I’ve now run three sessions of Blades in the Dark for my “Doskvol Dalliances” campaign. The first two had some serious pacing problems in various ways; the third session finally started to get the pacing I want. The short version of this is that you should leave plenty of time for the Downtime phase. In a three-hour session, I like this to take about 30 to 45 minutes.
Session Structure
The game includes three main phases per score (which, in my game, is also equivalent to a session). The first phase is free play, which is general roleplay: the characters might deal with their personal connections, talk about how their crew has changed, etc. I like this to take about 45 minutes or so, though this isn’t super precise and largely depends on how much the players want to do this sort of thing. My players like it, and I’m happy to indulge that. This is part of letting the game breathe: explore the characters and the world and focus on the fiction we’re collaboratively creating.

However, at some point, we move into the score phase. This is what most folks seem to think Blades is about, because it mostly includes the heist itself. I like this to take no more than an hour, though we can go up to an hour and a half if free play didn’t take too long.
The score starts with selecting a plan & detail (potentially requiring a bit of gathering information). Unlike other games, Blades really pushes back hard against detailed planning because it has other mechanics to handle that. Figure out your broad approach, specify an element that the GM can use (e.g. the point of attack for an assault), and the GM makes an engagement roll. This abstracts out all the little things that can go wrong as the crew gets started, takes into account the differing advantages each size may have, and even incentivizes daring plans. “Fortune favors the bold” is a literal mechanic in this game! You could consider this a small mini-phase but I mentally track it as part of the score.
Then the full score proceeds, starting the crew at the first interesting bit. Maybe they get a “desperate” result on the engagement roll and start out with somebody already onto them, or maybe they get a “good” result and have gotten past some initial safeguards before spotting the next danger. During prep, I come up with a list of possible dangers or complications that I can throw in as appropriate, and I get them to help me with details to flavor things (or even get additional suggestions on complications from them).
Once they’ve (hopefully successfully) completed the score, we move to downtime. This is where pacing can really become a problem if I haven’t managed it properly up to now. This should take 30-45 minutes – it’s not like handing out XP and gold at the end of a dungeon crawl. Don’t give this part short shrift, because it’s not just about the actual payoff (including entanglements and heat). This is when the players start making plans: what long-term projects do they want to run? Does anybody want to deal with their heat or wanted level? Pay close attention because they are telling you want they want in the game. I have players who want to focus on the spooky, haunted elements, and that shows as they’re trying to learn rituals and study ancient tomes owned by societies of antiquarians.
Experience
In our second session (embedded below), we did a better job of selecting the score before we played. Unfortunately, we spent so long on the score that the downtime didn’t have time to breathe. Because we play on Saturday nights, running over time is not exactly ruinous, but it can put us in a difficult position if somebody’s getting tired or we have a hard stop.
For our third session, not only did they select a score in advance (something I really really recommend), I focused on getting us to complete the score with some time to spare. This meant I could really zoom in on what they wanted to do with their long term projects and let them talk a bit about the traumas their characters had experienced during the session. This felt much more enjoyable for me; take a look if you like!