The last-day gamble and the bad start

Two moments in an evaluation produce more account-ending decisions than the rest combined: the day you believe is your last chance, and the first week when it goes wrong immediately. Both have arithmetic, and the arithmetic disagrees with the instinct almost every time.

The last-day gamble

Setup: a 30-day evaluation, day 30, you are at $1,500 of a $3,000 target, $1,000 daily limit, $200 risk per trade. The instinct is to size up to $500 and "go for it".

First question: is it really the last day? Many futures evaluations have no time limit at all, so the "last day" is one you invented. Several forex firms extend the window free or cheaply if the account is in profit at expiry. If either applies, there is no gamble to take; you keep trading the plan.

Second question, if it really is the last day and expiry means a reset fee: what does the gamble cost and what does it pay?

Plan	Needs	Rough probability (55% win, 1:1)	Outcome if it fails
Trade normally, $200 risk	8 net winning trades in one day	Under 1%	Expiry; buy reset, account history intact
Size up to $500	3 straight wins before 2 losses	About 17% for three straight; roughly 25% allowing for orders	Daily limit breach; buy reset
Size to $1,000 (max)	2 straight wins, no losses	About 30%	Daily limit breach on first loss; buy reset

The gamble does raise the probability of passing today from near zero to 25% or 30%. In that narrow sense it is rational: expiry and breach cost the same reset fee. But two things sit outside the table.

Some firms treat a breach differently from an expiry. A breach may cost a full new evaluation rather than a cheaper reset, may reset your "trading days" count, or may be logged as a risk-rule violation that affects funded-account trust later. Check.
It trains the behaviour that fails funded accounts. The funded account has no last day, but it will have days that feel like one, and the trader who has practised sizing to the daily limit under pressure will do it again with a payout on the line.

The plan that removes the dilemma is not to arrive at day 30 needing $1,500. Module 3's trades-needed table tells you, on day 1, whether 30 days is realistic at your expectancy. If it is not, the correct move was to choose a firm without a time limit or to keep building the log.

Key idea: A last-day gamble is a small edge in pass probability bought with the habit that ends funded accounts. Choose evaluations where the last day does not exist, and if you find yourself there anyway, accept the reset rather than practise the breach.

The bad start

Setup: "$50,000" account, $3,000 target, $2,000 drawdown (static or locked for simplicity), and in the first week you lose $1,200. You have $800 of room and need $4,200 of profit. The ratio, which started at 1.5, is now 4,200 / 800 = 5.25.

What are the options?

Grind at small size. Drop to $50 risk per trade. Target is now 84R, room is 16R. From the gambler's-ruin formula:

Win rate (1:1)	P(pass) at $50 risk from here	Expected trades to +84R
50%	16%	Never in expectation
52%	About 72%	Over 2,000 at a 1:1 payoff; about 840 at 0.1R
55%	About 96%	About 840 at a 1:1 payoff; about 420 at 0.2R

With a real edge and no time limit, the grind can work, but look at the trade counts. 400 to 800 trades at 2 to 3 a day is six months to a year on one evaluation, most of it spent one bad streak from failure.

Reset. Pay the reset fee (often $80 to $150) and start again with $2,000 of room and a 1.5 ratio. For the same 52% trader at $100 risk, the fresh account's pass probability is about 81% (Module 3) in roughly 150 trades.

Option	P(pass), 52% trader	Trades	Cash cost
Grind, $50 risk	About 72%	800+	$0
Reset, $100 risk	About 81%	About 150	Reset fee

For most traders with an edge, the reset is the better purchase: higher probability, a fifth of the time, for the price of a modest fee. For a trader without an edge, neither option is good, and the honest use of the bad start is as evidence.

Cut size but do not reset yet. A reasonable middle path if the loss is smaller (say $600 of $2,000). Halve the risk, keep the plan, and set a decision point: if room falls below 50% of the original allowance, reset.

How to tell variance from a broken plan

Before resetting, read the log for the bad week:

Finding	Read	Action
All losses were valid setups at planned size	Variance	Reset, same plan
One or more losses were outside the setup or oversized	Discipline	Reset; add the rule that would have stopped it
Losses were valid but the average loser was bigger than logged	Costs or slippage changed	Reset; re-measure the loser before restarting
Win rate over 40+ evaluation trades is far below the log	Edge may be gone or never was	Do not reset; go back to sim and re-measure

Try it: Compute the ratio (target + loss so far) / remaining room for your evaluation at three points: 25%, 50% and 75% of the allowance lost. Decide now, in writing, at which of those you will reset rather than grind. That decision made calmly is worth more than any made on the day.

Recap

Most "last days" are self-imposed or extendable; where a real one exists, a sized-up gamble buys a 25% to 30% chance at the cost of a habit that fails funded accounts.
After a bad start, recompute the ratio from where you stand; at 5:1 the grind takes hundreds of trades even with an edge.
A reset restores the original ratio for a modest fee and usually offers a higher pass probability in far fewer trades.
Read the bad week's log first: variance means reset the same plan; broken discipline or a missing edge means fix that before paying again.