An interesting debate to which I have arrived late. However, I am intrigued by the details of your combat.

If I have understood correctly, a complete formation of 8* Scorpions (+ Exarch), mounted in Wave Serpents (giving 21x 4+ hits), sprang a mounted assault on a formation of 6x enemy (Marine tacticals??) that had at least one BM. The Scorpions ended their movement in contact with their opponents but failed to inflict any damage from the (reduced?) number of hits inflicted, while they failed to save the 5x hits returned.

The Scorpions anticipated being +4 in the resolution (No BMs, enemy BMs, Inspiring, Numbers), but now found the resolution -1 against them due to the 5x casualties. Then to add insult to injury, they must have thrown two "1"s while their opponent threw a "6", wiping out the remaining 7x units from the Scorpion formation.

If I have got this right, the summary of the statistics is as follows:-

• 1/1024 An average number (10x) of hits from the Scorpions are all saved against 4+ armour (or ruins etc)
  
• 6/64 The marines score 5 hits
  
• 1/32 The scorpions fail to save any
  
• 1/36 Scorpion resolution of two "1"s
  
• 11/36 Opponents resolution of a “6” from two dice
  
  That works out at approximately 1/ 247,082,636

I may have got some of the details slightly wrong, but as you can see, statistically this was extremely unlikely, so you can rightly feel very aggrieved at the completely unexpected result - and possibly write about it in some article or even use it to claim that you are even less "lucky" than me .

But I suggest that it is so unlikely that you are probably never going to repeat it again - hardly the basis for revising the current rule mechanics.

Ha ! It was something along those lines - can't remember all the details now (trying to block out the memory.) Wasn't all scorpions and both units had taken a few casualties - i forget the details now. I think we did a quick approx calculation at the time and it came out around the 20,000:1 mark - I'm unlucky, but not that unlucky !

using a mean in place of a full distribution to calculate p(all_marines_survive) is likely to be off

calculation is along the lines of... p(0 hits) + p(1 hit)*.5 + p(2 hits)*.52 + p(3 hits)*.53 + ... ... ... + p(21 hits)*.521

I'd refuse to play with such a house rule myself. I really, really like the current system and the way assault resolution dice can give victory to the unlikely side. It reflects the effects of chance and keeps things exciting.