Friday 5 November 2010

Conqering Citites Part 1 - Timing

Timing

One of the most crucial skills you need to learn is how to time your waves perfectly. Easier said than done - I know the pain! We are talking here about the anti-timing mechanism which will alter the travel time +/- 30 seconds at launch. It affects attacks and supports. In my world and unit speed equal 1 I never really experienced 30s but 1-20s most of the time. 

Clearing Attacks

To be able to squeeze few clearing waves just before your CS ETA (Estimated Time of Arrival) you need to be bit lucky and help your luck as well. Also to get best results you will need few waves - with only one wave you cant do much. Play safe and send it 30s before your CS ETA.

Strategy with many waves:
Make your first LS wave a sure hit - open the attack window at least a minute before sending your clearing wave. Then refresh it few times while watching the timer showing ETA (while watching it just refresh your window by closing it and opening it again). When it shows 1 minute to your CS ETA send your attack. That means if your CS ETA is 20:00:00, you will click the attack button when the timer shows 19:59:00. Then do the same with your land nuke (use breakthrough!) to have a definite land hit but send it around 19:59:30. By sending your LS nuke a minute before you should create big enough window for your land attack, so it will hit after navy but still close to the CS (if luck is not on your side and land clearing wave is going to hit after LS, then you will have to send another early LS wave - add 30s to the land clearing wave ETA, so if it hits at 19:59:10 then send LS at 19:58:40).

From now on you can start sending your waves around 19:59:30 or even 19:59:45 if you want to be more risky or know how much time is usually added/taken (after sending hundreds of waves I knew it was 1-15s, max 20s for me). 

Napkin math - what are the chances your attacks will miss your time window
I'm not a mathematician but if an attack can be moved by anti-timing mechainsm by 30s in one way and 30s in other, then its 60s total. That 60s is equal to 100% possibilities. When I send my attack at 19:59:40 then it can hit between 19:59:10 and 20:00:10 (+/-30s). Therefore only if the system adds 20s-30s I will be out of my time window. That is 10s out of 60s, which is 1/6 of the all possibilities, which is 100%/6=16.6%. So in other words when you are sending your waves at 19:59:40 (or 20s before ETA) there are roughly 85% chances you gonna make it and 15% that your wave will land after the CS. If you send your attack at exact ETA it will be 50/50 chances you get there before the CS.

So after this cool risk management lesson from Grom you should know what to do with your waves :P The more offensive specialised cities you have, the more chances you will get close to your CS ETA. Each time check ETAs of your waves. That way you will see how much they are changed on average and you can make your decision to risk or not. If all are still 20s away from CS ETA why not to try and send one much later?

 When I was desperate for a "good shot" i would split my last LS nuke into 2 parts and send one around 19:59:40 and the other one at 19:59:45, risky? Maybe, but I was sure that 15-45s before my CS lands, there will be plenty of LS waves hitting that city anyway so I can waste one and try to get a good shot. If you don't have that many offensive cities then you cant really afford to risk that much. It worked for me because almost 9/10 times my waves were not altered more than 15-20s.

I was risking little bit by sending my waves around 20s to CS ETA (using the previous example at 19:59:40) but in this way I could put majority of my waves in the most dangerous time interval for the defender - between 19:59:10 - 19:59:59 but mostly between 19:59:20-19:59:59. That 40 seconds were loaded with my clearing waves and made a defenders life so much harder. He could split his defending units and dodge with many of them risking that these little groups will get wiped by my nukes, or dodge with one big wave risking missing my CS. 

So first I was making sure that my LS and land nukes hit the target before my CS and then was gradually sending more risky attacks (at 19:59:35, 40, 45) until I got one very close to the CS ETA that would make me happy. From my experience most of the attacks that I was sending around 19:59:40 were good ones even though I knew that in theory it can be altered as much as 30s. By watching ETAs I learnt to trust my instinct and was able to get 19:59:59 hit couple of times and 19:59:50-55 on regular bases. 

Naturally things might totally change if you don't send your CS first but sometime between sending your attacks. For example you planned your CS to land at 20:00:00 but it landed 19:59:50. It might be good as your best hit was 19:59:49 or it can be bad if you already managed to put couple of waves after 19:59:50. If you want to avoid that situations send your CS very late (around 20:00:00 if the defender is not a sniper and you dont care much about recalling and resending your support again) and make sure that your support waves don't come earlier as they will defend against your CS. All of your attacks (not teammates) that hit after CS won't hurt it (they will drink coffee and have some cakes...literally).

As Wellagain says in her instruction for traveltime calculator:
If you are sending your CS as the last wave (it's travel time is the shortest, and you have to send all other waves before actually knowing where anti-timing will put your CS), knowing how exactly your queue looks like will allow you to make the best decision on where to aim your CS at (remember, attacks landing after your CS will not harm it, so it's better to put it before the last attack rather than after first support).

There is nothing more to that really - no way to go around anti-timing mechanism. If you have only one or two waves you cant get a close hit without risking everything, so I would play safe and send it at around 19:59:30, maybe 25.

Now go and make some tests - send your troops at a ghost town or farm - keep sending and recalling troops to see what sort of times you get. You will get a feeling of how anti-timing mechanism works for you, how often it adds or takes 30s and how often its stays in closer range between 0-20. Some people reported that the mechanism never added more than 15s for them. Is it true? I dont know. Go and check by sending many waves and recalling them.

Supporting CS 

Supporting is  much easier, because if things go wrong you can recall it and send it again. It depends on how many waves you can send but good news are that you can split your forces in as many parts as you want to! This is not attacking anymore, we are supporting, so we don't care about size of our sent waves.

First prepare yourself - decide how big "chunks" you want to send (divide your forces into smaller groups). So lets say I have 350 Biremes in my city, I decide to send them in 7 groups of 50. I will type number 50 in, select it and copy (Ctrl+C). Now when its around 19:59:40 I will send first group, then fast Paste (Ctrl+V) number 50 and send next one, and again and again 7 times (7x50=350). I copy/paste as it saves me time from typing, You can type these numbers manually if you wish of course. After that I will check all ETAs and these groups which land before CS will be cancelled and resent back as soon as possible. If you have to send troops and ships then you can split them into 2 groups - just prepare first wave and then second one will be easy to send straight after without typing in numbers (just select all).

I would do the same with all my other supporting cities. Enemy would probably see 100s of incoming support waves but that doesn't matter. What matters is that with so many shots you can put many of your support waves 1-5 seconds after your CS lands. If for example I was sending support from 3 cities and sent only 3 big waves there was chance I would get times like 20:00:08, 20:00:10 and 20:00:15. But if I split them as I proposed above I will get 21 different waves and lot of them might hit magic 20:00:01.

I did that many times and no-one ever sunk my CS because my support was to late. Snipers had a hard life against me. Especially that anti-timing mechanism applies to defenders as well. For worlds older than 2.0 there is some short info about dodging (from grepo forum):
"Such precision is not possible in this game in any case, owing to time randomization factor that is built in to any troop movement. This can add or subtract up to 30 seconds at the time you click the attack button."
"Farming villages are exempted from the antitiming rule when outbound. That is the strengthening waves will land before the looting attack. They don't always return in the same order. With a farming village they arrive at the appointed time but lose a few seconds or gain a few on their way back, the real anti-timing rule takes over when you use it against a player." 
I can add that cancelling troops while they are moving to strengthen a farm will not work - anti timing is applied when cancelling.


If you have any comments or questions post them below, thanks.







3 comments:

  1. Good article, thank you.

    ReplyDelete
  2. do you know if anti-timing mechanism depends on size of the wave?

    ReplyDelete
    Replies
    1. I have never heard of it so I cant tell. I would be surprised if it does. It is hard to check though, as you will always get some random numbers.

      Delete