See Long-Term Analysis.

Imagine for a moment this is the only gambit card in your deck. What do you get? A free Murder* every turn for the rest of the game.

Even if this had a restriction like only shuffling it into the gambits if cast from hand, this would still be a Murder* with rebound, which is a major blow worth well beyond mana value 4.

This ought to demonstrate a weakness in the gambit concept: The gambit deck is too consistent. Shuffling cards into it once er turn gives you fairly good control. Your opponent would have to mess with it to even make the randomization by shuffling matter, and even then a free card per turn is a free card per turn. E. g. Long-Term Analysis doesn't seem better, because it is just a delayed free random spell from your deck, which can be manipulated to be whatever is the most like an Eldrazi titan in your format.

The solution seems to be to make it really more delayed/risky by adding other cards to the gambit pile. It cannot be just random cards from your deck, because those could be Ulamogs, and we don't want to make the effect even more swingy, so I suggest something like Gunk (see Habitual Overthinking) - worthless or low-worth virtual cards.

Now you can have some blanks in your gambits, but maybe even mix a drawback card, a virtual One with Nothing (maybe not quite as bad) in there.

So what about any of:

• "shuffle ~ and three blanks into you gambits. (Blanks have no effect.)"
• "shuffle ~ and a misfire into you gambits. (Misfires exile the top card of your remaining gambits.)"
• "shuffle ~ and a two hazard into your gambits. (Hazards deal 2 damage to you.)"

Wordiness is an issue, but I feel something is required to make a gambit a gambit and not a safe single-turn investment.

A reverse word-saving option is to just pay the mana cost. That's what we do to balance cards. Even then this card is still by itself essentially a Murder* with cantrip/(infinite) flashback/buyback etc.

Agree; this is undercosted.

It certainly could be cost appropriately, to not need the casting cost every time. Just - that cost is probably 7 or 8.

An activaiton cost for the gambit seems the simpler solution for that. Though you do still have the ridiculous repeatablity, without the downsides epic had. "Oh what a surprise, I drew my 'kill a cretaure a turn card' - I'll do that" every single game? Not so fun.

I significantly changed the mechanic based on a fusion of my own thought process and SecretInfiltrator's feedback. As written, it was definitely not a good idea previously; it may still not be.

Hmmm. Ok, this is now very different. Let's see - you only ever get to cast your gambits twice at most. It gambits a random card too, that's gonna be odd. Ending up with a land in the gambit deck is a dead gambit, which will be annoying.
And you do now have to pay to cast it. So this is now kind of like a less good variation on flashback - in that you don't get to choose what you're getting back each turn. But it's also pseudo card-draw, since you also get to ((at some point) cast the top card of your deck, if you didn't overload on it..

Ok, this is pleasingly different from other mechanics; and has a real "Ok, let's just throw stuff and see what happens!" kind of randomness, which is pleasing. Not so much "I have planned for every outcome" gambit, more of a "Bask in my glorious chaos!" gambit. So yeah, I like it. But I'm a small demographic :) Whether the card advantage (and it's a good advantage! This is, what, removal-twice-and-cantrip?) is going to outweigh the "Oh noes! Randomness!" feel, and make those boring other demongraphics choke it down, I don't know.

There's certainly tuning arguments to be made for the costs. But they're probably in the right ballpark here; and it'd need testing to see whether it needs to go up or down a bit.

Paying mana for the cards you play always helps with balance. :)

I'm not certain the gambit cost is a good idea. It adds a whole line to the already long reminder text. Heaving that valve for balance is nice, but the word count currently is the bigger issue.

The shuffling is an issue here, especially since the "put on bottom" part of the resolution could help you get o top of the chaos, but gets undone easily. Unfortunately I see no easy solution right now.

Removed separate Gambit cost.

The risk of adding a land to your gambits is meant to be a built-in drawback/balancing method.

I know it's been commented on that the shuffle is a mechanical downside (which it is, shuffling is slow and annoying) - but it's also a nice outlet for people to somewhat control the randomness. Don't like the gambit available to you this turn? Gambit something else an see what you can shuffle up.

The Gambit deck would generally be much smaller than a full deck. I imagine shuffling would be quite a bit faster.