[Two Companies Running Together Real Verification] AnomalyEA Operation Report Issue 1 (7/24〜8/20)

Hello. I am Leo FX, developing the USDJPY-specific EA "AnomalyEA".

In this series, I will公開 the operation results of the EA I created,whether I win or lose, 그대로will be published.

There are other places that simply list results, so I added one feature to this series.

The same EA is run simultaneously on two brokers: OANDA Securities and Gaika Finest (Gaiten Finest).

By running with the same logic, at the same time, with the same lot size,we'll quantify how much the results differ by brokerin numbers. I couldn't find any records of this kind when I searched, so I decided to make this the centerpiece of the series.

The reason I do this is that in the past I myself have been deceived by two aesthetically pleasing backtests. So when I become a seller, I decided to show the entire verification process.

There are three promises for this series.

1.Publish both wins and losses as they are(there's no point in a series that only reports convenient weeks)

2.Run concurrently on two brokers and publish broker differences

3.Write about failures(in fact, we already messed up twice this time)

Now then, here are the results from July 24 to August 20.

■ Summary (starting with the conclusion)

・5 wins, 1 loss in 6 events

・OANDA Securities +67.7 pips ／ Gaika Fineste +70.7 pips

・Win rate 61.5%（8 wins, 5 losses out of 13 trades）／ Profit Factor1.91

・In 1.0 lot termsapproximately +70,700 yen(actual amount on the Gaika Fineste account)

■ Operating conditions (same for every entry)

・EA: AnomalyEA (time-based, USDJPY only,no averaging up・SL 25 pips, time-based exit）

・Accounts: Running in parallel on 2 brokers: OANDA Securities Demo / Gaika Fineste Demo

・Lot size:1.0 lots fixed（※On 7/24, OANDA side was 3.12 lots. Details below）

・Entries occur 2–3 times per day and are settled in a matter of minutes to 3 hours. There is always only 1 open position

■ All trades (13 trades, two brokers comparison)

Viewed by day, it looks like this.

【Cumulative P/L Trend】

■ Realized spread section (this series' flagship feature)

This is the main topic from here.

This EA enters every day at 8:30 and 9:55. The 9:55 entry isright after the nominal rate for the Tokyo time is set, a time when real demand buying/selling concentrates.

In other words,it places orders at the moment when the spread is most likely to widen in a day. I recorded the spread every minute believing broker differences would appear.

The result is this. The maximum spread during the nominal rate time window (8:25–9:55).

【Spread Comparison】

On the OANDA side, in 5 of 6 cases the spread widens to around 2.0 pips at most.And the widening occurs specificallyduring 9:53–9:55, a narrow 3–4 minute window. Gaika Fineste, except on the market-wide volatile 7/31 at month-end,never exceeds 1.5 pips.

What’s interesting is that from a daily average perspective, the difference is hardly noticeable.・Average for 8/5: OANDA 0.90 pips ／ Gaika Fineste 0.85 pips

This EA faces an unfavorable moment exactly at the instant of order placement.If you compare daily average spreads, you would never notice this difference.

There are actual damages as well. The AnomalyEA has a safety feature that it doesn’t place orders when the spread exceeds 1.5 pips, buton July 24, orders were deliberately skipped 10 times by OANDA at 9:55. They eventually filled, but the entry was at a price worse by 0.6 pips compared to Gaika Fineste. Gaika settled immediately at the same moment without any skipping.

(This is data from 6 demo accounts. I won’t make definitive claims yet. I’ll draw conclusions after accumulating more data.)

■ Three memorable days

① July 31: +70 pips

08:30  159.92  ← Long position
09:53  160.62  ← Exit (+70 pips)

End-of-month last business day and a go-to dayA day when real demand for dollars clustered. Moreover USDJPY had been falling since the previous day, and with month-end buying, it recovered +70 pips in 83 minutes.

② August 10: First negative day

Conversely, on this day after the nominal rate passed, it did not reverse but continued rising,the short leg reached stop-loss.

09:55  157.92  ← Short position
10:31  158.17  ← Stop-loss (−25 pips)
11:45  158.26  ← The actual exit time would be here (−33 pips equivalent)

The stop-loss limits the loss to 8 pips.That is what stop-loss is for.

③ August 14: I forgot to start

Honestly put.I forgot to start MT5, so I missed the entire Go-To day on August 14.

Replaying with historical data later, both legs were losses on that day,−9.4 pips (about −9,400 yen). In other words, I avoided a loss in the end.

However,if this had been July 31, I would have missed +70 pips, which is not funny at all. As a preventive measure, I set MT5 to auto-start on PC login.

■ Announcements and one honest note

On August 10, AnomalyEA began selling on GogoJungle.

And on the product page, GogoJungle showsan independent forward test run on their own environment, with their own accounts and default settings. I, the seller, do not touch it at all—an independent measurement.

That number is currentlynegative (PF0.68).

The reason is simple: this measurement started in August and has only 4 trades so far. The August 10 loss is included as one of those four trades. As shown above, July results are not included.

There’s no point in hiding it, so I’ll say it myself.If you’re thinking of buying, please consider that number is still based on only 4 trades.And as the number grows, I will report whether it moves up or down in this series.

■ Future plans

・Remaining Go-To days in August：8/25 (Tue)・8/28 (Fri)・8/31 (Mon)

・8/31 is the last business day of the month, so the “End of Month London FIX” leg that caused the July 31 stop-loss will fire a second time. A rematch.

■ Honest corner this time

I will write three things.

1) The lot size was off.On July 24, balances between the two demo accounts differed, so the compounding calculation caused the lot sizes to be misaligned (OANDA 3.12 lots vs Gaika 1.00 lots). Since a monetary comparison could not be valid, on July 29 I unified both to1.0 lot fixed. This is why pips are the main focus in this article.

2) I forgot to start [the EA].As noted above. I was lucky, but it did create a data gap.

3) This is the most important.Looking at the spread table, Gaika Fineste appears favorable, buton July 31 stop-loss execution, OANDA actually had smaller slippage and was more favorable (−26.1 vs −26.5).

It’s not a simple case of “Company A is all good.” If you cherry-pick favorable parts, you can write articles praising either broker.That is why I present all 13 trades without omission.

This article is a record of automated trading program verifications and is not investment advice. The displayed results are historical performance and do not guarantee future profits. FX trading may incur losses exceeding the initial capital.

・USDJPY-only EA “AnomalyEA” (GogoJungle) → https://www.gogojungle.co.jp/systemtrade/fx/82678

・Free tools
Go-To Day Calendar
https://www.gogojungle.co.jp/tools/indicators/82759?utm_source=share

Account Suitability Checker
https://www.gogojungle.co.jp/tools/indicators/82760?utm_source=share

・Live account records (Real Trade) → https://real-trade.tech/accounts/83716