put 20% of your current total on the table each round. Not a fixed ₹20,000 forever — but 20% of whatever your total is at that moment. As your wealth grows, the bet grows. As it shrinks, the bet shrinks. This is what makes it a compounding strategy.
What Happens Over 10 Rounds
Now imagine three investors playing the same game. Same coin, same flip sequence. Only their bet sizes differ.
- Kelly bets 20% of his current total each round — exactly what the formula says
- Over-bettor bets 40% — confident in his edge, doubles down
- Under-bettor bets 10% — cautious, plays it safe
The flip sequence over 10 rounds looks like this: W, L, W, W, L, W, L, W, W, L — 6 wins and 4 losses in a realistic, mixed order. Not arranged to prove a point.
Here is exactly how each round plays out, starting with ₹1,00,000:
|
Round |
Outcome |
Kelly (20%) |
Kelly bet |
Over-bettor (40%) |
Over-bettor bet |
Under-bettor (10%) |
Under-bettor bet |
|
1 |
Win |
₹ 1,20,000 |
₹ 20,000 |
₹ 1,40,000 |
₹ 40,000 |
₹ 1,10,000 |
₹ 10,000 |
|
2 |
Loss |
₹ 96,000 |
₹ 24,000 |
₹ 84,000 |
₹ 56,000 |
₹ 99,000 |
₹ 11,000 |
|
3 |
Win |
₹ 1,15,200 |
₹ 19,200 |
₹ 1,17,600 |
₹ 33,600 |
₹ 1,08,900 |
₹ 9,900 |
|
4 |
Win |
₹ 1,38,240 |
₹ 23,040 |
₹ 1,64,640 |
₹ 47,040 |
₹ 1,19,790 |
₹ 10,890 |
|
5 |
Loss |
₹ 1,10,592 |
₹ 27,648 |
₹ 98,784 |
₹ 65,856 |
₹ 1,07,811 |
₹ 11,979 |
|
6 |
Win |
₹ 1,32,710 |
₹ 22,118 |
₹ 1,38,298 |
₹ 39,514 |
₹ 1,18,592 |
₹ 10,781 |
|
7 |
Loss |
₹ 1,06,168 |
₹ 26,542 |
₹ 82,979 |
₹ 55,319 |
₹ 1,06,733 |
₹ 11,859 |
|
8 |
Win |
₹ 1,27,402 |
₹ 21,234 |
₹ 1,16,171 |
₹ 33,192 |
₹ 1,17,406 |
₹ 10,673 |
|
9 |
Win |
₹ 1,52,882 |
₹ 25,480 |
₹ 1,62,639 |
₹ 46,468 |
₹ 1,29,147 |
₹ 11,741 |
|
10 |
Loss |
₹ 1,22,306 |
₹ 30,576 |
₹ 97,583 |
₹ 65,056 |
₹ 1,16,232 |
₹ 12,915 |
|
Final |
|
₹ 1,22,306 |
|
₹ 97,583 |
|
₹ 1,16,232 |
|
Sit with those final numbers for a moment.
All three investors had the same coin — a genuine 60% edge. All three experienced the identical sequence of wins and losses. The only difference was how much they put on the table each round.
Kelly ends at ₹1,22,306. A healthy gain.
The over-bettor ends at ₹97,583. He actually lost money — despite winning 6 out of 10 flips with a biased coin in his favour. You can see how it happened: rounds 2 and 5 hit him hard because he had bet 40% of a now-larger base each time. The losses ate faster than the wins could rebuild.
The under-bettor ends at ₹1,16,232. He never faced a scary loss, but the timid 10% bet meant his wins never compounded to their full potential. He left money on the table quietly, round after round.
Notice also that the over-bettor led after rounds 1, 3, and 4. For a while, his aggression looked like brilliance. This is exactly what happens in a bull market — over-concentration feels like skill, right up until it doesn't. In this particular sequence, the over-bettor lost money despite having an edge. One of Kelly's surprising findings is that over betting is punished more severely than under betting. Betting twice the Kelly amount can reduce long-term growth to zero, even when you have a genuine edge
The bet size is not a minor detail. It is the whole game.
From Gambling Tables to Your Portfolio
The formula translates directly to investing. Bet size is simply how much of your total portfolio you allocate to one stock or instrument.
Say you are considering a stock. You genuinely believe there is a 55% chance it delivers meaningful gains, and a 45% chance it disappoints. Your expected upside is roughly equal to your downside — for every rupee at risk, you stand to gain a rupee. Real investments are far more complex than coin flips because gains and losses are not fixed amounts. Investors often use Kelly as a framework for thinking about position size rather than applying the formula literally,
Kelly says: 0.55 − (0.45 ÷ 1) = 0.10 → allocate about 10% of your portfolio.
Not 30% because you are excited about it. Not 5% because you are nervous. 10%, because that is what your honest assessment of the edge actually supports.
The number almost always comes out smaller than your gut suggests. That is not a flaw in the formula. That is the lesson.
The Honest Caveat
Kelly requires you to know your probability of being right. In a coin flip, that is calculable. In investing, it never truly is. Your estimate will always carry uncertainty — and if your estimate is off, your bet size will be off too.
This is why most serious practitioners use fractional Kelly — betting half of what the formula suggests. You sacrifice some theoretical optimum, but you gain protection against your own estimation errors.
A simple rule of thumb: whatever Kelly tells you, consider using half.
The Money Vichara Reflection
Think about your current portfolio — not the what, but the how much.
Is there a position significantly larger than others — not because you have more edge, but because you have more excitement?
Is there a position very small — not because you have less conviction, but because you are holding back from habit or fear?
The table above showed something important. The over-bettor did not lose because he picked the wrong coin. He lost because he ignored the question of how much to bet. The under-bettor did not fail dramatically — he just quietly under-earned, year after year.
Both are costly. Both are common. And both are avoidable — not by finding better investments, but by thinking more honestly about how much of yourself you are putting into each one.
Kelly assumes repeated opportunities with a measurable edge. Most retail investors do not possess a reliably measurable edge in individual stocks. For many investors, Kelly's greatest value is not as a formula to calculate, but as a reminder that position sizing deserves as much thought as stock selection
The question is not just "Am I in the right investments?"
It is also: "Am I invested in the right proportions?"
That second question, asked honestly and regularly, is where a great deal of wealth is quietly built — or quietly lost.
That is the real vichara.

Comments
Post a Comment