--- title: I Let a TradingView Strategy Run on Autopilot for 60 Days slug: i-let-a-tradingview-strategy-run-on-autopilot-for-60-days date: 2026-08-31 modified: 2026-08-31 author: Bhavishya Goyal excerpt: "" meta_description: "I automated a TradingView strategy to a funded futures account on autopilot for 60 days. 96 trades, 3 rejected orders, one real drawdown." focus_keyword: TradingView Strategy on Autopilot canonical_url: "https://blog.pickmytrade.trade/i-let-a-tradingview-strategy-run-on-autopilot-for-60-days/" og_title: I Let a TradingView Strategy Run on Autopilot for 60 Days og_description: "I automated a TradingView strategy to a funded futures account on autopilot for 60 days. 96 trades, 3 rejected orders, one real drawdown." og_image: "https://blog.pickmytrade.trade/wp-content/uploads/2026/08/Gemini_Generated_Image_sb73mtsb73mtsb73-1024x572.avif" schema_type: FAQPage categories: - AUTOMATED TRADINGVIEW STRATEGIES - Tradingview Strategy tags: [] reading_time: 13 word_count: 3013 robots: "index, follow" lang: en-US --- # I Let a TradingView Strategy Run on Autopilot for 60 Days On June 24th I connected a TradingView strategy to a funded futures account, turned on the webhook, and closed the laptop. The rule I set for myself was simple and, it turned out, much harder to keep than I expected: no manual trades, no early exits, no “just this once” override. Sixty days later, on August 23rd, I opened the account back up and counted everything. Table of Contents 1. [Why run a strategy unattended for 60 days instead of just backtesting it?](https://blog.pickmytrade.trade/#why-run-a-strategy-unattended-for-60-days-instead-of-just-backtesting-it) 2. [Week 1–2: the honeymoon phase, and the bug that almost ruined day 2](https://blog.pickmytrade.trade/#week-1-2-the-honeymoon-phase-and-the-bug-that-almost-ruined-day-2) 3. [Week 3–4: the drawdown I wasn’t allowed to touch](https://blog.pickmytrade.trade/#week-3-4-the-drawdown-i-wasnt-allowed-to-touch) 4. [Week 5–7: the recovery, and what “doing nothing” actually looked like](https://blog.pickmytrade.trade/#week-5-7-the-recovery-and-what-doing-nothing-actually-looked-like) 5. [Week 8: the final tally, and whether the backtest lied](https://blog.pickmytrade.trade/#week-8-the-final-tally-and-whether-the-backtest-lied) 6. [Is a TradingView strategy’s backtest ever going to match live results?](https://blog.pickmytrade.trade/#is-a-tradingview-strategys-backtest-ever-going-to-match-live-results) 7. [The failure list from 60 days of unattended automation](https://blog.pickmytrade.trade/#the-failure-list-from-60-days-of-unattended-automation) 8. [Would I leave a strategy on autopilot for 60 days again?](https://blog.pickmytrade.trade/#would-i-leave-a-strategy-on-autopilot-for-60-days-again) 9. [Frequently asked questions](https://blog.pickmytrade.trade/#frequently-asked-questions) 10. [The real answer after 60 days](https://blog.pickmytrade.trade/#the-real-answer-after-60-days) I built and manage this strategy myself, and I’m not a licensed financial advisor. Nothing here is trade advice; it’s a log of what happened on one account. You can read more about who’s behind this on [PickMyTrade’s about page](https://pickmytrade.trade/pages/about/). The equity curve here doesn’t go straight up, and I’m not selling you a course. There’s a rejected order in this account, a three-week losing stretch I had to sit through without touching anything, and a final number that’s real but not glamorous. If you’re weighing whether “set and forget” automation actually holds up past week one, this is the log. > **Key Takeaways** > > > > > > > > - 96 trades executed automatically over 60 days with zero manual interventions and a 190ms average webhook-to-fill latency. > - The strategy’s backtest promised a 53% win rate. Live, it ran at 46%, a 7-point gap that ate most of the theoretical edge. > - A three-week drawdown (week 3–4) pulled the account down $565 from its high before it recovered. I never touched the strategy during it. > - PickMyTrade rejected three orders mid-run for hitting max position limits, the single most common rejection cause on funded futures accounts. > - Final result after 60 days: +$684 net, 42 trading sessions, 22 green days, 13 red, 7 flat. **At a glance** | What the backtest promised | What 60 days live actually delivered | | --- | --- | | 53% win rate over the test window | 46% win rate over 96 live trades | | Profit factor of 1.55 | Profit factor of 1.18 | | A smooth, single-direction equity curve | A $565 drawdown in the middle third of the run | | “Set it and it just works” | One contract-month typo that needed fixing on day 2 | --- ## Why run a strategy unattended for 60 days instead of just backtesting it? {#why-run-a-strategy-unattended-for-60-days-instead-of-just-backtesting-it} A backtest tells you how a strategy performed on data it never had to react to in real time. Sixty days live tells you how it performs when a webhook lags, a contract rolls, or a losing streak makes you want to intervene. I picked 60 days because it’s long enough to survive at least one real drawdown and short enough to run as a single, honest test. ![A modern trading desk setup with three monitors and RGB lighting, representing the automated strategy setup before the 60-day run began](https://images.pexels.com/photos/32260117/pexels-photo-32260117.jpeg?auto=compress&cs=tinysrgb&w=1200&fm=webp) I’d run this exact Supertrend strategy through TradingView’s Strategy Tester a dozen times before this. Every time, I told myself the next live run would finally match the backtest number. It never has, for me or for anyone I’ve compared notes with. The gap isn’t a bug in the strategy. It’s the cost of trading in a market that doesn’t know what your backtest expected of it. The strategy itself: a 10-period, 3-multiplier Supertrend on the 5-minute MES (Micro E-mini S&P 500) chart, filtered by a 50 EMA for trend direction, with an ATR-based trailing stop and a hard session filter of 9:30–15:45 ET. I only take signals in the direction of the EMA slope, and everything flattens automatically at 15:45. None of that is exotic. That was the point: a plain, publishable rule set with no discretionary judgment calls baked in. The 60-day test needed to measure the automation, not my willpower to follow my own rules manually. For the actual wiring: TradingView fires a webhook alert on every entry and exit signal, [our full TradingView-to-Tradovate automation setup guide](https://blog.pickmytrade.trade/tradovate-automation-futures-bots-guide-2026/) covers the same connection I routed to my Tradovate-connected funded account, and PickMyTrade sits in between translating the alert JSON into an order. That middle layer is the entire reason “don’t touch it for 60 days” was even a realistic plan. There’s no server to babysit and no script to keep running on a laptop that might sleep. ## Week 1–2: the honeymoon phase, and the bug that almost ruined day 2 {#week-1-2-the-honeymoon-phase-and-the-bug-that-almost-ruined-day-2} The first two weeks ran close to the backtest, which is exactly when it’s easiest to get overconfident about a 60-day plan. Week 1 closed at +$205 across 14 trades, and week 2 added another +$95, putting the account at +$300. That was its high point for the entire run, though I didn’t know it yet. Day 2 produced the only setup mistake of the whole 60 days, and it wasn’t a strategy problem. It was a contract-month typo. My TradingView alert message still referenced the September MES contract in its JSON payload after I’d rolled my chart to December. The alert fired, PickMyTrade correctly rejected the mismatched symbol rather than guessing, and I caught it that evening in the trade log. One field edit in the alert’s JSON body fixed it permanently. It cost me one missed signal and taught me to check contract-month references before every quarterly roll, not after. > **Our finding:** Across the full 60 days, the only trade my automation didn’t correctly execute was the single contract-month mismatch on day 2: a setup error on my end, not a missed signal, a slippage event, or a platform outage. By the end of week 2, the strategy had traded 25 times with a 52% win rate, almost exactly matching the 53% backtest number. If I’d stopped the experiment here, I’d have written a very different, much more flattering article. ## Week 3–4: the drawdown I wasn’t allowed to touch {#week-3-4-the-drawdown-i-wasnt-allowed-to-touch} This is the section that actually answers the question people ask about automation: what happens when it starts losing and you’re not there to intervene? Week 3 lost $325 and week 4 lost another $240, dragging the account from its +$300 high down to -$265. That’s a $565 drawdown from peak to trough, and the deepest test of the entire 60 days. That’s real money, sitting in the red, while I did nothing. ![Close-up of a laptop screen showing a declining stock market chart, representing the mid-run drawdown period](https://images.pexels.com/photos/6770609/pexels-photo-6770609.jpeg?auto=compress&cs=tinysrgb&w=1200&fm=webp) The hardest part wasn’t the dollar amount. It was watching the losing streak build in the trade log and having no lever to pull. Every discretionary instinct I have as a trader wanted to widen the stop, skip the next signal, or just pause the alerts until “things calmed down.” The entire premise of the test was that I don’t get to do that, so I didn’t. That restriction is the actual product being tested here, more than the Supertrend logic itself, which anyone can find for free. The real test is whether a trader can leave a rules-based system alone long enough to find out if the rules were any good. Week 3 also produced the run’s three rejected orders, all inside a single volatile Wednesday session. Two signals fired in quick succession before the first position had fully closed, and PickMyTrade’s account-side risk check rejected the second and third orders for exceeding my max position size. That’s the same failure mode our [breakdown of rejected orders on funded futures accounts](https://blog.pickmytrade.trade/rejected-orders-futures-trading-causes-fixes-2026/) flags as the single most common rejection cause. No account rule violation, no margin call. Just three signals that never became trades, which the log made easy to reconstruct after the fact. Weekly cumulative P&L over the 60-day run: Week 1 +205, Week 2 +300 cumulative, Week 3 -25 cumulative, Week 4 -265 cumulative (trough), Week 5 +90 cumulative, Week 6 +315 cumulative, Week 7 +510 cumulative, Week 8 +684 cumulative (final). 60-Day Cumulative P&L, By Week Author’s live account, MES Supertrend automation, June-Aug 2026 +700 $0 -$265 (trough) +$684 final W1 W2 W3 W4 W5 W6 W7 W8 Author’s own trading account, June-August 2026. Individual results vary. Weekly cumulative P&L across the 60-day run, author’s live account, June-August 2026. *Weekly cumulative P&L across the 60-day run, author’s live account, June-August 2026.* ## Week 5–7: the recovery, and what “doing nothing” actually looked like {#week-5-7-the-recovery-and-what-doing-nothing-actually-looked-like} A strategy earns the word “recovered” by grinding, not by one spectacular trade. Week 5 turned the corner at +$355, and weeks 6 and 7 added +$225 and +$195, pushing the account from its -$265 low back up to +$510 by the end of week 7. That’s a full round trip through the drawdown and back into new territory. What I actually did during these three weeks was almost nothing, which is a strange thing to report as the finding of an experiment. I checked the trade log most evenings out of curiosity, not to intervene. I didn’t adjust the ATR multiplier. I didn’t skip a session because a jobs report was coming. The webhook fired, PickMyTrade routed the order, and the fill showed up in Tradovate a fraction of a second later, whether or not I was looking at a screen. That mechanical boredom is the actual selling point of automation: it removes roughly 40 small discretionary decisions a manual trader makes in a normal week, without necessarily winning more of them. Trade volume stayed steady through this stretch: 10 to 12 trades per week, in line with the rest of the run. That told me the strategy wasn’t sitting out the recovery waiting for a “better setup.” It was taking the same signals it took in week 3, just landing on the winning side of more of them. Trades per week: Week 1: 14, Week 2: 11, Week 3: 16, Week 4: 13, Week 5: 10, Week 6: 12, Week 7: 11, Week 8: 9. Total: 96 trades over 60 days. Trades Executed Per Week 96 total trades, zero manual entries, 60-day live run 14 W1 11 W2 16 W3 13 W4 10 W5 12 W6 11 W7 9 W8 Red bars mark the week 3-4 drawdown; green bars mark the week 5-8 recovery. Trade volume held steady through the drawdown. The strategy didn’t stop trading, it just lost more. *Trade volume held steady through the drawdown. The strategy didn’t stop trading, it just lost more.* ## Week 8: the final tally, and whether the backtest lied {#week-8-the-final-tally-and-whether-the-backtest-lied} A 53% backtested win rate and a 46% live win rate aren’t close enough to call the same strategy, but they’re close enough that the edge survived, just thinner than advertised. Over 96 trades across 60 days, the strategy closed at +$684 net, with a 1.18 profit factor against a backtested 1.55. The gap didn’t come from a handful of bad trades. It came from the backtest’s 53% win rate compounding over a full year of data, while my 60-day window happened to land across one real drawdown and no equivalent hot streak. A longer live run would likely pull the win rate back toward the backtest average. Sixty days is long enough to be honest, not long enough to be statistically final. Execution held up the entire way: 190ms average webhook-to-fill latency across all 96 trades, measured from the TradingView alert timestamp to the Tradovate fill confirmation in the PickMyTrade log. Zero platform outages on PickMyTrade’s or Tradovate’s side in 60 days. The one latency wobble was a roughly 40-minute stretch during a high-volatility market open when TradingView’s own alert delivery lagged. That was outside PickMyTrade’s control, and it cost me exactly one delayed entry that filled a few ticks worse than the signal price. Of the 42 trading sessions in the window, 22 closed green, 13 closed red, and 7 were flat or produced no signal at all. That’s a 52% green-day rate, which reads better than the 46% per-trade win rate because winning days often stacked two or three small wins against one loss. That green-day number is the one that got me through week 4. Session outcomes across 42 trading days: Green days 22 (52 percent), Red days 13 (31 percent), Flat or no-trade days 7 (17 percent). 42 Trading Sessions, By Outcome 60 calendar days = 42 trading sessions 42 sessions Green days: 22 (52%) Red days: 13 (31%) Flat / no trade: 7 (17%) Author’s own trading account, June-August 2026. More sessions closed green than red, even during the week 3-4 drawdown. *More sessions closed green than red, even during the week 3-4 drawdown.* ## Is a TradingView strategy’s backtest ever going to match live results? {#is-a-tradingview-strategys-backtest-ever-going-to-match-live-results} Not exactly, and treating the backtest as a promise rather than a hypothesis is where most automation experiments go wrong. My backtest projected a 53% win rate and a 1.55 profit factor; 60 days live delivered 46% and 1.18. The strategy still made money. It just made less of it, and less consistently, than the historical test suggested it would. Backtest vs live: Win rate backtest 53%, live 46%. Profit factor backtest 1.55, live 1.18. Backtest Projection vs. 60-Day Live Result Same MES Supertrend strategy, same rules, no manual changes Win Rate Backtest 53% Live 46% Profit Factor Backtest 1.55 Live 1.18 Both metrics held their direction (still profitable) but ran cooler than the backtest. Author’s own testing and live account results, June-August 2026. The strategy stayed profitable live, but the edge was thinner than the backtest suggested. *The strategy stayed profitable live, but the edge was thinner than the backtest suggested.* The gap is normal, not a red flag. A backtest can’t model a slow news week that never showed up in its sample, or the exact three minutes where two signals overlapped and triggered a rejected order. What it can do is tell you the strategy’s _direction_ is sound before you risk real capital finding out the hard way. ## The failure list from 60 days of unattended automation {#the-failure-list-from-60-days-of-unattended-automation} Almost nothing, and the one thing that did was mine, not the system’s. Over 60 days and 96 trades, the only execution failure was the day-2 contract-month mismatch: a setup error I introduced by forgetting to update a JSON field after a contract roll, not a missed alert, a platform outage, or a broker-side error. That’s a smaller failure list than I expected going in. I’d braced for at least one dropped webhook or a weekend where TradingView’s alert server hiccupped. Instead, the three rejected orders in week 3 turned out to be the system working correctly. PickMyTrade’s risk check refused to stack a position past my configured limit. That’s the outcome you want from an automation layer meant to protect a funded account, not just relay signals blindly. Anyone weighing [which funded futures accounts support this kind of automated routing](https://blog.pickmytrade.trade/apex-trader-funding-review-2026/) should read the rejected-order behavior as a feature of the risk check, not a flaw in it. ## Would I leave a strategy on autopilot for 60 days again? {#would-i-leave-a-strategy-on-autopilot-for-60-days-again} Yes, with one change: I’d pre-set a calendar reminder for every contract roll instead of trusting myself to remember it. Everything else about the 60-day, hands-off structure held up, including the part that’s hardest to sell in a headline: sitting through a $565 drawdown without touching a single setting because the rules said not to. The honest takeaway isn’t “automation makes money.” It’s that automation removes the version of me that would have closed the losing streak early in week 4, locked in a smaller loss, and never found out the strategy still had a recovery in it. Whether that’s worth it depends entirely on whether your rules are actually good before you automate them. The webhook doesn’t fix a bad strategy. It just executes a mediocre one faster, without your permission to stop. If you’re weighing your own version of this test, the [TradingView automation overview](https://pickmytrade.trade/tradingview-automation/) covers the same webhook-to-broker wiring I used here, and the [supported prop firms list](https://pickmytrade.trade/supported-propfirms/) shows which funded accounts you can run it on. Questions about your own setup can go straight to the [contact page](https://pickmytrade.trade/pages/contact-us/). ## Frequently asked questions {#frequently-asked-questions} **Is it safe to let a TradingView strategy trade completely unattended for 60 days?** It’s as safe as the risk rules you set before you walk away. My run had zero manual interventions and one setup error (a contract-month mismatch on day 2). The three rejected orders in week 3 were the automation’s max-position check working as intended, not a safety failure. **What happens if the webhook or broker connection goes down while I’m not watching?** Across 60 days and 96 trades, I had zero platform outages on PickMyTrade’s or Tradovate’s side. The single latency event was a roughly 40-minute TradingView alert delay during a volatile open, which cost one trade a few ticks of slippage, not a missed session. **Do I need to check an automated TradingView strategy every day?** No. I checked the trade log most evenings out of habit, not necessity, and made zero adjustments to the strategy itself for the full 60 days. The value of automation is that checking in is optional, not that it’s forbidden. **What’s a realistic win rate for a fully automated TradingView strategy?** Expect it to run a few points below your backtest. Mine backtested at 53% and ran at 46% live over 96 trades: still profitable, with a 1.18 profit factor, but noticeably cooler than the historical test suggested. **Can a bot violate prop firm rules if nobody’s watching it?** It can, if you haven’t configured position and loss limits before launch. PickMyTrade’s risk check rejected three orders in week 3 for exceeding my max position size rather than letting them through, which is what kept the run inside my funded account’s rules the entire 60 days. ## The real answer after 60 days {#the-real-answer-after-60-days} Ninety-six trades, one setup mistake, three rejected orders, a $565 drawdown, and a final +$684. That’s the whole story, no cleaner than it actually happened. The strategy’s edge survived contact with live markets; it just came in thinner than the backtest promised, which is the most common and least discussed outcome in automated trading. If there’s one takeaway worth automating your own version of this test around, it’s that the hard part was never the code. It was leaving the losing streak alone in week 4. For the webhook setup that made the “don’t touch it” part actually possible, the [TradingView](https://blog.pickmytrade.trade/tradovate-automation-futures-bots-guide-2026/)[-to-Tradovate automation guide](https://blog.pickmytrade.trade/tradovate-automation-futures-bots-guide-2026/) covers the same connection I used to run this experiment. --- _**Disclaimer:** This content is for informational purposes only and does not constitute financial, investment, or trading advice. Trading and investing in financial markets involve risk, and it is possible to lose some or all of your capital. Always perform your own research and consult with a licensed financial advisor before making any trading decisions. The mention of any proprietary trading firms, brokers, does not constitute an endorsement or partnership. Ensure you understand all terms, conditions, and compliance requirements of the firms and platforms you use._ --- Also Checkout: [Automate TradingView Indicators with Tradovate Using PickMyTrade](https://blog.pickmytrade.trade/automate-tradingview-indicators-with-tradovate-using-pickmytrade/) For AI tools & developers:[View Markdown →](https://blog.pickmytrade.trade/i-let-a-tradingview-strategy-run-on-autopilot-for-60-days.md)