Trading a Falling Crypto Market: A 4-Bot Case Study
Bitcoin is down 28% on the year and the Fear index reads 27. One trader ran four bots through it: three COMBO Shorts at 10x and one unleveraged spot DCA, +$8,678 on $7,581 committed. Here's the setup, bot by bot, and where it would have failed.
Trading bots can profit in a falling market when they are configured for short exposure. Over 8 to 10 days of live trading in July 2026, one trader ran three COMBO Short bots at 10x leverage alongside one unleveraged spot DCA bot, committing $7,581.63 in total and finishing the month at $8,678.33 in realized return — a gain of 114.46%. The leveraged short side produced 98% of that.
What a COMBO bot is
A COMBO bot combines DCA and GRID logic inside a single leveraged futures position. DCA orders average the entry price when the market moves against the position; GRID orders take profit in slices when it moves in the intended direction; a trailing stop-loss follows the move once the position is in profit. Direction — long or short — is set at launch.
Market conditions during the test
Updated 3 August 2026. Bitcoin opened August near $63,000, roughly 28% lower year to date, with most large-cap altcoins down 32% to 44% on the year. The Crypto Fear & Greed Index read 27. The FOMC held rates unchanged, the CLARITY Act stalled in the US Senate, and spot Bitcoin ETF flows turned negative again in the closing days of July. The test ran through the second half of that month.
How the test was set up
Four bots, launched for diversification and to see how the systems behave under opposite conditions.
One spot DCA bot, long, no leverage. This one answers a long-horizon question: how does unleveraged accumulation perform when left alone for months rather than days? It buys on a schedule, averages the entry price, and has no liquidation price attached. Slow by design.
Three COMBO bots, short, 10x leverage. These test the other end — how quickly the bots can generate profit on short and medium-term horizons, and how much drawdown they absorb while doing it. Short direction, because the market was already falling and the bots were configured for the move in front of them rather than a reversal someone was hoping for.
Splitting the account this way is the strategy. One portion carries leverage and volatility, one portion carries neither, and neither depends on the other being right.
Results by side
| Side | Bots | Invested | Bot profit | Avg. daily |
|---|---|---|---|---|
| Spot DCA, Long, no leverage | 1 | $2,172.42 | +$160.82 (+7.40%) | 0.83% |
| Futures COMBO, Short, 10x | 3 | $5,409.21 | +$8,517.51 (+157.46%) | 11.43% |
| Total | 4 | $7,581.63 | +$8,678.33 (+114.46%) |
The DCA bot did what it was built to do: under 1% a day, no leverage, nothing to liquidate.
The three COMBO Shorts absorbed partial drawdowns and produced the rest. Individually, WLD returned +283.86% over 10 days 18 hours, SUI returned +101.32% over 8 days 20 hours on a much smaller base, and HYPE returned +41.88%.
Why the short side worked
Returns like these come from trade frequency rather than from one correct call about where the market ends up. A COMBO Short accumulates its position by averaging the entry price upward as price rallies against it, then sells back into the position partially and proportionally each time price falls. The bot places many trades a day. A trader making the same decisions manually would need to be awake for all of them.
Three things follow, and they are the part worth copying rather than the number:
Leverage is a risk dial, not a setting to max out. This test used 10x deliberately, to find out what the bots would survive. The same configuration at 3x or 5x produces a smaller result and a far wider margin before an adverse move approaches liquidation. That is the trade most traders should be making.
A leveraged COMBO is a short-horizon tool. One to three weeks is the practical window, after which the bot is closed and the profit locked in rather than left open to keep deciding the month.
The unleveraged side is not decoration. Under 1% a day looks unremarkable next to 11%, and it is the portion of the account that cannot be liquidated while the leveraged portion is being tested.
Test three COMBO bots without a subscription
The trader in this case was running a $100 PRO plan. Testing the same bot type does not require one.
EVEDEX connections allow up to three COMBO bots on the Free plan — the same count used on the short side here. The exchange is non-custodial and requires no KYC, and the connection uses a trade-only API key that cannot withdraw funds.
What this result does not prove
A leveraged short book in a month when large-cap altcoins fell 32–44% is a strategy meeting its ideal conditions. The same four bots configured the same way into a sustained rally would have lost money, and the DCA logic that averages into an adverse move would have kept committing reserve margin while doing it. At 10x cross margin, losses can exceed the initial margin.
The distribution matters as much as the total. Remove WLD from this account and the remaining three bots made roughly $1,580 instead of $8,678 — still a good ten days, and a completely different headline. Leveraged directional bots do not spread returns evenly; a small number catch the move properly and the rest do somewhere between nothing and a little. Planning around the average of four bots means planning for a result that no individual bot produced.
What the setup bought was exposure to the chance that one of them would run, at position sizes small enough that no single bot could take the balance down, with an unleveraged reserve that survives either way.
Running this configuration yourself
Choose direction from the market in front of you rather than the one you want; there is a full walkthrough of that decision if that is the step where you stall. Backtest the configuration against recent price history for the pair, then run it in demo mode against live prices with virtual funds before committing capital. If you are allocating across several bots at once, the multi-bot sizing framework covers how to stop them from quietly becoming the same bet. The step-by-step EVEDEX setup guide covers API key scope, margin and leverage fields, and the backtest step.
Three COMBO bots on the Free plan
Connect EVEDEX with a trade-only API key and launch up to three COMBO bots. Backtest first, demo second, capital third.
FAQ
Can trading bots make money when crypto is falling? Yes, when configured for short exposure. A COMBO Short opens a leveraged short, averages the entry upward if price rallies against it, and takes profit in slices via GRID orders as price declines. In this case, three COMBO Shorts returned 157.46% on $5,409.21 of margin over 8 to 10 days of trading in July 2026.
What leverage should a COMBO Short use? This test used 10x deliberately, to measure drawdown tolerance. Lowering leverage to 3x or 5x reduces both the return and the speed at which an adverse move approaches the liquidation price.
How long should a leveraged COMBO bot run? One to three weeks is the practical window, closing the bot and locking in profit at the end rather than leaving a leveraged position open indefinitely.
How many COMBO bots can I run on a free plan? Up to three on an EVEDEX connection. Backtest and demo mode are available before any live position.
Is running a long and a short at the same time a hedge? Not automatically. Four bots across four different assets are four independent positions, each carrying its own risk. The unleveraged spot DCA here was a survivability allocation, not an offset.
Are these results typical? No. This is one account over one month, in a market where large-cap altcoins fell 32–44%. Individual results vary widely, and the same configuration in a rising market would have produced losses.
