I mean, was the combat engine actually doing anything it wasn't supposed to, or are you just giving into the salt?
One thing we should bear in mind is that while
any given event that has only, say, a 5% chance of happening is by definition unlikely (such as
Challorn being targeted 14 times out of the first 20 Sydraxian shots fired in that last battle)...
There are a lot of separate things that can happen which
all have low probability.
Some of them are bound to happen in any given battle (e.g. an enemy ship dodging three shots in a row, or dodging 5 total when you'd expect it to dodge 3, or one of our ships taking a disproportionate number of hits during a specific phase of the battle).
This is why large data sets invariably contain clusters and anomalies. When you have thousands of data points, you have millions of ways that two data points can be
related. Of those millions of possible relationships, 1% of them will be relationships that only had a 1% chance of occurring on any given trial... which means thousands of events of low probability,
if you exhaustively mine the dataset for all the improbable coincidences.
The only time you can definitively look at something like Oneiros's combat engine results and say there is a bug is when something that
obviously should not be happening happens- for example, when
Challorn was firing shots into the battle after retreating a little after Turn 70.