pells wrote:
- I wasn't inform of the goals of the PCs. Well, in fact I guess they didn't have a real plan in head. They decided to poison some orcs after being in the camp for a while. Which is fine with me : that part was really fun to play as I had to describe the whole camp, creating everything on the spot with the help of the players. But using TSOY, which stakes should we roll for ? I guess the one we are confortable with ...

Ask the players what do they want to roll for. If they don't have any idea what they're going to do at the camp, or they want to make sure they succeed at sneaking, then you can roll about whether they manage to sneak in undetected. If they want to poison orcs, then roll for that. If they want to scout the camp, then roll for that, including successful exit. There are many different ways a scene like this could go, all depending on what the characters actually want. This is the "principle of the camera", that a story focuses only on those things that are important, not those that are not.

- Let's say the PCs decided to announce me their goals as "come invisible into the camp and put poison into the leaders' food", then how many dices do I roll ? Against what do the PCs roll ?

If I didn't have any statistics for anything prepared here, then I'd whip up some for the leaders of the orc camp, whoever is getting poisoned here. Then I'd have the players roll their sneaking skills or whatever they're using here. I'd probably have each individual sneaking character roll sneaking, with successes as bonus dice for the others (the characters help each other stay undetected). If any fail, then the orcs at the camp have noticed them. However, if all succeed, then the final successes (the success of the final sneak roll) are compared with some manner of alertness roll from the orc leader who is being poisoned. Succeed in that, and the players are pretty much home free; the initial success in the sneaking checks established that the characters have the basic capability of moving about in the orc camp without being seen. The reason we compared the check with the leader orc's was that the leader was the one who was most affected by what the characters were doing - he deserved a chance to notice it and avoid the poison.

If there were a player character guarding the camp or there was some other reason for the camp security itself being a dramatic issue, then there could have been two separate rolls for the orcs: one of the leader and one for whoever put up the camp security. This is because in this case we would have been interested in whether the special security worked successfully. But if there was nothing special in place in that regard, then the simple success is sufficient for just wandering around the camp without doing anything too untowards.

So, in summation: a simple success in sneaking for actually going into camp undetected, then a suitable resisted check for actually doing something important there. If the important thing is trivial in itself, like slipping pre-prepared poison in food, then the initial sneaking check has double duty in that regard. But if the character was trying to do something difficult, like contort himself inside a tiny chest in camp, then he might have to roll more checks.

That said, let's say the PCs, who would play black orcs, come to the gates of Murdithem and choose to "bypass the brigands and get outside the stronghold", how many dices do I roll ???

Do you mean "inside the stronghold", perhaps?

Depends on what the PCs are doing, exactly. If we assume that the brigands want to stop the characters from getting in, then you need the statistics of the "average brigand" here. There are two cases:
- The PCs do something where the respective number of the brigands and PCs does not matter. In this case just roll a resisted check between the PC and the average brigand. This would be the case if they tried to bullshit themselves in, for example.
- The PCs attack or otherwise do something where the respective size of the forces comes to account. Again roll a resisted check between the PCs and the average brigand, but now give a penalty die to whichever side is the tactical underdog in terms of force size, quality, or whatever.

The basic principle here is that the number of people, especially statists or otherwise identical people, does not affect checks in a fundamental way in TSOY. The check to get in the stronghold is the same whether there is one brigand or a hundred guarding it. The number of people comes in only in some very definite ways:
- The force disparity could be so large that it affects the stakes themselves. For example, if the PC is alone and there are a hundred enemies, then the SG is within his rights to declare that the best stakes the player could hope for would be to "escape" or "delay" the opposition; the force disparity makes it unbelievable for the PC to, say, "triumph". Likewise, if the opposition is split into two different commands, a single PC command could only hope to stop one of them, most of the time.
- If the disparity is not enough to affect the stakes, or even if it is, the SG may assign one or two penalty dice to the side that is acting in significantly difficult conditions. One of those conditions could well be that the opposition has set up a prepared defensive perimeter, or that there are more of them than us.

Note that if there's a bit less of those brigands and they're a bit more important as persons, then they'll all have their own statistics. In that case it might make sense to roll a series of independent ability checks for bonus dice among the group. But that's a different situation.

The above principle of not caring about the number of people, only their importance, is what makes TSOY a fundamentally dramatic game, as opposed to realistic. The guiding principle of TSOY is that it's always about individual passions clashing in some manner. The system of Secrets, on the other hand, is there to let the players give dramatic weight to different aspects of the setting they wish to give the power to constrain the events. Thus, when somebody is not satisfied with the simplistic approach above, I recommend different Secrets to tailor the situation to their tastes. For example:

Secret of Discipline (activity)
The character acts as one with his group of peers, who all have the same Secret of Discipline. The group acts as one character in conflicts, with Pools set at the level of lowest personal Pool in the group, +1 per additional character. The group has the lowest Ability rating of the participating members in a conflict. The group can use any Secrets known to all group members. The group may, however, only act on matters that it is intended for, so a military organization couldn't do scholastic research, for example. The group members suffer all Harm caused to the group individually. The group Pools are formed when the group gets together and changed whenever the size of the group changes. The group Pools are only replenished by individual characters sacrificing their own Pool points to it in the form of impromptu leadership scenes. The same character can be a member in several groups at once (for example, a trooper as a member of his squad and his battalion) and he can leave the group for short periods of time, but only as long as group cohesion does not suffer. Cost: 3 Instinct from each joining group member.

The above Secret is suitable if you feel that a group of statists should reflect their numbers somehow. While in a game without the above Secret a large group of individually insignificant characters is mostly meaningless, unless led by an important individual, the above Secret makes the little people matter in large numbers and when uniformly trained. They will have huge Pools which they can use with their necessarily narrow set of Secrets.