I think this is potentially heading for a record as FOGs longest stream

The starting point for all of this is some stats. I don't want to spend ages boring everyone with the details of how to simulate it properly. But just to say that in setting up the rules we simulation tested things to death at times by two different methods - Terry built a pretty cool Monte Carlo simulator, and I have a Statistical Spreadsheet set up that streams all the probabliities and stacks them. To be fully informed you need to:
1. Set up the probabilites for 30 different possible IMPACT outcomes streams - a combination of bad loss, loss, draw, win, big win x 3 possible CT results x 2 possible DR results. Hence 30 outcome streams.
2. You then need to do the same for each stream for the MELEE - so another 30.
3. You then need to run it long enough to get to a break as what is important is not who survives round 1 but who wins the tie overall - generally this means 1 more MELEE round for anything reasonably uneven, and 2 for anything even
So the output streams of odds are 30 x 30 x 30 (x 30) = 27,000 (or sometimes 810,000) outputs to then recombine.
Even this is not enough to test it properly as you then need to try it a few times for
different realworld scenarios involving generals, rear support etc to make sure you are comfortable with the whole profile. This is the process we went through at times for anything we were worried about - columns in combat being probably the first obvious potential problem we thought about. Having been through all this I would suggest that " a little simulation can be a dangerous thing" is not a bad motto.
To leap to conclusions that relate to the dialogue here:
1. Going in in column is plain bad vs combat troops. When you run the overall stats their chances of winning are so compormised that they end up losing badly vs a simple stand up fight. 2 years of gaming experience prove this to be the case empirically as Roger says. But anyone not convinced is welcome to give it a try.
2. Graham Briggs scenario is different. If you are going against missile armed troops who get extra dice at impact then the trade-off is more even and it is potentially viable to take the risk in column. But the real world problem is that it often pushes you into suffering CT tests from shooting at -1 for 1HP2 as only 3 bases count and often the same firepower will hit the column as would hit the 3 wide BG anyway (when you would need 3 hits rather than 2). Internstingly this part is actually more even the less armour you have as if you are unprotected MF you are likely to suffer 3 hits anyway!
3. Going in 2 wide 3 deep is a very sound tactic. We are happy with this as it dcoesn't have the potential volatility of the 2 by 2 combats. The risk here is tactical in that you havea narrower line and can end up very exposed. But that bit is a game of tactcal skill. Depth is also useful vs firepower of course.
4. All the above have odds that change considerably when you add generals and rear support behaviour to the equation. Put simply if you think of a column of poor foot with no generals then prima facae they seem better in column because their odds of failing is very very high when deployed. If you add a general to average troops then they are more likely to survive deployed out anyway, if you go in column then the benefit of the other side putting a general into combat is much higher as you are very unlikely to get 2 hits and have a shot at killing the general, so your odds in practice change a lot.
There is much more to it than meets the eye ... hope that helps reconcile some of the instincts on this stream
Si