Same Setup, Crypto vs Stocks

This video takes one setup — built mostly from what this channel has already measured — and runs it on three stocks and three coins, over the same years.

The plan was to show what changes in the arithmetic when you carry a method across markets. That part is here.

But the setup lost money. On both markets, and the best setting I tried only broke even. And how I found that out is worth more than the comparison. By the end you'll be able to check any rule you own for the three accounting errors that made this one look profitable.

Start with the setup, stated completely, because a setup you cannot state completely cannot be measured.

Entry: price returns to an untested level — a turn that has not been touched since it formed — within three daily ranges of price.

Stop: half a daily range beyond the level. Target: two stops away. Time exit: ten days.

Untested came from the ranking video, the distance filter from the markup video. The trade itself is new here: a target two stops away, a ten-day exit, and a stop of half a daily range — five times the tenth used earlier in this module, because this trade is held for days, not hours.

First run: sixty five per cent of trades won, and the average trade made plus zero point nine two of an R.

That number is wrong, and here is how it was wrong.

The simulation only started checking the stop on the day after entry. Every trade where the stop was hit on the entry day itself quietly survived to the next morning.

Fix that, and check both the stop and the target on the entry day. Result: seventy five per cent winners, plus one point two two R.

Better, and even more wrong.

Because a bar gives four numbers and no order. If price reached the target on the same day it touched the level, the likely sequence is that it was at the target first and touched the level afterwards.

So the entry day counts asymmetrically: the stop can fire on it, the target cannot. Worst case only. That is what the bar can honestly support.

Result: sixty two per cent, plus zero point eight five of an R. Still too good.

Two more gifts to myself were hiding in the entry.

The touch was counted with a tolerance of a quarter of a daily range. So the trade filled at the level price even when the market never got there.

And the approach direction was never checked. A sale at resistance was counted even when price had been trading above that level for weeks — when it was not resistance at all.

Remove both. Fill only if price actually reached the line, and only if it approached from the correct side.

Twenty eight to thirty per cent winners on the three stocks. Twenty six to twenty seven on the coins.

Average trade: minus zero point one eight of an R on Bitcoin, minus zero point two two on Ethereum. Every instrument negative.

Four versions. Plus zero point nine two, plus one point two two, plus zero point eight five, and then minus zero point one seven. Nothing about the market changed between them. Only the honesty of the accounting.

That is the most useful thing in this video, and it is not a fact about crypto.

Now, since the setup is measured, the comparison it was built for is still worth making.

First: does a wider stop rescue it? Half a range, one, one and a half, two.

On Apple: minus zero point one one, minus zero point zero seven, minus zero point one, and minus zero point one again. On Bitcoin: minus zero point one eight, minus zero point two three, minus zero point one two, minus zero point zero seven.

Wider is not steadily better: Apple is best at one range, Bitcoin at two. Not one of the twenty four settings ends above zero, even before costs. There is no stop distance at which this becomes a business.

Second: what does not change between the two markets.

The share of winners: twenty eight to thirty per cent on stocks, twenty six to twenty seven on crypto. Close enough to call the same.

The sign of the result: negative on all six at half a range. And a wider stop helps on both only at the far end — at two ranges every instrument loses less than at half a range, and none of them becomes a business.

So the method transfers. Unfortunately, so does its unprofitability.

Third: what does change, and this is the part worth writing down.

The stop as a percentage of price. On the index fund it is zero point four one per cent. On Solana, two point eight.

At the same one per cent risk, divided by that stop, that means a position of two hundred and forty five per cent of the account on the index fund, and thirty six per cent on Solana. On the index fund that is more than the whole account.

Seven times the difference, from the same rule, on the same day, with the same risk.

Which is the practical point of this whole video: the rule is portable, the number it produces is not.

Fourth: frequency. The coins produced fifty eight to sixty two trades a year, the stocks thirty six to forty two.

Not because crypto has more setups per day, but because it has three hundred and sixty five trading days against two hundred and fifty two.

The same strategy gives you about fifty per cent more trades on crypto, for the same rules.

And fifth, the one that surprised me: costs.

A round trip costs about zero point zero five of a per cent on a stock and zero point two zero on crypto — four times more, and that is the number everybody quotes.

But what matters is the cost in R, and that depends on how close the stop is.

On the index fund, with a stop of zero point four one per cent, a round trip eats zero point one two two of an R. On Tesla, with a stop of one point eight, it eats zero point zero two eight.

Four times more expensive — on the same market, with the same commission.

Bitcoin, with the four times higher fee but a wider stop, comes out at zero point one two one. Almost exactly the same as the index fund.

So the expensive thing is not crypto. It is a tight stop, wherever you place it.

Now three honest limits, and the first one is large.

This is daily bars and one setup. A losing result here does not mean levels do not work; it means this particular mechanical version of them does not, at this timeframe, over this window.

Two: the fills are optimistic even now. A real order at a level does not always get filled, and slippage on a stop is worse than the number I used.

Three: two years and nine months, six instruments. Enough to see a sign, not enough to date it.

What all this leaves you with is short.

Rules carry across markets. Numbers do not: measure the stop in per cent of price on the instrument you actually trade, and let position size follow from it.

Cost is not a property of the market. It is the ratio of your fee to your stop, and you control the second one.

And when a backtest of your own idea comes out positive — a test on past prices, like this whole video — spend an hour trying to break it before you spend a dollar trading it.

So, what to go and do. It takes an evening, and it is the most valuable hour in this module.

Take any rule you believe in and write it out completely enough that a stranger could run it. Then check three things: does the stop fire on the entry bar, could the target have been hit before your entry, and did price actually reach your fill.

Most home-made edges live entirely inside those three questions.

The method travels. The profit did not. Educational content only. Nothing here is financial advice.