Choosing your Deck wisely might be the most surprising step for your enemy in any fight. While the enemy knows what library you use and can see your procedures in action he will never be able to see your deck details. So he will always be worrying what deck you might be using. Is it that extreme fast one that is quite unstable. Or is it the versatile one? He might get some clues once the fight runs, but he can never be sure.

To make comparison easier all decks share a common template form.

Click to open sample deck in separate window.

First there is the name of the deck.

Next you see the market price of a deck. A deck can be purchased at various shops for this amount of money. No negotiating about the price, sorry. One might be able to buy a deck used however or steal it using less money or none at all. This however is not the norm.

Then there is most important stat of a deck: The Main Processor.
This is the stat that describes the raw CPU power a deck can provide at maximum efficiency. Of course hardware can only be as effective as the user. So this value is directly linked to the Power Level of its user. A computer cannot be more efficient then the person that works with it. And similar a hacker cannot outdo the physical limitations of his deck. Both have to work in a symbiosis.
Under normal circumstances a deck can produce a peak performance of 10 times its Main Processor power or the Powerlevel of the user - whatever is lower.
So a deck with a CPU of 1 can load a library worth of 10 or start procedures that need up to 10 power combined (given a sufficient CPU cache). The average power output is however only 5 times the CPU level. Peaks can only be reached for a very short time and always are paid in later waitstates.
The 5 points are however only the average. They are hugely dependent on the state of the Kernel. On a slow system with a kernel speed of none, a CPU level is only worth 3. On the other hand if you get a very fast system with a kernel speed of extreme a CPU level can be worth 7 points. So monitoring the state of the Kernel can be a very important thing to do.

The next line(s) describe the coprocessor efficiency of a deck. While the Main Processor is very important it is not the only valuable stat. The Main CPU got so many different tasks it is impossible to compile all procedures in real time as well. This is where the Co-Processors kick in. Just like the main CPU, the Co-Procs form a symbiosis with a hacker's skills. When a hacker cannot use the full power of one it will be limited by his weakness. But again a hacker cannot outperform deck limitations. Even a cybergod has to wait the cycles a compile needs on a given Coprocessor.
So the true efficiency is again the minimum of the deck's and the hacker's powers.

The Instruction Cache is your main memory. Of course your deck got an insane big amount of base memory, but it is too slow for the real thing when speed is of essence. The only part that counts here is the IC. This valuable memory determines (together with your CPU output) how many action points a deck can store for later use. So an instruction cache of 30 offers the ability to store 30 action points given that both the CPU and the PL are 3. If the CPU output would be limited to 2 (either by a damaged CPU or by a lower PL) the effective IC would decrease to 20 as the CPU cannot use more of that memory.

Next stat... I/O Channels.
All is great and fine when you prepare the procs on your deck, compile them and are ready to release them. Just one problem - the procs have to be uploaded fast into the target system. Here the I/O Channels come into effect. Each channel enables the upload of one procedure to a fight. You can however re-use an uploaded procedure in the same turn. So with an I/O of 1 you can use 4 NOP Bombers in a turn, but you cannot use a Nop Bomber and an ANB. This would need an I/O of 2 - so a low I/O while not limiting your power directly, limits your diversity a lot.
Why not use 10 or more I/O lines then? Well, for one they are not that cheap. But there is also another reason: As the name suggests it is Input and Output. As fast as the programs can be uploaded just as fast malicious code can enter.
So a high I/O isn't only a good thing. It also opens up your deck to the risk of multiple damages on a crash or similar attack.

The precompiler section is pretty straight forward. It enables to run procedures at a higher speed rate, but with less flexibility. Only very expensive decks offer such a compiler.

The probably most important value of a deck is the "Auto-Disco Efficiency". It does not do anything for you in respect to power. But it will save you big bucks. The master fuse will hopefully disconnect your deck before it can be more seriously harmed. The percentage shown is the percentage the fuse will blow on each hit.
Be warned this master-fuse is extremely expensive, especially on higher efficiency ratings.

Second last there are the normal fuses. These are your armour. Evenly distributed over the system parts, these jewels make sure that they go first before some valuable part takes damage. They are cheap as well, should they ever be damaged (as long as you don't have a special 20 fuse deck or so). Now aren't that good news? Here are the bad ones however. Only if all fuses are present they'll 100% divert damage. Once the fuses start to be blown there is a chance that damage can already hit the deck even if some fuses are still there. So if you are unlucky enough and the same subsystem(s) are targeted twice already one blown fuse might give you problems.

Finally there is a description of the deck. Where it comes from. It's history. Manufacturer. Stuff like that. Also special properties of hardware will be noted here.