Why the broker matters more than usual
LongRunner makes one long trade a day on a stock index. Each trade earns a small amount, and in our MetaTrader 5 tests the profit factor of the profitable runs was only 1.03 to 1.12. At that size of edge, trading costs are not a side note – they decide whether a market makes money at all.
The clearest example is the Dow Jones. With the same EA, the same times and real ticks, it was profitable at BlackBull (profit factor 1.12, Aug 2020 – Sep 2026) and lost at RoboForex (profit factor 0.97, Dec 2017 – Jul 2026). Our working assumption is the spread right after the daily open at 01:05 on RoboForex's Dow symbol; a test with a later start time is still open, so that is not proven. The Nasdaq-100, on the other hand, held in every test.
What to check before you choose
1. The index spread at 01:05 and 23:50 server time
The advertised "from" spread or a daily average says little here. What counts is the spread at the two moments LongRunner trades: right after the daily session opens (01:05 on the US sets) and shortly before the close (23:50). We do not publish index spreads – our cost data only covers EUR/USD and gold – so check your broker yourself:
- Watch the spread in Market Watch or in the footer of LongRunner's dashboard at 01:05 server time on a few days.
- Run the set file in the Strategy Tester with "Every tick based on real ticks" on your broker's data. That model uses the broker's recorded spreads, so the result already includes them.
- Then set "Max spread at entry". With a limit, LongRunner waits for the spread to come down until the latest entry time instead of buying into a wide spread.
Our broker reviews list the FX commission of the raw accounts. For index CFDs, check in the contract specifications whether a commission is charged on top of the spread.
2. Server time UTC+2 / UTC+3
The set files assume a server on UTC+2 in winter and UTC+3 in summer, so that the New York close falls on midnight. IC Markets, BlackBull and RoboForex run such servers, as do many other brokers. With a different server time, shift the start and close time by the difference. On start, LongRunner compares both times with the symbol's trading sessions and warns in the log if one lies outside.
3. Hedging account
A hedging account is recommended. On a netting account all EAs and manual trades share one position per symbol; LongRunner then skips the day rather than buying into someone else's position. All three brokers below allow EAs, scalping and hedging without restriction according to our reviews.
4. Symbol names
The set files are written per market, not per symbol name. The NAS100 set goes on NAS100, USTEC or US100, the US30 set on US30, DJ30 or WS30, the DE40 set on DE40, GER40 or DAX40. Some brokers add prefixes or suffixes – in our tests at RoboForex the symbols were called .USTECHCash and .US30Cash.
5. What you do not need
- No high leverage. A position worth 0.5 to 0.7 times the balance is an effective leverage below 1. Retail leverage limits do not get in the way.
- No swap-free account. With the close time on, LongRunner is flat before every rollover and pays no swap.
- No co-location. One market order a day at a fixed time is not latency-sensitive. A stable connection at 01:05 and 23:50 matters more than milliseconds.
The brokers we actually tested
How this order is builteditorial · partner · as of 04 October 2026
- Selection: the 3 brokers this article compares (BlackBull Markets, RoboForex, IC Markets) — an editorial selection by the article’s topic.
- Pick — partner influence: yes: No ranking: the order follows our tests – first the two brokers we ran LongRunner’s MT5 real-tick backtests on, then the broker of the live cross-check account. All three are partner brokers (active affiliate link), so this selection is not independent of those partnerships, and the linked reviews contain partner links.
- Rating — partner influence: no: Weighted average of Costs 20%, Execution 20%, Algo suitability 20%, Regulation 15%, Conditions 15%, Deposits & support 10%, the same for every broker. Commissions do not count, and no rating is raised for partners. Methodology →
- As of: 04 October 2026
| Broker | What we ran | Market (symbol) | Result |
|---|---|---|---|
| BlackBull | MT5 real ticks, 08/2020–09/2026 | Dow Jones (US30) | every day 0.67 ×: profit factor 1.12, +36.7% · trend filter 0.7 ×: 1.03, +4.7% |
| BlackBull | MT5 real ticks, 10/2025–09/2026 | DAX (GER40) | every day 0.67 ×: profit factor 1.12, +5.5% – one year only |
| BlackBull | demo forward test since 1 Oct 2026 | Dow Jones (US30), M15 | running, too short for a result |
| RoboForex | MT5 real ticks, 12/2017–07/2026 | Nasdaq-100 (.USTECHCash) | trend filter 0.5 ×: profit factor 1.10, +38.8% · every day 0.67 ×: 1.05, +41.0% |
| RoboForex | MT5 real ticks, 12/2017–07/2026 | Dow Jones (.US30Cash) | every day 0.5 ×: profit factor 0.97, −11.5% |
| RoboForex | MT5 real ticks, 12/2017–07/2026 | S&P 500 (.US500Cash) | profit factor 0.98–1.00 – no edge, hence no set file |
| IC Markets | live account, 12 Feb – 30 Sep 2026, another EA with the same daily idea | Nasdaq-100, Dow, DAX (USTEC, US30, DE40) | profit factor lot-adjusted 1.28 / 1.11 / 1.07 · in money 1.22 / 1.01 / 0.99 |
For readers in the EU: BlackBull and RoboForex Ltd are not EU-regulated, so an account there comes without EU investor protection (details below). IC Markets has an EU entity regulated by CySEC – but the live account in the table runs at its international entity Raw Trading Ltd, and we have no LongRunner backtest on the EU entity's data, so run the set file in your own Strategy Tester with real ticks first.
BlackBull Markets
BlackBull is where the Dow made money in our tests and where LongRunner has been running as a forward test on a demo account since 1 October 2026 (US30, M15) – the dashboard screenshots come from there. The Dow test with the trend filter shows the filter's price in a period without a long bear market: profit factor 1.03 instead of 1.12. The DAX run covers only one year. History quality was 67% for US30 and 100% for GER40.
- Regulation: NZ FSP register (Black Bull Group Ltd, FSP403326), FSA (Seychelles, BBG Ltd, SD045)
- Trading servers: Equinix London (per BlackBull) · also NY4 / TY3
- EAs, scalping, hedging: unrestricted
Important for readers in the EU and the UK: clients outside New Zealand are served by BlackBull's Seychelles entity – that means no EU investor protection: no ESMA leverage limits, no EU compensation scheme, jurisdiction outside the EU. Details in the BlackBull review.
RoboForex
RoboForex supplied the longest history: almost nine years of real-tick data on three US indices. The Nasdaq-100 held with and without the trend filter, and the filter halved the drawdown (15.7% instead of 31.0%) at almost the same return. The Dow lost money here (profit factor 0.97), and the S&P 500 had no edge at all. One caveat on the data: in the older years part of RoboForex's tick history was generated by the tester.
- Regulation: FSC Belize (licence 9759600)
- Trading servers: Amsterdam (ECN server) (VPS measurement)
- EAs, scalping, hedging: unrestricted
RoboForex Ltd itself states that it does not target clients from the EU/EEA or the UK; German clients are served by a different group company. Details in the RoboForex review.
IC Markets
At IC Markets we have not run LongRunner's MT5 backtest yet. What we have is a live account that has traded the same daily long idea on three indices since February 2026 – with another EA, not with LongRunner. Its original setting (every day, 0.67 × balance per index) is what the "every day, 0.67 ×" rows in our backtests reproduce, and the server time of the Python study (New York + 7 hours) is IC Markets' server time.
From 12 February to 30 September 2026 the account made 167 trades on the Nasdaq-100, 167 on the Dow and 164 on the DAX. Adjusted for lot size, the profit factors were 1.28 / 1.11 / 1.07 – no contradiction to the backtests. In money they were weaker (1.22 / 1.01 / 0.99), because the lot sizes were raised in July, just before the weakest phase: the sizing followed the balance of the whole account, which other strategies had grown. All three indices were weaker from July on than before. Seven and a half months are a cross-check, not proof.
- Regulation: ASIC (Australia, AFSL 335692), CySEC (EU, 362/18), FSA (Seychelles, SD018)
- Trading servers: Equinix NY4 (MT4/MT5) · Equinix LD5 (cTrader)
- EAs, scalping, hedging: unrestricted
More in the IC Markets review and in the head-to-head IC Markets vs RoboForex.
How to test your own broker
- Download the free demo of LongRunner from the MQL5 product page and the set files from its comments section.
- Strategy Tester: your broker's index symbol, M15, "Every tick based on real ticks", 10,000 USD, the matching set file, as many years as your broker's data allows.
- Compare the profit factor and drawdown with the table above. Clearly below 1.0 on the Nasdaq-100 means: look at the spread at 01:05 before anything else.
- Then run it on a demo account at that broker for a few weeks, and only then live. Why one does not replace the other is explained in our guide on backtesting and forward testing.
If you are still choosing a broker, our lowest-spread broker comparison is a starting point – keep in mind that it ranks FX and gold costs, not index spreads at 01:05. Where a broker's trading servers sit is listed under broker server locations.
Bottom line
There is no single best broker for LongRunner – there is the broker whose index spread at the start time leaves the edge intact. In our tests that held for the Nasdaq-100 at RoboForex and for the Dow at BlackBull, not for the Dow at RoboForex. Check UTC+2/UTC+3 server time, use a hedging account, and run the set file on your broker's real ticks before trading live.
LongRunner on MQL5· 99 USDOur own product · not a partner link, we earn from salesBacktests are simulations; past results do not guarantee future results. Trading CFDs on margin carries a high risk of loss.
Further reading
- LongRunner explained – idea, inputs, set files, risk and all backtests
- Running LongRunner on a VPS
- Brokers with a free VPS