October 12, 2026 · Free trading journal
There is a version of this story that would be easy to write.
Thirty-six trading bots. Forty-two completed trades. Twenty-two winners. Two standout systems with a 75% win rate.
Add a confident headline about beating a falling market, point toward the next opportunity, and publish.
But that would leave out the part that matters most.
These were paper trades, not verified live profits. The numbers were displayed session counters, not a reconciled full-day ledger. Some market data was approximately 40–48 minutes old. And different parts of the dashboard disagreed about profit and loss.
That makes this a story about building a down-day trading process—not a claim that I verified the market’s closing direction or finished the day with money in my pocket.
Still, there was something worth studying.
Two systems recorded six wins against two losses apiece. A short-momentum bot recorded two wins from three completed trades. Other systems struggled. Twenty-eight reported no completed trades at all, although some still showed open positions.
The snapshot gave me a reason to be curious, but not a reason to celebrate prematurely.
My question became:
What can I learn from the winners without telling myself a story the evidence doesn’t support?
That question led me through the bot results, into the attached futures report, and eventually toward a more useful approach to forecasting potential profit.
Not a promise of bigger returns.
A plan for finding out whether an opportunity is real.
1. The number I wanted to see—and the number I needed to understand
The first number that caught my attention was 75%.
Two bots reached it:
bot_es_micro_trend— September 28 version: six wins, two losses.bot_mes_fed_momentum— September 16 version: six wins, two losses.
Eight completed trades each. Sixteen combined. Twelve winners.
On a dashboard full of competing strategies, those two records stand out.
My first instinct was to ask how to get more of that.
Could I give the leaders more room? Could I run similar configurations? Could I identify what they had done differently and repeat it?
But the snapshot could not answer those questions.
It did not show the complete entry-and-exit sequence. It did not establish whether the winners were large or small. It did not tell me whether the systems took the same trades at nearly the same time.
And a name containing “trend” or “Fed momentum” did not prove which market condition produced a particular result.
That distinction changed my reading of the leaderboard.
I could say the two bots led by number of reported wins.
I could not say they led by verified net profit, execution quality, or risk-adjusted performance.
I also could not say their 75% win rates would continue.
For me, that became the first useful lesson of the session:
An encouraging result earns a closer investigation. It does not automatically earn more capital.
I wanted the winners to be meaningful. The next step was to give them a fair test rather than a flattering explanation.
2. The full scoreboard, including the parts that don’t make a great headline
Here is the completed-trade snapshot I am working from.
Bot / version Completed trades Wins Losses Win rate ES micro trend — Sept. 28 8 6 2 75.0% MES Fed momentum — Sept. 16 8 6 2 75.0% MES Fed hawkish momentum — Sept. 16 10 4 6 40.0% MCL long — Sept. 23, generation 2 6 3 3 50.0% MES momentum short — root folder 3 2 1 66.7% ES micro momentum — root folder 3 1 2 33.3% ES micro trend — Oct. 2 2 0 2 0.0% MES mean reversion, generation 2 — Sept. 29 2 0 2 0.0% Total 42 22 20 52.4%
The folder dates identify versions, not the dates when these trades occurred.
Across the full snapshot, eight of the 36 bot instances reported completed trades. The remaining 28 reported zero wins and zero losses.
Their win rates were undefined, not zero percent.
That matters because “did not complete a trade” and “completed only losing trades” are entirely different observations.
The overall win-to-loss count ratio was 1.10:1. That describes the number of wins relative to losses. It is not a profit factor or a reward-to-risk ratio.
I want that distinction visible because it is easy to make the scoreboard sound more impressive than it is.
Twenty-two wins gave me useful material.
Twenty losses gave me equally useful material.
Neither group told me the whole financial story by itself.
3. What the two leaders actually contributed
The two leading bots produced 12 of the fleet’s 22 reported wins.
That is approximately 54.5% of all wins, from approximately 38.1% of completed trades.
Remove those two systems from the snapshot, and the remaining active bots recorded:
26 completed trades;
10 wins;
16 losses;
a combined win rate of approximately 38.5%.
That is a noticeable difference.
It also creates a temptation: keep the leaders, switch off everything else, and imagine the cleaner result I could have had.
I do not want to do that.
Selecting the winners after seeing the results would turn this session into an exercise in hindsight. It would not show that I could have identified them before the trades occurred.
Instead, I want to preserve the observation and test it forward.
Were these two systems responding to different opportunities, or were they effectively making the same bet?
Did they keep their losing trades small?
Did their reported wins survive a realistic allowance for costs?
Did their strongest trades occur when the data feed was current?
Those questions matter more to me than the ranking.
There is also a version-control lesson here. The September 28 trend bot recorded six wins and two losses, while the October 2 version recorded two losses and no wins.
The newer date did not correspond to the better displayed result.
That does not establish that the newer version is worse. Two trades cannot carry that conclusion.
It does remind me to evaluate the change itself rather than assume that newer means improved.
For now, the leaders have earned priority in my research queue.
Nothing more—and nothing less.
4. The short bot that made me think differently about a down day
The short-momentum bot recorded two wins and one loss.
Three trades are not enough to declare an edge. But the result belongs in a conversation about down-day trading because it reminds me not to build every strategy around waiting for prices to recover.
A short futures position can benefit when the contract price falls and lose when it rises. That is the basic directional mechanism—not a guarantee that shorting a weak market will be profitable. (cmegroup.com)
The important limitation is that my snapshot does not include the actual fills.
I cannot reconstruct a precise sell signal, failed rally, or breakdown entry from a bot name and three outcomes.
I can only identify what deserves further examination.
For a future paper test, I would want to distinguish between two very different short entries:
A preplanned signal that appears after a rally fails.
A late entry taken simply because the screen already looks frightening.
Those are not equivalent decisions in my research process.
My down-day plan would require a defined trigger, a defined reason to exit, and fresh data before the simulated entry.
I would also write down what invalidates the bearish idea.
If the market recovers the level my setup depends on, I do not want the bot—or my interpretation of it—to keep arguing with the recovery.
The useful idea is not “short everything on a red day.”
It is:
Give bearish setups a place in the process, but demand the same evidence and risk discipline that I would demand from a bullish setup.
The short bot offered an interesting starting point. The ledger still has to explain what happened.
5. The losing bots deserve more than a quick dismissal
The busiest bot in the snapshot was the September 16 hawkish-momentum version.
It completed ten trades: four wins and six losses.
Another momentum system finished one win and two losses. Two additional bots recorded two losses each without a win.
Those records are less attractive than six-and-two.
But I do not want to use win counts to declare them financially unsuccessful when the average gains and losses are missing.
Consider an invented example—not a description of these bots.
Four wins averaging two risk units each would produce eight units of gross gain. Six losses averaging one unit each would produce six units of gross loss.
That would leave two units before costs, despite a 40% win rate.
Change the winning-trade average to half a unit and the same four-and-six record becomes a very different result.
This is why I want the losing systems examined rather than hidden.
For each one, I want to know:
Did the entry follow the intended rules?
Was the data current?
Did the exit behave as designed?
Were repeated entries responding to the same underlying move?
How large were the losses relative to the winners?
Was the system operating in a condition it was designed to trade?
I am not looking for excuses.
I am trying to separate a losing trade from a broken process.
A valid strategy can have a losing observation. An invalid implementation can produce a winning observation.
My job is to determine which problem—or success—I am actually looking at.
The scoreboard tells me where to begin. It does not finish the diagnosis.
6. Twenty-eight quiet bots—and a lesson about what silence means
Twenty-eight bot instances reported no completed trades.
At first glance, that sounds like a lot of unused machinery.
But the snapshot gives me no basis for calling all of them disciplined, inactive, broken, or unproductive.
Some showed open positions.
Others may have been waiting for an entry condition. Some may have encountered a data or configuration issue. I do not know without inspecting their state.
That means I need a more useful classification than “zero wins.”
For the next review, I want every bot assigned to a clear category:
Flat and waiting for a valid signal.
Holding an open position.
Blocked by a risk limit.
Blocked by stale or missing data.
Experiencing an execution or configuration problem.
Finished for the session.
Each category leads to a different action.
A bot that correctly rejected an invalid setup should not be pushed to trade merely to improve its activity count.
A bot holding a position needs exposure monitoring, even if it has not closed anything.
A bot with a stale feed needs technical attention, not a more aggressive entry rule.
That is an operational lesson I can apply immediately.
I do not want to reward activity for its own sake.
But I also do not want to romanticize inactivity. “It stayed out of trouble” is not an explanation if I have not checked whether it was functioning.
The 28 quiet bots are not a footnote to the session.
They are an unfinished section of the review.
Before asking how to produce more winning trades, I need to understand what the majority of the fleet was actually doing.
7. The detail that interrupted the victory story
The market-data age was approximately 40–48 minutes for bots displaying data.
That is the detail I cannot move into fine print.
The processes could be alive. The heartbeat timestamps could look healthy. The console could keep updating.
None of that established that the price information was current.
For my intended intraday testing, that distinction is fundamental.
I cannot confidently interpret a signal as timely without knowing the age of the information that produced it.
At the same time, I should not overstate the finding. A stale-data snapshot does not prove that every earlier completed trade used stale prices.
It means I need the market-data timestamp associated with each individual decision.
That is a testable question.
My next review would connect:
Market event → received timestamp → signal timestamp → simulated order → simulated fill → exit.
If those records line up, I can examine the trade.
If they do not, I need to repair or qualify the result before including it in a performance claim.
The CFTC specifically cautions that hypothetical trading assumes executions at particular prices and may misrepresent performance when those assumptions have not faced actual market conditions. That warning belongs close to a paper-trading leaderboard, not buried beneath it. (cftc.gov)
My proposed fix is straightforward: make data freshness an explicit condition for new simulated entries, with a threshold appropriate to the strategy.
I also want a visible stale-data status and a log entry explaining every blocked trade.
I would rather discover fewer valid trades than publish more questionable ones.
8. Three P&L readings—and no reason to pick the nicest one
The accounting problem was just as important.
For PID 36236, the supplied summary showed:
Realized P&L: $0.00
A result-log Running Total: $14.62
A console header showing R −$4.86 / U −$4.86
I cannot combine those figures responsibly without understanding their definitions.
They may refer to different periods, scopes, or accounting states. One may have reset. Another may include a position the first excludes.
The snapshot does not settle the matter.
What I will not do is select $14.62 because it supports the most attractive article.
Nor will I select the negative header and declare the whole fleet unprofitable.
Both conclusions would outrun the evidence.
My next accounting task is to reconstruct results from the underlying trade records:
Identify each unique completed trade.
Match its entry and exit.
Confirm quantity and contract specification.
Calculate gross simulated P&L.
Apply the stated cost assumptions.
Separate closed results from open-position marks.
Reconcile the result with every dashboard field.
I also want the session boundaries documented.
“Today,” “since startup,” and “since the last reset” need not describe the same interval.
Until that work is complete, I can publish the displayed win-and-loss counts with their limitations.
I cannot publish a verified net-profit figure.
That feels less satisfying than a clean green number.
But I would rather have an unresolved accounting question in a paper account than carry the same ambiguity into a live one.
The most useful result of this session may be identifying exactly what the dashboard still needs to explain.
9. Why 52.4% cannot forecast profit by itself
Once I stopped treating the win rate as the answer, I could ask the more useful question:
What is the average financial outcome per trade after costs?
CME’s trading-education material explains expectancy in terms of winning frequency, losing frequency, and the average size of each outcome. A strategy can win less often than it loses and still have positive expectancy if its payoff structure supports it. (cmegroup.com)
For this journal, I would write the calculation as:
Net expectancy = probability of a win × average gross win − probability of a loss × average gross loss − average trading cost.
The snapshot supplies an observed win frequency.
It does not supply trustworthy values for the other inputs.
I therefore cannot calculate the fleet’s verified expectancy.
I can, however, show how sensitive the answer is to assumptions.
Suppose I define 1R as a fixed illustrative risk unit. This is a teaching device, not the actual risk taken by the bots.
Suppose every losing trade averages 1R, and each completed trade incurs an illustrative cost of 0.08R.
Holding the reported 22 wins and 20 losses constant gives:
Hypothetical average winner Gross result Costs across 42 trades Net result 0.80R −2.40R 3.36R −5.76R 1.00R +2.00R 3.36R −1.36R 1.20R +6.40R 3.36R +3.04R
Same win count. Same loss count. Three different financial outcomes.
These are arithmetic scenarios—not measured performance, probability estimates, or forecasts.
They show why I need the missing information before talking about profit.
10. My forecast starts with conditions, not a dollar target
The user-friendly forecast would be a number.
A daily target. A weekly estimate. A monthly projection built by multiplying one promising session.
I do not have the evidence for that.
What I can offer is a conditional outlook.
If the reported win frequency remains broadly similar in a much larger, properly recorded sample; if average winners are sufficiently large relative to average losses; and if costs do not consume the margin, then the strategy could produce positive net results.
Those are meaningful conditions, not small qualifications.
Using the illustrative assumptions from the previous section, 22 winners would need to average slightly more than 1.06R to cover 20 losses of 1R plus 3.36R in total costs.
That threshold comes from dividing 23.36R by 22.
It is not a target the real bots have demonstrated.
It tells me what I would need to measure.
My route toward potential improvement therefore has four parts:
Verify the price data.
Reconcile the accounting.
Measure the payoff distribution.
Test changes on subsequent observations rather than rewriting the past.
I also want the possibility of failure stated plainly.
The next sample could have fewer wins. Losing trades could be larger. Simulated fills could prove too favorable. The apparent leaders could lose their advantage.
A useful forecast should survive those possibilities being mentioned.
My working outlook is cautiously constructive about the research opportunity, not bullish about an unverified profit stream.
There are enough encouraging observations to justify continuing the experiment.
There are not enough trustworthy financial measurements to justify extrapolating income.
That is where I stand today.
11. Then I opened the futures report—and found both ideas and contradictions
The attached Futures Trading Strategy Report, generated October 12, 2026, covers a wide range of potential drivers:
Interest rates and central-bank policy.
Equity-index momentum.
Oil supply risks.
Gold and uncertainty.
Currency strength.
Volatility and cross-asset positioning.
That breadth is useful for generating questions.
But I cannot treat the document as a verified market feed.
For example, it places December gold futures around $4,230 on page 8, $2,450 on page 12, and $4,200 on page 17, without explaining how those widely different figures fit one dated analysis.
It also presents inconsistent policy-rate assumptions and several conflicting market readings.
On page 15, the stated 10-year yield of 5.30% and two-year yield of 5.10% imply a positive 20-basis-point ten-year-minus-two-year spread—not the negative spread described there.
Those are reasons to pause before converting the prose into orders.
The repeated references to December 2026 Eurodollar futures are another warning. CME documented the conversion of eligible Eurodollar futures and options into SOFR equivalents in April 2023. A 2026 report cannot simply carry the old trade examples forward as though that transition did not happen. (cmegroup.com)
My response is not to discard every idea.
It is to separate themes worth researching from numbers and instruments requiring verification.
I can use the report to ask what might matter next.
I cannot use its confident tone as evidence that its prices, institutional-positioning claims, or trade labels are correct.
12. The first research scenario: persistent equity weakness
The report repeatedly links a hawkish policy outlook with pressure on equity markets.
I am treating that as a scenario to test—not as a verified description of October 12.
My first proposed down-day study is simple:
What happens when the index stays weak after an attempted recovery?
For a future paper test, I would define the setup before observing the outcome.
That means choosing a consistent time window, specifying the level that represents a recovery attempt, and deciding what counts as a failure.
I would then compare how the short-momentum system and the leading trend system behave under those same conditions.
Do they enter together?
Does one enter earlier but suffer more reversals?
Does another wait for confirmation but surrender too much of the move?
Those are practical questions the summary cannot answer.
I would also define the opposite case: a successful recovery that invalidates the bearish setup.
A down-day process should not require me to remain bearish after its own conditions disappear.
The goal is not to find language that explains every outcome after the fact.
The goal is to establish a decision rule whose successes and failures can both be counted honestly.
The two reported wins from the short bot make this a reasonable place to investigate.
They do not prove that failed rallies were the source of those wins.
That distinction will stay in the journal.
If the next properly recorded sample supports the idea, I can continue testing it. If it does not, I will have learned something specific instead of merely discovering that a dramatic market narrative sounded convincing.
13. The second research scenario: oil strength while equities struggle
The report’s energy sections repeatedly raise the possibility of supply concerns supporting oil.
Again, I am not accepting its stated prices or geopolitical claims as independently verified facts.
I am taking the broader scenario and turning it into a research question:
Could an oil strategy provide a different opportunity while equity-index strategies face weakness?
The active MCL long bot recorded three wins and three losses across six completed trades.
That is not confirmation.
It is a small, mixed result that deserves inspection.
My next paper study would ask whether its entries were associated with sustained strength or merely with temporary price spikes.
I would want to compare the winning and losing trades by:
Entry timing.
Size of the move before entry.
Time held.
Adverse movement after entry.
Exit behavior.
Data freshness.
I would also examine what the equity bots were doing at those same times.
I do not want to label the oil system a diversifier just because it trades a different instrument.
In this account, diversification needs to show up in the exposure and results, not merely in the bot names.
The report also contains inconsistent spread descriptions, so I would verify contract months, units, and spread direction before attempting any calendar-spread study.
For now, my interest is narrower.
Can the MCL bot identify a repeatable long opportunity with acceptable losses when the expected continuation does not happen?
If the answer is yes, that could become part of a broader down-day toolkit.
If the answer is no, an appealing oil narrative will not rescue it.
14. The third research scenario: uncertainty without a clear direction
Not every uncomfortable session gives me a clean directional setup.
The futures report proposes gold exposure and volatility hedges in several places, but it also supplies conflicting price levels and volatility readings.
That makes me reluctant to borrow its strikes, targets, or numerical risk rules.
Instead, I take a more modest lesson from it:
Uncertainty can be a reason to reduce activity, not just a reason to add another trade.
For my own paper process, this becomes a separate regime to study.
What happens when price repeatedly reverses, the trend systems disagree, and the short setup fails to follow through?
My proposed response is not an automatic switch into gold, options, or volatility futures.
It is to mark the condition clearly and measure whether the existing systems should participate at all.
I would compare results during clean directional periods with results during repeated reversals.
If a strategy’s apparent strength disappears in the latter condition, I want that limitation documented.
The report’s fixed percentage reductions tied to volatility thresholds are not rules I can adopt simply because they are printed in an institutional-looking document.
I would need to test their relevance to the particular instrument and strategy.
There is a personal discipline here as well.
When one approach stops producing understandable results, I do not want to compensate by adding unfamiliar instruments until something works.
My next potential profit opportunity may come from a better filter.
It may also come from deciding that a particular environment is outside the system’s useful range.
Either conclusion is valuable if the evidence supports it.
15. The hedge I would not add just to make the article sound sophisticated
The report contains many options structures: put spreads, call spreads, risk reversals, straddles, and more.
That vocabulary can make a plan sound complete.
But my snapshot does not establish that these bots traded those structures, that the required contracts were verified, or that I have measured their cost and behavior.
I will not retrofit an options success story onto a futures-bot leaderboard.
There is also an important difference between a directional position and a hedge.
A hedge has to be evaluated against the exposure it is intended to offset.
For example, a purchased put provides downside exposure, but it also requires a premium; the buyer can lose that premium. Selling a put creates a different obligation and potentially a long futures position if exercised. Those are not interchangeable ways to “get paid for protection.” (cmegroup.com)
For this research process, I would first identify the actual exposure.
Am I trying to offset an existing long position?
Am I trying to profit from a decline?
Am I trying to reduce the impact of a sudden move?
Those are different objectives.
I would then examine the proposed instrument, premium, expiration, and worst-case outcome before adding it to the paper study.
The same caution applies to the report’s suggestions about volatility.
I do not want to assume that a trade becomes protective merely because “VIX” appears in its name.
For now, the simplest improvement is to understand the positions already present.
Complexity can wait until it solves a measured problem.
16. My next experiment needs rules that I cannot rewrite afterward
The snapshot has given me several hypotheses.
The two leaders may be worth prioritizing.
The short-momentum system may deserve a focused down-day study.
The MCL long system may offer a separate opportunity under particular conditions.
Now I need an experiment that can disappoint me honestly.
I would freeze the relevant configurations and document any later changes.
I would keep the same accounting conventions across systems.
I would label market conditions using rules chosen before reviewing the final results.
Most importantly, I would retain the weaker outcomes alongside the stronger ones.
NFA’s discussion of hypothetical performance warns about the role of hindsight and the limits of simulations, including their inability to fully reproduce financial risk and execution conditions. That is particularly relevant when selecting attractive systems after observing their results. (nfa.futures.org)
My next scorecard would therefore include more than win rate:
Measurement What I want it to establish Net simulated P&L Whether gains remain after stated costs Average winner and loser The payoff structure Largest loss and drawdown The severity of adverse outcomes Data age at each decision Whether signals used timely information Concurrent exposure Whether bots duplicated the same risk Rule adherence Whether the implementation matched the design Performance by condition Where the strategy worked or struggled
I am not setting a magical trade-count threshold and calling everything beyond it proven.
I want a broader sample across different conditions, with uncertainty still visible.
The aim is not to graduate the most exciting bot as quickly as possible.
It is to discover whether the promising result survives a fairer test.
17. Where I think potential improvement could actually come from
After reviewing the snapshot and the report, I see three possible routes toward better results.
They are hypotheses, not demonstrated improvements.
First: remove invalid observations
If some decisions were made using unsuitable data, I need to identify them, repair the cause, and repeat the test.
That may make the historical scoreboard less attractive.
I am willing to accept that.
A smaller set of interpretable results is more useful to me than a larger set whose timing I cannot verify.
Second: understand the losses
I want to know whether losing trades reflect normal strategy behavior or avoidable implementation problems.
Repeated entries, stale signals, incorrect resets, or misunderstood position states would require different fixes from a valid signal that simply did not work.
I do not want to change the strategy when the accounting is broken.
Nor do I want to blame the accounting when the strategy is weak.
Third: improve selection without pretending hindsight was foresight
The leaders deserve prospective attention.
But I would evaluate them under fixed rules rather than repeatedly promoting whichever bot most recently looks best.
I would also check whether their gains overlap.
Two successful systems could still represent one concentrated exposure if they enter similar positions simultaneously.
My forecast for potential improvement rests on those practical questions.
I am not predicting that the next session will produce more wins.
I am predicting that, if I complete this work, I should have a better basis for deciding whether the wins matter.
That may sound modest.
For a trading system, better evidence is a substantial improvement.
Without it, a larger position would merely make an unresolved question more expensive.
18. What I want the next journal entry to say
I would like the next post to contain a clean profit figure.
But only if I can explain where it came from.
I want to be able to say that the trade ledger reconciled with the dashboard, that every decision used data of an acceptable age, and that simulated costs were applied consistently.
I want to show the losing trades without treating them as an embarrassment.
I want to explain why a bot entered, why it exited, and whether those actions matched its rules.
If the leading systems continue to perform well, I want that progress documented.
If they fall back toward the rest of the fleet, I want that documented too.
That is the kind of trading story I want to publish for free: not a sequence of carefully selected screenshots, but a record readers can follow as the evidence changes.
The attached futures report can contribute ideas.
It cannot supply missing fills, repair contradictory prices, or establish my account’s profitability.
The bot counters can identify patterns.
They cannot replace the ledger.
And a 75% win rate can be encouraging without becoming a promise.
That is the balance I am trying to keep.
I want enough optimism to continue the work and enough skepticism to know when a result has not earned my confidence.
A down-day trading plan needs both.
It needs the imagination to recognize opportunity outside a rising market.
It also needs the discipline to leave an opportunity alone when the data, execution, or accounting cannot support it.
The takeaway: I found winners. Now I have to find out what they are worth.
October 12 gave me 42 reported completed paper trades.
Twenty-two were labeled wins. Twenty were labeled losses.
Two bots led with six wins and two losses each. A short-momentum bot offered another interesting result. The remaining systems showed a mixture of losses, balanced outcomes, open positions, and no completed activity.
That is the honest starting point.
I am encouraged by the leaders.
I am interested in the report’s scenarios around persistent equity weakness, possible oil strength, and changing uncertainty.
But my outlook for potential profit remains conditional.
Before forecasting income, I need trustworthy prices, reconciled results, realistic costs, and a broader prospective test.
Simulation is useful, but it is not equivalent to live trading. NFA explicitly cautions that hypothetical records cannot fully capture the experience of financial risk or assure similar future outcomes. (nfa.futures.org)
So the next step is not to turn the size up.
It is to turn the quality of the evidence up.
That means verifying the leaders rather than worshipping them. Investigating the losers rather than hiding them. Understanding the quiet bots rather than counting them as failures. And using the futures report as a source of questions rather than a substitute for current, verified information.
The headline result was 22 paper wins.
The more important result was a clearer understanding of what I still need to know.
I found winning trades in the snapshot. Now the work is to determine whether the process behind them can produce repeatable, net-positive results.
That is the story I plan to keep telling—on the encouraging days, the disappointing days, and the days when the most useful decision is to stop and fix the system.
This post discusses simulated trading and educational research. The supplied statistics are session counters, not an independently verified full-day ledger. No verified net profit is claimed. All payoff scenarios are hypothetical, and future losses remain possible. The attached report contains inconsistencies; its market levels, instrument references, and institutional claims should not be treated as verified trading instructions.
Subscribe free to follow the next review: fresh data, reconciled results, and an honest look at whether the leaders can repeat.