Building the Bear Case Against My Own DraftKings Position

Holding a position is not the same as having an opinion. Positions come with a price and a point in time when you are right or wrong. Early in my career, I learned that being too early, even if ultimately right, is still being wrong; for example, being short Autonomy when it was acquired by HP, only to be “exonerated by accounting fraud” a year later. The discipline that separates conviction from wishful thinking is simple to describe and hard to practice: build the case against your own position with the same rigor you bring to the case for it.

Constructing a credible bear thesis is slow and tedious. It is unpleasant to spend an afternoon arguing against the money you have already put to work. Often, we skip it and maintain conviction in a position we never stress-tested.

I have been bullish on DraftKings for a while. This week, I leveraged the Daloopa Bull/Bear skill in Claude at the ticker and asked it to argue with me.

What the skill does

The Bull/Bear skill is one of Daloopa’s Skills — pre-built, specialized instructions that run on top of Daloopa’s MCP. The skill takes a ticker and does the work that a careful analyst would be required to do before a demanding PM considers initiating a position. It answers the question, “Where could you be wrong?” It pulls eight quarters of source-linked fundamentals. It selects the KPIs that matter for the business model rather than using a generic template. For DraftKings, that means sportsbook handle, hold, monthly unique payers, and ARPMUP. It does not report ARR or net revenue retention because it is not a SaaS company, and those metrics are irrelevant. It pulls management guidance and the latest filings. Then it constructs three bottom-up scenarios with probability weights and per-share targets.

Every number links back to the exact line in the filing. I did not open a spreadsheet. The whole thing ran from a company lookup to a finished framework in minutes.

The output, in one table

DraftKings closed at $25.70¹. Enterprise value lands near $14.0bn². That is 20.4x trailing Adjusted EBITDA of $685mm³ and 2.2x trailing revenue of $6.29bn. The most recent quarter showed the inflection the bulls point to – revenue of $1,646mm up 17% year over year, with Adjusted EBITDA of $167.9mm.

Scenario outputs are computed by the skill from source-linked inputs.⁴

The base case is the most likely outcome. The distribution skews positive, mostly because the stock has already compressed from the mid-forties into the mid-twenties. None of that is the point of this post.

The bull case I am underwriting

Before the bear column earns its keep, here is the case it has to beat. It is the thesis I already hold.

Start with the inflection. Sportsbook revenue ran at $1,095mm last quarter on handle of $14.1bn. iGaming added $461mm and has climbed steadily. iGaming is the higher-margin half of the book and has grown steadily, even as it gets less attention than the Sportsbook headline numbers. Management guided full-year revenue to $6.5bn to $6.9bn and Adjusted EBITDA to $700mm to $900mm. The base case asks for $7.6bn and $1,140mm in FY27. That is a step up from the guide and not a leap.

Then the leverage. Cost of revenue and product and technology barely move with handle. When hold grows, incremental sportsbook revenue is close to pure margin flowing down to the bottom line. That is why one good hold quarter can swing the number by hundreds of millions. The bull case is that parlay mix and live betting and trading discipline lifted structural hold for good with more to come. The bull case is also that one large state legalizing online casino re-rates the entire iGaming line.

The point is the bear column

Here is what the bear case forced me to sit with. Four things, all of which I had been quietly waving away.

First, hold is key. Hold is defined as gross revenue divided by handle. For example, if $100 million was wagered at DraftKings during the Super Bowl and DraftKings kept $8 million after paying out winning tickets, that is an 8% hold. It does not account for promotions, operating expenses or state taxes.

Sportsbook hold increased from 6.4% in Q1 2025 to 7.8% last quarter. That climb is most of the bull case. But the climb was not a straight line. Hold printed 8.7% in 2025Q2, fell to 5.2% in 2025Q3, then recovered to 8.0% and back to 7.8%. The 2025Q3 quarter is one that shows the house sometimes loses. On handle of $11.4bn a hold of 5.2% pulled sportsbook revenue down to $596mm from $998mm the quarter before. That one bad run of outcomes flipped Adjusted EBITDA to negative $126mm and operating income to a $272mm loss. A few weeks of football going the customers’ way did that. Hold is not a slow dial I can count on.

On roughly $54bn of trailing handle every 50bps of hold is about $270mm of near-pure-margin revenue. If 7.8% is a flattered run and not a structural floor, the model breaks. I want to believe parlay mix and live betting made the gain permanent. The bear case reminds me that I do not know that yet.

Second, the profitability is an Adjusted EBITDA construct. GAAP operating income in 2026Q1 was effectively break-even on $1.65bn of revenue. The Adjusted EBITDA of $167.9mm is derived by adding back $65.2mm in stock-based compensation and a new advocacy line of $26mm in lobbying spend. SBC is substantial. Stock-based compensation totaled about $326mm⁵ across the four trailing quarters, close to half of the $685mm of trailing Adjusted EBITDA the headline multiple is built on. My true economic margin of safety is thinner than the headline margin suggests.

Third, the Predictions bet is genuinely two-sided. Management wants leadership in sports prediction markets by year-end. As a bull, I read that as TAM expansion. The bear case reads the same sentence as a regulated, competitive, cash-consuming distraction against deep-pocketed competitors- Kalshi, Polymarket, and Robinhood. The advocacy and other legal costs line ran $2mm in 2025Q4 before jumping to $26mm in 2026Q1. That is a thirteenfold step up in a single quarter before the product has a revenue line to show. Both readings fit the facts. That is exactly what makes it a risk and not a footnote.

Fourth, the tax line is moving the wrong way. Cost of revenue carries the gaming taxes the states keep raising, and none of it gets added back. In 2026Q1 alone, gaming taxes rose $86mm year over year, after a $142mm increase in 2025Q2. An $86mm tax increase in one quarter is about half of the $167.9mm of Adjusted EBITDA the same quarter produced. Every new state tax hike and per-wager fee lands here and compounds against the hold story. A higher-tax map and a federal wagering excise are the version of the bear case that does not need a single bad Sunday.

Why this matters for holding the position

Reading the bear case did not make me less bullish. It made the bullishness legitimate. I walked away with a list of metrics that I knew I had to closely monitor to ensure my bullish thesis should not be re-evaluated.

1. Watch hold for two quarters. If net revenue margin reverts toward 6.5%, the operating-leverage story is flattered, and I am early and/or wrong.
2. Watch iGaming legalization. A single large state legalizing online casino is the cleanest path to the bull case. If no new states legalize, the base case caps the upside.
3. Watch the Predictions burn against Predictions revenue. Advocacy and legal spend climbing without a revenue line to show for it is the bear case arriving in slow motion.
4. Watch the gaming-tax line as a share of revenue. If state tax hikes and a federal excise keep lifting cost of revenue faster than hold lifts the top line, the margin structure erodes no matter how the games break.

If hold reverts and the Predictions spend climbs without revenue, that’s the signal to revisit the thesis.

It is hard to hold a position honestly if you have only ever argued one side of it. The Bull/Bear skill builds the other side for you in minutes and sources every claim. But the real job is knowing the list of things that would change your mind — and the skill hands you that list.

Methodology

Every figure in this post links to its underlying filing line item through Daloopa. The trailing baselines, the enterprise value build, and the scenario math are computed from those linked inputs. The framework took minutes to generate. The confidence it gave me is the part that lasts.