Last Updated: October 25, 2026
A DCA bot configuration is an order plan, not a return forecast. Before connecting an account, write down what opens a deal, how much each additional order can spend, when the plan stops and how you will reconcile an unknown order state. This guide explains those decisions with invented spot examples; it does not document personal trading results or recommend a live configuration.
Averaging a trade is different from scheduled investing
A scheduled purchase allocates a chosen amount at calendar intervals. A long averaging bot adds to an existing position when specified conditions are met. Lower purchase prices can reduce the average cost per unit while increasing the amount exposed. A falling market can exhaust every planned order and continue falling.
For an arithmetic example, buy $100 at a price of $100, then $200 at $80. You hold 3.5 units bought for $300, so the average cost is $85.7143, before fees. At a market price of $70 those units are worth $245: the unrealized loss is $55, despite the lower average entry. A 1% target above average cost would be about $86.5714 before costs; neither recovery nor execution at that price is assured.
Make two separate tables: money and price levels
Order-size scaling and spacing are different inputs. Write the first averaging order as order 1, so its exponent is zero.
order k amount = first averaging amount × volume scale^(k − 1)
planned capital = base amount + sum(all averaging amounts)
This invented ladder preserves the original guide's useful size example. It assumes quote-currency sizing, six additional orders and no leverage. Displayed cents are rounded; calculate totals from unrounded values and then apply the venue's actual rounding rules.
Scroll sideways to see every column. Keyboard: focus the table region and use the arrow keys.
| Order | Amount | Cumulative planned spend |
|---|---|---|
| Base | $30.00 | $30.00 |
| Averaging 1 | $30.00 | $60.00 |
| Averaging 2 | $39.00 | $99.00 |
| Averaging 3 | $50.70 | $149.70 |
| Averaging 4 | $65.91 | $215.61 |
| Averaging 5 | $85.683 | $301.293 |
| Averaging 6 | $111.3879 | $412.6809 |
With a scale of 2 instead, the sixth additional order is 32 times the first: the sequence is 1, 2, 4, 8, 16, 32. It is not 64 times. This is why the number of orders matters as much as the first order's size.
For a separate illustration, a $25 base plus five orders starting at $30 and scaled by 1.2 requires:
25 + 30 × (1 + 1.2 + 1.44 + 1.728 + 2.0736) = $248.248
That is about $248.25 before fees and rounding, not the $213 quoted in an earlier edition. Three simultaneous deals with this same ladder would require $744.744 before those additions. A balance used by another bot or locked in an open order is not spare funding.
Spacing: distinguish the next gap from total deviation
For an illustrative cumulative ladder with a first gap of 1.5% and a gap multiplier of 1.2, sum the gaps rather than treating each gap as the total decline:
Scroll sideways to see every column. Keyboard: focus the table region and use the arrow keys.
| Additional order | Next gap, percentage points | Cumulative deviation from initial reference |
|---|---|---|
| 1 | 1.5 | 1.5% |
| 2 | 1.8 | 3.3% |
| 3 | 2.16 | 5.46% |
| 4 | 2.592 | 8.052% |
| 5 | 3.1104 | 11.1624% |
These are arithmetic inputs, not recommended market levels. Check the platform's preview for the selected averaging mode and trigger reference; a custom ladder or indicator condition need not use this simple model. Do not confuse a gap measured from the preceding order with a percentage measured from the original price or a changing average entry.
Prepare a configuration record before activation
Scroll sideways to see every column. Keyboard: focus the table region and use the arrow keys.
| Decision | Record explicitly | Failure to check |
|---|---|---|
| Account and instrument | Eligible account, exact spot pair, quote currency, minimum size and size increment | Wrong product, rejected order or unintended leverage |
| Start condition | Immediate, scheduled or specified signal; timestamp and expiry | A default immediate entry or a stale repeated signal |
| Capacity | Full ladder × maximum concurrent deals, plus costs and other commitments | Later orders cannot be funded |
| Exit | Target reference, fees, order type and treatment of partial fills | A displayed percentage is mistaken for net profit |
| Stop rule | Trigger reference, cancellation behavior and what remains open | Stopping the bot is mistaken for selling all holdings |
| Unknown state | Stable order ID, reconciliation method and pause condition | A timeout is retried as a second order |
The current 3Commas interface documentation describes multiple sizing and averaging modes. It also says saved bot changes apply to new trades, while existing trades need separate attention. Compare the preview and open-deal settings with your written record instead of assuming a saved form changed every position. Documentation checked October 25, 2026; this is not a funded platform test.
Connection permissions are a limit, not insurance
Use the provider's current connection instructions for the exact exchange and account type. A dedicated integration key makes revocation and audit easier. Grant only the permissions needed by the intended workflow; do not grant withdrawals for an order-management task. Use supported restrictions and verify the documented service addresses rather than guessing an IP allowlist.
Disabled withdrawals do not make funds safe from all losses. Trading access can still create unwanted exposure or costly trades. Exchange custody, account compromise, liquidation on leveraged products and integration failures are separate risks. Do not put keys into screenshots, prompts or shared test fixtures. This guide does not certify any provider's security.
Evaluate the whole account, including unfinished deals
A low take-profit target can close some trades sooner, but it does not establish better returns: costs, fill frequency, inventory and time under water all matter. Trailing exits can reverse or slip. A stop trigger is not a guaranteed execution price. No fixed weekly or monthly return is supported by the examples here.
In a paper study, preserve the input data and configuration, include rejected and partial orders, charge the assumed fees and software costs, and value open inventory at the end. Compare the same dates, starting capital and cash flows with a stated baseline. Keep the period used to choose settings separate from the period used to assess them. A simulator can check mechanics while still modeling real liquidity poorly.
Useful failure cases include zero available funds after an earlier order, the entire ladder filling before a further decline, two signals arriving together, a response timeout followed by a late fill, a target reached only on an untradeable quote, and a test ending with an open losing position. A positive sum of closed-deal profits does not settle any of those questions.
What to verify on the provider's site
Check current subscription limits, supported exchange/product, account eligibility and any simulation feature directly before paying or connecting. The previous plan-price table and recommended portfolio split were not backed by current evidence and are not used here. The worksheet works without choosing a paid plan or placing a trade.
Affiliate disclosure: 3Commas provider link. This is an affiliate link: AI Trading Ranked may earn a commission from qualifying activity. Compensation is a commercial conflict to consider, not evidence of testing or performance. You can instead navigate directly to the provider.
Risk warning: Educational information, not individualized financial advice. All numerical cases above are constructed illustrations; capital committed to trading can be lost.