Project- Influences of Roman military on modern day riot control

[Bryan]

Comparing modern riots to ancient warfare is like comparing driving to work with racing in a Formula One race because both are in cars. One might learn a few lessons about racing by observing certain aspects of commuting on a highway to work that also apply to racing but the overall focus is night and day different.

[MonsGraupius]

What I'm interested to know what factors affect the depth of a police line and the decisions to form up in loose and tight formations.

The name of the game is containment, so they will typically keep calling riot control until the riot is contained. 

Interesting word "contained" - not how we would describe a Roman line.

No one formula exists for numbers, NYPD operates differently than say Middle of Nowhere, Idaho. Shield walls are used because they create mobile walls to contain the rioters, they offer the best frontal protection from missiles, and they allow officers to gain the advantage of close order formations for the psychological effect of group protection.

The psychology would be very similar.

Their depth is usually limited out of necessity, the rioters rarely try to charge the riot control line and break through it, if they do there are better means of controlling the crowd then shield and batons, CS grenades, concussion grenades, rubber ball grenades, paintballs, rubber bullets, ultrasonic weapons, high pressure water hoses, actual firearms for lethal force (rarely used), which are stationed behind the front rank, so they can quickly be used by opening the ranks or used from the heights of a vehicle roof, or thrown over the top of the shield wall. 

I'm surprised you don't specify anything on depth. When you get somewhere above 10 rows, there's a serious danger of crush injuries or even death. But too few and the crowd could overwhelm the police rows - and literally trample over them.

I'd imagine the closest Roman tactics would be use cavalry or light infantry in a shock tactic. However, I suppose there could also have been archers or slingers who would come to the front to pick off individuals in the opposing line.

That's interesting as I was imagining a very static line with very much the same troops. It would need a lot of training to open up the ranks on command - without killing someone in the charge.

Shield walls aren't used because somebody read Polybius, Livy, Vegetius, or Adrian Goldsworthy books. They are used because the size of the shield (usually rectangular body shields) is most protective in a unified in a group by having everyone stand really close to one another, helped since they don't really need to move dynamically, and rarely use batons either (which by and large are not effective weapons in a fight). Cops don't show up to riots hoping to get hurt, departments cant risk cops getting hurt, which is why they give them so much padding and armor, and the big plexiglass (sometimes metal) shields. 

The Romans were pussies as well (Joke!!!) But seriously, no Roman goes to battle wanting to die either. The same/similar conditions often lead to very similar equipment and tactics. Romans also wanted effective shields for similar reasons. Big difference is that they were up against weapons intended to kill.

Loose formations are best for offensive crowd dispersal tactics, riot control personnel can rush the rioters and just go to town beating the crap out of the slow ones who don't run away fast enough. Its dangerous because the loss of cohesion and order can make a rioter counter attack dangerous, and command and control is basically impossible. But it rarely happens, because most agencies aren't willing to deal with the repercussions of dispersing rioters with violence in the age of 24 hours news and cell phone videos recording and uploading nasty incidents to social media. Same reason they don't just shoot them anymore like the old days or modern 3rd World countries.  

Thanks, yes, the police seem to have the two types of formations: loose and tight - as do the rioters. This just seems to be natural, but I couldn't understand why. So, good to get an insight into the rational.

I've never much seen the point of light infantry (loose): supposedly they were "skirmishers" who engaged the enemy first. That never made much sense to me, unless the other side were also light infantry, in which case the battle was like two women fighting whilst the men watched (and couldn't advance because the light infantry were in the way).

It makes more sense to use them as an offensive force once enemy heavy infantry are retreating and/or have lost formation. Then the faster speed of light infantry enables them to easily catch up and massacre the heavies without the risk of a counter attack from heavy (or at least they can run away quickly)

If you want to see fierce riots you have to look at Japan and Korea (they are mental!)

Normal chain of command exists for the forces. Sergeants are sometimes in the line but usually in the back of it, to better give command and control while monitoring radios and being given commands from the rear, they oversea riot control certified Patrolman. These will in turn be controlled by Lieutenants (who rarely will be included in the line, stationed behind for command and control), Captains, Chiefs, and any other senior officers in attendance (usually its a goat rope, with LOTS of brass, all giving conflicting commands, leading to more confusion and chaos).

I've never believed the idea that centurions were in the front line, let alone the corner of the checkerboard pattern which was almost suicidal as a position. It's pretty obvious that if the "brains" were in the most vulnerable position, then they'd be targetted right from the start leaving a broken chain of command and control.

[Bryan]

Interesting word "contained" - not how we would describe a Roman line.

Because they weren't Romans. If police didn't want to contain, they'd use their service pistols and simply kill everyone rioting. Like how the Romans used the best weapons they had access to, swords and javelins and spears. 


The psychology would be very similar.


No, it wouldn't. South African and Syrian riot police routinely open fire with assault rifles on rioters. American cops in San Antonio will not act the same as St. Louis in a riot because they have different officers, different SOPs, different equipment. Cops and soldiers are different. Different jobs, different technology, different standard operating procedures, different rules of engagement, different mindset, different everything. Modern people and Romans are different in everything. 


I'm surprised you don't specify anything on depth. When you get somewhere above 10 rows, there's a serious danger of crush injuries or even death. But too few and the crowd could overwhelm the police rows - and literally trample over them.

When do rioters or riot control forces every fight it out when each side has 10 or more ranks? I'd like to see this battle, because it most certainly wouldn't be a riot at that point. 

The crush you're describing almost never happening because the protesters aren't trying to push cops away, because to do so means eating mace right in the face like this woman or taking a baton to the head, or getting shot in the temple with a beanbag from a shotgun.

You should read Paul Bardunias' new book on hoplites, he covers crowding, and mentions the Ukraine Maidan Riots of 2014. 


I'd imagine the closest Roman tactics would be use cavalry or light infantry in a shock tactic. However, I suppose there could also have been archers or slingers who would come to the front to pick off individuals in the opposing line.

Why do you keep feeling a need to find similarities when the people creating this stuff don't know a damn thing about Roman warfare? You think cops are sitting around reading RAT on their free time envisioning ways to improve the efficiency of riot control forces by replicating Sallust' description of the battle of the Muthul River? No. They know shields work well to stop getting hurt from thrown beer bottles. They know that the best way to increase the safety of the group is to stack the shield carrying cops next to one another, to create a wall of protection. That second rankers with shotguns and rubber bullets have clean shots at rioters without worry of catching a beer bottle, because the shield wall protects them. If anything, riot control is closer to archaic Greek warfare, or even Scandanavian/Germanic warfare of the Migration period than Roman. 


The Romans were pussies as well (Joke!!!) But seriously, no Roman goes to battle wanting to die either. The same/similar conditions often lead to very similar equipment and tactics. Romans also wanted effective shields for similar reasons. Big difference is that they were up against weapons intended to kill.

Do riot police believe in Virtus? Do riot police decimate police departments who show cowardice during a riot? Do the chiefs have cowards beat to death in front of the whole department the day after a riot as a way of demonstrating the punishment for cowardice? No? Then I guess they weren't the same. There are no same/similar conditions. 


Thanks, yes, the police seem to have the two types of formations: loose and tight - as do the rioters. This just seems to be natural, but I couldn't understand why. So, good to get an insight into the rational.

No, rioters don't have formations because they have no real cohesion, command and control, officers, discipline, standards, etc. 


I've never much seen the point of light infantry (loose): supposedly they were "skirmishers" who engaged the enemy first. That never made much sense to me, unless the other side were also light infantry, in which case the battle was like two women fighting whilst the men watched (and couldn't advance because the light infantry were in the way).

Armies didn't just appear in full battle array, they had to maneuver into battle formations from order of march, a quite complicated job that the enemy would try to prevent, delay, or disrupt. That is the primary point of skirmishers, they stop the enemy from properly forming and advancing in a cohesive and organized manner by bringing them up attack before the main party advances and engages. They harass the enemy, force them to hide behind their shields, while the other force is still psyched up and not scared. 


It makes more sense to use them as an offensive force once enemy heavy infantry are retreating and/or have lost formation. Then the faster speed of light infantry enables them to easily catch up and massacre the heavies without the risk of a counter attack from heavy (or at least they can run away quickly)

They do that too. But they are still useful before the rout, so they will get used before the rout. But skirmisher forces only make up about 10% of total strength, that isn't enough enough to rely on to massacre large amounts of forces, they aren't that light, not that fast. Cavalry is better suited, and that was one of their primary roles in battle in the past, chase down and slaughter those running away. 

I've never believed the idea that centurions were in the front line, let alone the corner of the checkerboard pattern which was almost suicidal as a position. It's pretty obvious that if the "brains" were in the most vulnerable position, then they'd be targetted right from the start leaving a broken chain of command and control.

So the massive casualties that Caesar's centurions faced, that was from what, accidents? Centurions were chosen for steadfast ability to lead, martial ability, and valor, they were the studs and competed heavily with one another in bouts of valor in order to receive better promotions. So of course they would have fought in the front ranks, there is no better place to put a well armed, highly skilled, highly motivated individual, who is trained to lead by example. What is a centurion in the rear supposed to do once the century is committed? Climb up onto the shoulders of someone in front of him to see? From six to eight ranks behind the fighting, how can they command and control the fight? What needs to be commanded and controlled?

Yours in italics.

[Densus]

I am about 5 posts behind the conversation so I will try to answer the main questions without 101 quotes. If I miss anything significant let me know and I will try to answer.

Depth - As deep as you need it be for the circumstances provided you have the numbers. It is very rare that a truly violent crowd intent on killing or seriously officers will close with and try to break through a line, they are much more likely to keep there distance and attempt to inflict casualties through throwing missiles or striking officers with long poles/sticks. In situations where they actively trying to break through the line the level of violence is generally much lower. Therefore (counter-intuitively) you might be dealing with a full on riot with a single line of officers and be 10 deep for a much less violent crowd that are trying to push through to gain access to something.

Crushing - Serious crushing caused by two groups of people pushing against each other is almost non-existent. One side or the other will generally give up/be pushed back before it gets to dangerous levels. It is usually only once you add in a static non-movable object like a wall that crush incidents become dangerous.

Shields - Their primary function is protection from missiles and in a situation where a crowd is trying to push through the line they can often be more of a hindrance than a help so when a crowd is at the low level of violence where pushing is going on normally you wouldn't have shields deployed at all, unless the situation has already been violent and de-escalates. In a few countries like Germany and Denmark it is now very rare that they use shields, they rely on fast aggressive movement and if they absolutely have to stand still in front of a violent crowd they use water cannon to keep the crowd back out of throwing range. They also get away with using fists and feet a lot more than we do in the UK so they can actually get in tighter with a crowd because they don't need space to use a baton to the same degree.

Open order/Close order - Simple answer to this is that when we are static or moving backwards we try to be shield to shield for extra protection from missiles. If we are going forward (and therefore more likely to need to use batons) we move into open order. It isn't a hard or complex manoeuvre, I can teach it to a brand new unit in 40 minutes. The bigger challenge is getting them to actually do it when there is a violent crowd right in front of them. The more scared they are the closer to their mates they want to be. It takes a gradual build up through progressively more violent training exercises to give units the confidence to do it in a real incident. When we do move forward in open order it is not an ill-disciplined charge, we may be in open order

Horses - The horses will go into a crowd BUT I can't think of a single incident where the crowd has actually stood, they always run. I have been on the business end of a horse advance once and it is properly terrifying.

Level of violence - Any time a comparison is made between rioting and ancient warfare you need to bear in mind that the primary aim of the police is not to kill rioters. That doesn't necessarily mean that there aren't people in the crowd who aren't trying to kill you but we are trying to resolve the situation with a reasonable level of force that is appropriate to the overall situation. In a battle you are trying to resolve the situation as much violence as you can physically direct at the enemy.

Location of the commanders - Bryan and I have discussed this before and the answer we came to was it depends what the Romans saw the role of the Centurion in battle to be. If he was commanding his century (ordering formation changes, moving them forward or back, being aware of what was happening on the century's flanks, passing information to and receiving orders from his commander) he needed to be at the back. If his role was to inspire by example then he needed to be at the front. Personally I think it was probably both and dependant on the situation. If things were going as planned he was probably at the back, if it was all going wrong and the men needed the morale boost of seeing their Centurion standing his ground/moving forward into the enemy then he moved to the front.

I hope that helps.

[MonsGraupius]

I hope that helps.

Thanks. Very helpful.

On the level of violence - yes, it's incredibly difficult to know how much the threat of death would change the situation. However, there is some pretty appalling violence in some videos - including group knife attacks - which almost certainly resulted in death (when I posted links to them previously they were removed by a moderator - to my disgust). My perception is that the crowd/police behaviour & dynamics doesn't change massively even with extreme violence. Indeed, the same behaviour gets repeated time and time again in very different societies - so it appears to be instinctively human behaviour rather than socially learned norms.

So, I think we can get a good idea how a rabble attacking Romans might have behaved. What we cannot know is how training & leadership would affect that, and that is particularly a problem with the Romans and police who are trained to counter their instinctive behaviour.

However, the aim was probably very similar: to get the opposition to back down/run away.

On the depth. When you say "10 deep" I presume you are also saying "shoulder to shoulder". However, very few videos online show anything like that number actually engaged. I think the most I've seen is about five deep. And there are very good reasons why it couldn't be more, because to put it in perspective, if you have a rugby scrum (three deep) and you were to put a little old lady between the rows - the forces are such that she could die due to crush injuries. Obviously a scrum is a well organised pushing machine and it's unlikely most troops would have old women on the front line, but even if we assume a "half-hearted" scrum, in theory there could be seriously crush injuries starting to occur when the line in 6 deep. 

There may be something wrong with that calculation as I've not been able to find any accounts of such an incident. However, if you then added into the equation the extra adrenaline of battle and people literally pushing for their lives ...

[Bryan]

Who were the rabble attacking Romans?

[Densus]

I hope that helps.

Thanks. Very helpful.

On the level of violence - yes, it's incredibly difficult to know how much the threat of death would change the situation. However, there is some pretty appalling violence in some videos - including group knife attacks - which almost certainly resulted in death (when I posted links to them previously they were removed by a moderator - to my disgust). My perception is that the crowd/police behaviour & dynamics doesn't change massively even with extreme violence. Indeed, the same behaviour gets repeated time and time again in very different societies - so it appears to be instinctively human behaviour rather than socially learned norms.

So, I think we can get a good idea how a rabble attacking Romans might have behaved. What we cannot know is how training & leadership would affect that, and that is particularly a problem with the Romans and police who are trained to counter their instinctive behaviour.

However, the aim was probably very similar: to get the opposition to back down/run away.

On the depth. When you say "10 deep" I presume you are also saying "shoulder to shoulder". However, very few videos online show anything like that number actually engaged. I think the most I've seen is about five deep. And there are very good reasons why it couldn't be more, because to put it in perspective, if you have a rugby scrum (three deep) and you were to put a little old lady between the rows - the forces are such that she could die due to crush injuries. Obviously a scrum is a well organised pushing machine and it's unlikely most troops would have old women on the front line, but even if we assume a "half-hearted" scrum, in theory there could be seriously crush injuries starting to occur when the line in 6 deep. 

There may be something wrong with that calculation as I've not been able to find any accounts of such an incident. However, if you then added into the equation the extra adrenaline of battle and people literally pushing for their lives ...

I cannot think of a single incident of serious casualties arising from two groups of people pushing each other. One side or the other will break or give ground before the pressure gets intense enough to cause serious harm. Every incident I know of where there have been deaths or serious injuries there has been a wall, a fence or another immovable object involved. A couple of examples of deeper formations and what it really looks like in the videos below. As you can see one side always moves back before the pressure gets too intense.

https://www.youtube.com/watch?v=qzwATFfA1Lo

https://www.youtube.com/watch?v=E8JeL-T6RWk

[Bryan]

Here is a clip from the 2014 Maidan riot in Ukraine that shows deep depth between rioters and security (but neither side has actual cohesive ranks), They are stacked really deep, pushing and shoving, nobody died from it.

Winter on Fire, 23 minute mark

[Robert Vermaat]

Normal chain of command exists for the forces. Sergeants are sometimes in the line but usually in the back of it, to better give command and control while monitoring radios and being given commands from the rear, they oversea riot control certified Patrolman. These will in turn be controlled by Lieutenants (who rarely will be included in the line, stationed behind for command and control), Captains, Chiefs, and any other senior officers in attendance (usually its a goat rope, with LOTS of brass, all giving conflicting commands, leading to more confusion and chaos).

I've never believed the idea that centurions were in the front line, let alone the corner of the checkerboard pattern which was almost suicidal as a position. It's pretty obvious that if the "brains" were in the most vulnerable position, then they'd be targetted right from the start leaving a broken chain of command and control.

Sorry but two misconceptions here I think.
One, the role of a modern sergeant controlling his unit through information is totally different from that of a centurio. The speed and amount of information gathering during battle would be incomparable (meaning the centurio would hardly gather any by comparison).
Also, the chaos of battle would make the amount of control very difficult if a centurio was in the rear and would have to direct from there. The officer stood in front, next to the standard/cornicen, the troops watched these and acted accordingly.
Centurios stood in the front line, or maybe behind a line of antesignani, and had NCOs controlling the rear in case of defectors or a change of front.
Two,

[MonsGraupius]

I cannot think of a single incident of serious casualties arising from two groups of people pushing each other.  One side or the other will break or give ground before the pressure gets intense enough to cause serious harm.  Every incident I know of where there have been deaths or serious injuries there has been a wall, a fence or another immovable object involved.  A couple of examples of deeper formations and what it really looks like in the videos below.  As you can see one side always moves back before the pressure gets too intense.

https://www.youtube.com/watch?v=qzwATFfA1Lo

https://www.youtube.com/watch?v=E8JeL-T6RWk

Densus, thanks for the links I've now got ~70 videos so I can't say for certain, but I've got at least one of the miner's strike, but the Chicago one I think is new. It's also got a tremendous piece of action with the crowd pushing forward in a V.

On the crush injuries, you clearly must be right because it's amazing how few injuries there are.

I think that is because inherently the two sides are "play acting" to some extent. The crowds are during what the crowds think they should do in a protest and the police are doing the same - and it's all a bit like a big dance (at least some of the wilder Ceilidhs in Scotland)
I suppose when you're protesting, your aim is primarily to "be noticed" - so most people are just content to stand and be counted - or push half heartedly.

The miner's strike may have been different - in that if I recall correctly, they were being blocked from entering pits - and/or wanted to block others and lorries from getting in/out.

Can you give me names or some other way to locate those examples of incidents involving deaths.

---

The officer stood in front, next to the standard/cornicen, the troops watched these and acted accordingly.
Centurios stood in the front line, or maybe behind a line of antesignani, and had NCOs controlling the rear in case of defectors or a change of front.

And your evidence for this is what?

You've never played rugby have you? If you had, you'd know that a captain at the bottom of the scrumage is no longer capable of being a captain. If you are in the scrumage - there are only three things you can do: push forward, get pushed back or collapse. And it's almost impossible to talk to the guy next to you let alone organise the "rest of the troops".

And the last thing I'd want if I were a Roman solider, is some nancy centurion getting caught up in the scrimage of the front line, allowing the enemy to work their way around the flank and come at me from behind because THEY WERE NOT DOING THEIR JOB AND WATCHING MY BACK.

[Densus]

Sorry but two misconceptions here I think.
One, the role of a modern sergeant controlling his unit through information is totally different from that of a centurio. The speed and amount of information gathering during battle would be incomparable (meaning the centurio would hardly gather any by comparison).
Also, the chaos of battle would make the amount of control very difficult if a centurio was in the rear and would have to direct from there. The officer stood in front, next to the standard/cornicen, the troops watched these and acted accordingly.
Centurios stood in the front line, or maybe behind a line of antesignani, and had NCOs controlling the rear in case of defectors or a change of front.
Two,

I think people greatly over estimate the amount of situational awareness that someone in the front rank has. Your entire focus is on the people directly in front of you who are trying to hurt or kill you. If you were more than two files down from the Centurion you would have no idea what he was doing or saying, if he was at the front right of the Century as I have seen suggested then to the soldier eight files away he may as well have been on a different planet.

This video will give you an idea of how much someone in the front rank can actually see in a battle that involves over 500 people. https://www.youtube.com/watch?v=78509ehnnnM

[Nathan Ross]

Centurios stood in the front line, or maybe behind a line of antesignani, and had NCOs controlling the rear

I agree. Polybius (6.24) calls the centurio 'hegemon' (leader, or front ranker) and the optio 'ouragos' (tail man). Vegetius says centurions wore distinctive crests on their helmets, and their men used these crests as a mark to follow them - something they could not do if the centurion was behind them! The amount of tactical independence of the century in the battle line was minimal, and centurions led by example, from the front, following the commands of the tribunes and legati stationed in the rear.

However, we've discussed all this before...

Centurion Position in Battle

Two,

Did you get distracted at this point, Robert?

[MonsGraupius]

Centurios stood in the front line, or maybe behind a line of antesignani, and had NCOs controlling the rear

I agree. Polybius (6.24) calls the centurio 'hegemon' (leader, or front ranker) and the optio 'ouragos' (tail man). Vegetius says  ...

I liked the comment suggesting the leader would move around as required as that was based on actual experience in a very similar situation. I also like arguments from the available evidence (although I don't think being called "front ranker" meant you were in the front rank anymore than "commander in chief" - means the US president would fight in the battlefield. Or "Sir Branston", being a knight rides a horse with a ... what are those long sticks ... javelin?)

But I'm sick to death of books and websites that tell me that such and such happened, when I know, or am fairly sure, that there's no evidence to back up the assertions. (My pet peeve at the moment are authors who will state in no uncertain terms and without any evidence: "of course Tacitus was a liar" or "Tacitus cannot be trusted on this" .... usually just before/after they recount the whole of 1st Cent British Roman history using Tacitus as their primary source.

[Bryan]

You've never played rugby have you? If you had, you'd know that a captain at the bottom of the scrumage is no longer capable of being a captain. If you are in the scrumage - there are only three things you can do: push forward, get pushed back or collapse. And it's almost impossible to talk to the guy next to you let alone organise the "rest of the troops".

And the last thing I'd want if I were a Roman solider, is some nancy centurion getting caught up in the scrimage of the front line, allowing the enemy to work their way around the flank and come at me from behind because THEY WERE NOT DOING THEIR JOB AND WATCHING MY BACK.

And what we have here is a basic misunderstanding of ancient Roman combat by your comparison of it to a modern rugby scrim. You're doing the same thing by trying to interpret modern riots being "just like" ancient warfare because a few small similarities existed. Its a form of presentism, an anachronistic misinterpretation of past historical events by trying to explain them with modern concepts. The Romans didn't play Rugby and Rugby isn't especially similar at all to how a Roman century would be fighting in battle.  

A Roman milites isn't supposed to worry at all about his flanks, that is so far outside of his responsibilities its really not worth mentioning. So he shouldn't be worried about his "nancy" centurion getting distracted (your description doesn't even make sense, someone fighting in the front ranks is the last one to call a wimp, which is what Nancy means). The century isn't operating all alone in a large field like a rugby team. Apples and oranges. 

The flanks were the responsibility of the centurion to some extent, they would have had some ability to dictate their century's depth and frontage and choosing whether to advance recklessly across bad terrain without flanking unit protection. But a centurion wouldn't be "watching his soldier's backs" because other, some junior (the optio) and more more senior, that was their job, to command the whole of the battle lines. The centurion's only responsibility was his own person centurion, or his maniple, which would often be supported by many other centuries/maniples in their cohort or battle line, most of the time with other reserve lines close behind them (the reasoning to prevent breakthrough and small unit flank attacks), as well as having light infantry attached to them, often fighting from within the gaps between units. 

What modern military theorists call "Commanding" (vice Leading) was the role of the tribunes, prefects, legates, and Praetorian/Consular generals, who either rode or walked immediately behind the front line, their jobs to rally forces in the local area, control advances and retreats, orchestrate line relief, etc. They'd be the ones who would be watching the centuries involved in a fight from a superior position (immediately behind, often on horseback for height), and most importantly they also had the authority to command other forces and coordinate manipular and cohortal maneuvers (a centurion had no authority to order anything outside his own century/maniple, and there was little he could do anyway to his own unit once it was firmly committed to close combat). Meaning if a centurion in the back of his century somehow realized he was being flanked the tribune would have seen it too, the centurion couldn't alter his unit's formation while engaged in close combat anyway, nor could he order a nearby centurion to redeploy his century to support. A tribune could order a nearby century to support, he had the command authority and he was in a position to best get them to act. 

Let's look at the evidence:

- There is no evidence the Romans had sub officers in their front ranks, but there would have had to have been someone to lead from example in the front. 

- We know the most courageous rankers in the unit were made standard bearers, and that carried with it an increased risk of death. We know that it was the centurion's job to protect the standard bearer. 

- We know the "chosen man" Optio, assistant to the centurion, often served in the rear of the formation, to keep the men in line, force them to advance, prevent them from retreating. 

- We know the tribunes were immediately behind the leading battle line, they often jumped into hard fighting in the front lines, leading by example, they too suffered large casualties, similar to centurions, that they often served on horseback, and that because they wore gaudy uniforms from horseback within missile range of the front line and couldn't use evasive maneuvering to limit getting hit, it made their job more dangerous. 

- We know centurions often suffered extraordinarily high casualty rates in comparison to normal rankers, which means they would have put themselves in very dangerous positions. 

Since centurions weren't on horseback like many of the senior officers were, since they were the lowest true small unit leader with real meaningful authority gifted by the tribunes or general (representing SPQR), that they still were killed in large numbers despite wearing some of the best equipment available, and since most would have been, at the very least, mediocre skilled warriors with sword, shield, and pilum, since the standard was in the front to lead the men and they had to protect it, then it means the centurions led from the front, they didn't command from the rear. Not in every situation, but in most, especially pitched battle advances and clashes. 

Whether they were front center, in the line, front right, in the ranks, or kind of floating center or right in front of the leading rank, that's a guess, but I'd lean that they had a specific position within the leading rank, which was theirs, which they would fill when fighting in a pitched battle, leading right, the position of honor for a unit.

[Densus]

If we accept that the Centurion is in the front rank and has no control over his Century or any awareness of what is happening more than ten feet from him and if we accept (to take an idea from an older thread) that the front rank is made up the more experienced soldiers/better fighters who are going to hold that position until the battle is done because it is the position of honour.

It makes me wonder what, if anything, is different between a Roman Century in battle and a 'barbarian' warband of the same period?