How to roast a kabocha squash

Learning to use new ingredients can be daunting, especially vegetables where the preparation and cooking techniques are unfamiliar and the guidance available is minimal. Fortunately, kabocha is fairly forgiving. I’ve settled on a reliable technique that works well and requires no additional ingredients like oil or seasonings. My typical kabochas are around 1.5 kg, so as long as yours is somewhere close to this, the timings should work well.

Preheat an oven to 400 ºF (200 ºC). Place the kabocha on baking sheet with a wire rack and roast whole for 20 minutes. Take out the kabocha and, when cool enough to handle, since it down the middle, right through the stem (this make take some strength). The point of this initial roast is to soften the kabocha so it’s easier to cut.

Once cut, use a spoon to scoop out the seeds and stringy fibers from the middle.

Place each half cut side down and slice into in half again, again going through the stem. You should now have four quarters.

Now cut each one of these quarters in half again, so you have eight large wedges.

Turn each wedge face up.

Starting with one point of the wedge, slice it into thirds or fourths, depending on your preferred size and the size of your kabocha.

Place wedges on a wire rack on a baking sheet and roast for 40-45 minutes, until you start to see some dark brown areas at the tips of some of the kabocha wedges.

Take out the tray and let cool (if you want). Wedges can be reheated for 5-10 minutes in a 400 ºF degree oven if you prefer them warm, stored in the refrigerator, or frozen. If you freeze them, I suggest freezing on a baking sheet prior to storing in a bag, since they otherwise tend to stick.

Sourdough’s Slow Recovery

I’ve kept my sourdough starter in suspended animation for years, with periodic feedings. A couple attempts at using it for leavening largely failed, so it was relegated to the role of a flavor enhancer in my yeast-based breads. Given the surge in sourdough popularity associated with the COVID-19 pandemic, my interest was rekindled and I brought the starter back to life with a series of room temperature feedings. Through my web-based re-education, I realized that my usual approach of replacing half my starter with equal parts flour and water was amateurish and inadequate. Even after accelerating these feedings to twice a day, I learned that the starter would be unable to develop adequate strength unless I replaced a greater proportion of the starter at each feeding. The so-called 1:5:5 feedings (one part starter to 5 parts water and 5 parts flour) seem to be the gold standard for establishing a robust starter.

My starter was rising regularly on this regimen, so I decided to give it a try. I found some detailed instructions on YouTube from Full Proof Baking which seemed rather laborious, but the results looked impressive. I decided to give it a go. The next day, my first loaf didn’t look bad, though didn’t quite have the rise I expected. To be fair, I hadn’t gone through every last step of the various forms of folds, so perhaps some additional effort was warranted.

My next attempt was a bit higher-rising, but something was off. The base was dense and the crumb was a bit gummy. I had handled the dough somewhat roughly in the transfer from the banneton to the clay baker, and suspected this may have been part of the problem.

My next attempt proved more successful. I’m not sure what was due to better handling vs. a more mature and robust starter, but the differences were obvious.

Look at the difference in rise and texture. The taste was noticeably better, with great flavor and a nice bite to the middle. Loaf after loaf has yielded similar results, using either 75% or 80% hydration.

My dough prep is simple: I mix 360 g flour with 280 g water and autolyse for 30-60 minutes. I then add in 10 g of salt and 80 g of stater and mix by hand for a few minutes until well combined. After 30-60 minutes, I’ll pick up one end of the dough in the bowl, stretch it up, and fold it over the middle. I rotate 90 degrees and do the same. I continue this until I’ve performed 8 folds, the flip the whole mound of dough over and let it rest (covered) for an hour. I do this a total of 5-6 times, resting 45-60 minutes between each set of 8 folds. By this time, the dough is noticeably fuller and puffier. I turn it onto a floured counter and use a bench scraper to sweep it into a roughly round shape. I fold the right side across the middle, then do the same with the left. Starting at the top, I roll it into a cylinder and pinch the seam shut. I transfer it to a banneton well dusted with rice flour (seam side up) and cover with plastic wrap or a shower cap-like bowl cover. It goes into the fridge at least overnight or as much as two nights. I transfer it into a preheated covered baker and into the oven it goes.

Snce my family prefers a crispy, but not particularly thick crust, I bake covered at 500 ºF for 5 minutes, 450º F for 30 minutes, then 450 ºF for 5 more minutes uncovered.

Resurrecting sourdough

I have kept a sourdough starter in my fridge for at least five years or so. I knew very little when I started, but followed a range of web-based instructions for how to get started. I quickly tired of the twice-daily “feedings” of flour and water (not to mention the perceived wastefulness of discarding half of it each time). When I discovered the starter could be maintained in the refrigerator, with feedings spaced to as much as once a week, this seemed like a natural choice.

The feedings ended up coinciding with various bread-baking ventures, so I’d just add the half of the starter to my bread recipe whenever needed, and replenish the used flour and water. Since I used a 100% hydration starter (equal amounts of flour and water), the math was easy. As the refrigerated starter wasn’t particularly active (“unfed”, in sourdough terms), I used instant yeast in whatever amount was called for in the recipe, assuming the starter wouldn’t add much leavening power.From time to time, I’d try various experiments with using the starter as the main source of leavening, but it was so much weaker than the commercial yeast, it didn’t seem worth the effort.

With home sourdough mania spreading almost as fast as COVID-19, it seemed like the right time to give it another shot. After some re-reading, I realized that I had probably never been letting my starter become active enough to serve as the main sole source of yeast. I brought it back to room temperature and started the twice-daily feeding regimen again. Since we have started composting, the “discard” didn’t seem so offensive, and I’m spending a lot more time at home regardless. I had converted the starter to whole wheat (which I’m told favors rapid resurrection of the yeast), but started converting it back to white since I now have plenty of white flour and it’s reportedly more predictable once the yeast starts reactivating. Indeed, within a few days, the starter was visibly active, full of a bubbly texture that rose and collapsed between feedings. I tried the “float test” (adding a small amount of starter to a glass of water to see if it floats) and, whlie the results were mixed, I decided to charge ahead.

I added 100 g of starter to 350 g of bread flour, 8 g of salt, and about 2 g of diastatic malt powder (for some extra insurance), along with 230 g of water (for a 70% hydration dough). After some vigorous stirring and stretch-and-fold kneeding, I let it rest. It took about 2-3 hours to double (a bit longer than commercial yeast, but nothing egregious ). I shaped a batard and put it in a proofing basket. About an hour later, it was ready to go into my preheated covered baker. I baked covered at 500 ºF for 12 and a half minutes, 450 ºF for the same amount of time, then left it uncovered for the last 10 minutes to allow the crust to crisp a bit. The internal temperature was nearing 210 ºF at this point (a bit hotter than ideal), so I may cut it to 20 minutes of covered baking next time, but the results were much better than I had anticipated. The flavor was not particularly sour, but it had a depth and complexity that was just not possible with the instant-yeast equivalent.

I’m sold, at least for now. While commercial yeast can be more forgiving and predictable, I’m eager to try new experiments with this beast. In fact, my bagel recipe is currently, I hope, slowly leavening a batch of bagels in my refrigerator right now. The residual fear of a failed-rise led me to proof for 15 minutes prior to refrigerating and to extend the refrigeration time from 24 to 36 hours, but tomorrow will be the real test.

Roasting Kabocha: Fast or Slow?

Most instructions on the web suggest roasting kabocha at about 400° F. My usual protocol involves baking the whole squash for about 20 minutes at 400 °F, scooping out the seeds, cutting into 24-32 wedges (depending on size), then roasting for 40 more minutes. At this time of year, I’ve found a regular bounty of kabocha with the ideal, somewhat dry egg-like texture, but some have clearly been sweeter than others.

A common technique to increase sweetness of sweet potatoes (including Japanese sweet potatoes) is to bake them at a lower temperature for a longer period of time, giving the starches a greater opportunity to break down. I wondered if the same rules could be applied to kabocha.

With my most recent sqaush, I roasted some wedges at 325 °F for 80 minutes and some at 400° F for 40 minutes (the batch on the left of the image is the higher temperature). The 400 °F wedges were clearly darker with greater caramelization. To my surprise, if anything the higher temperature yielded sweeter squash. It’s possible that the lower temperature specimens simply needed more cooking, but for now, I’d suggest sticking to 400 °F.

Update: I did another test, this time cooking part of the squash at 400 °F for 45 minutes and part at 250 °F for 150 minutes. Both versions were well cooked and dry, but the texture and flavor was definitely superior in with the higher temperature. The lower temperature squash was grassier and less sweet compared to the higher temperature. Stick with 400 °F.

Half-Moon Cookies

While pre-packaged baked goods now dominate many childhoods, when I was growing up, most desserts came fro the neighborhood bakery. I have no idea how much time has enhanced my memories of these treats, but the most tangible losses are those that cannot be replaced in today’s markets. One of my favorites was the half-moon cookie. The oversized cake-like cookie had one half covered in white vanilla frosting with the other coated in chocolate frosting. On special occasions (like Halloween), the white frosting might be tinted a celebratory orange, but the concept was the same.

In our household (and hopefully most). there were rules to the half-moon cookie. While you could approach it in any way you wanted (chocolate first, vanilla first, or some combination), in no way were you allowed to selectively eat only one half. Nor could you save one half for later. You approached it when ready, then devoured the whole thing in one sitting. Now some may think they’ve seen this cookie before as the “black and white”. Nonsense. That is a completely different species. While there are superficial similarities, the black and white (at least those I’ve seen) are much smaller, with a denser cookie and a firm icing rather than a buttercream frosting. The half-moon is really more of a cupcake-cookie hybrid.

I thought I had seen my last half-moon years ago, but I some web-searching led me to a reportedly original recipe for the Hemstrought Bakery Half-Moon Cookie. While I don’t know anything about the Hemstrought Bakery, the pictures looked promising and it was certainly in the realm of possibility that my local bakery (Lederman’s in Newton, MA) had adopted a version of this bakery’s creation. The recipe looked doable, though I found a version that simplified the frosting recipe by deriving the chocolate frosting by adding cocoa powder and milk to half of the vanilla instead of creating a completely separate batch. Even this simpler recipe (which had been scaled down from the bakery’s published instructions) seemed to create an intimidating amount of cookies that risked overwhelming the capacity of even my ravenous children. I scaled it down further and set to work. The simplified recipe instructed me to scoop the cookie batter using a #20 (three tablespoon) scoop, but it quickly became apparent that this produced monstrosities that risked tempting eaters to violate the rules of engagement (see above). I downsized to a #40 (1.5 tablespoon) scoop which was considerably more reasonable.

The result was exactly as a remembered: a delicious chocolate cake base with just the right amount of frosting. They are not too sweet, with a texture that’s somewhere between a cake and a cookie. Here’s the recipe I finally settled on (updated on May 24, 2020).

Half Moon Cookies

  • Servings: 18 cookies
  • Difficulty: moderate
  • Print

A soft chocolate cookie, half with vanilla frosting, half with chocolate froisting


260 g all purpose flour
1/2 teaspoon baking powder
1 teaspoon baking soda
112 g butter
225 g sugar
30 g cocoa powder, sifted
1/8 teaspoon salt
1 egg
2.5 g (1/2 teaspoon) vanilla extract
180 g (3/4 cup) milk

Vanilla frosting:
84 g (6 tablespoons) butter, at room temperature
300 g powdered sugar, sifted
50 g milk
5 g vanilla extract

Chocolate frosting:
Half of the vanilla frosting
20 g cocoa powder
9 g (1 tablespoon) milk


  1. Preheat oven to 350 degrees.
  2. Line two baking sheets with parchment paper.
  3. Whisk together flour, baking powder, baking soda and salt in a bowl and set aside.
  4. In a large bowl or the bowl of a stand mixer fitted with the paddle attachment, beat butter at medium speed until smooth.
  5. Add in sugar and cocoa powder. Mix at low speed until combined, then increase to medium-high until light and fluffy (3 minutes).
  6. Scrape down the bowl with a rubber spatula.
  7. Add in egg and vanilla and beat at medium speed until combined (30 seconds).
  8. Scrape down the bowl.
  9. While mixing at low speed, add in 1/4 of the flour mixture, followed by 1/3 of the milk, and continue alternating until all the flour and milk is added. Do not overmix.
  10. Scrape down bowl and fold a few times with a rubber spatula to ensure mixture is homogenous.
  11. Use a #40 (1.5 tablespoon) scoop to portion the dough in mounds on the baking sheet, no more than 8 cookies per sheet.
  12. Optional: chill the pan and cookies for about 15 minutes prior to baking. I find this helps the cookies from spreading too much during baking.
  13. Bake for about 12 minutes until edges are set and the center has a slight resistance. Your finger should leave an impression.
  14. Let cool for a few minutes on the baking sheet, then transfer to a rack to cool completely before frosting.
  15. While the cookies are cooling, make the frosting: you can use a stand mixer, but I find it easier with a hand mixer.
  16. Combine the milk and vanilla for the vanilla frosting in a small bowl.
  17. Beat the butter until smooth, then add the powdered sugar.
  18. Start mixing at slow speed, and slowly add the milk and vanilla mixture until the a thick, but spreadable frosting is achieved.
  19. Portion out half the frosting into a separate bowl, add the additional milk and cocoa powder, and beat until well combined.
  20. Flip the cookies over so that the flat side is facing up, and frost half the circle with vanilla frosting. Once all the cookies are frosted, frost the remaining half with chocolate frosting.

Good crusty bread is clocher than you think

Unlike my youngest child, who frequently requests the crust of even the softest breads removed before she will touch it, my search for intensely crusty bread seems boundless. The crustier, the better, even if it leaves my mouth ravaged by a trauma only rivaled by several bowls of Cap’n Crunch. Since crusty breads are so ephemeral, they make the ideal home baking project.

Steam is key to developing a crust, and I’ve created reasonable results with a hot panful of water accompanying the loaf in my oven, but I’ve been curious if a cloche (a clay baker) would come…clocher to achieving my goal. It seems so simple – the dough in the confined space of the cloche serves as its own steam source, no precarious shuffling hot cast iron required.

After acquiring a batard-shaped clay baker, I set to work. Using a simple lean dough (flour, water, yeast, salt), I preheated the cloche for 30 minutes to 500 ºF then carefully transferred my shaped loaf into the hot vessel before making a few quick artistic slashes. I replaced the cover and baked for 20 minutes covered, then reduced the temperature to 450 ºF for 10 minutes. For the final 10 minutes, I removed the lid to allow the curst to firm up, then let the resulting loaf cool

I’m impressed. The crust is fantastic and, unlike some uncloched versions, doesn’t fade away completely after a rest. The bread itself retains sufficient moisture despite the long bake, as it spends much fo the time enclosed in clay.

Cookie Pie

In the midst of the COVID-19 lockdown, my family has been inhaling desserts at an alarming rate. I have a simple, scalable cookie dough that has been the basis for many forms of cookies:

  • 112 g (one stick) unsalted butter, slightly cooler than room temperature
  • 100 g light brown sugar
  • 60 g sugar
  • 1 teaspoon vanilla extract
  • 1 large egg
  • 1/2 teaspoon baking soda
  • 1/8 teaspoon salt
  • 175 g bread flour (all-purpose is fine also)
  • 100 g chocolate chips, M&Ms, or other add-ins

The butter and sugars are creamed together until fluffy, followed by the egg, vanilla extract, and a mixture of the flour, salt, and baking soda. Just as the flour mix is almost incorporated, the add-ins are mixed in until evenly distributed. The dough does well if rested in the refrigerator for at least a few hours. It can keep for several days if needed, and can be scooped into balls and frozen as well.

To make cookies, you simply scoop out your desired size onto a parchment-lined cookie sheet, and bake at 350 ºF for 10-15 minutes depending on how cold the batter is and how you like your cookies. They do best when taken out of the oven as they just starting to brown on the edges, but not quite finished cooking (they can finish with a 3-5 minute rest on the hot baking sheet prior to being moved to a cooling rack.

Sometimes, cookies get a bit old, and an alternative is to just pack the dough into a pie plate and cook at 350 ºF until it’s set around the edges and not totally unbaked in the middle, about 25-30 minutes. This “cookie pie” has a unique texture and is a big hit most. The texture is best when you give it a chance to full cool and set.

Kabocha vs. Buttercup

Squash season is in full swing. The shelves usually limited to bland butternut and unispiring acorn varieties are now bursting with a varied and colorful bounty. I typically ignored these generally unattractive gourds until the past year or two, when I discovered the unique taste of the kabocha. Dense, flavorful, and almost potato-like in texture, it is a far cry from the mushy and stringy butternut that I typically associated with squash.

But not all kabocha are the same. In prime season (now), it’s considerably easier to find the large, heavy kabocha that are brimming with a modest sweetness, chestnut-like flavor, and dry, dense composition. Out of season, kabocha can occasionally develop a wetter texture with a mushy texture more reminiscent of the pedestrian varieties that contaminate the squash section year round. The challenge that it’s often difficult to identify whether a kabocha will be fantastic or uninspired. Heavier squashes seem to be superior, and I typically find that rougher skin and larger skin seem to be predictive of the better versions. I typically roast my squash by cooking it at 400 °F for 20 minutes, then halving it, scooping the seeds out, and cutting into thick slices. These thick slices are then roasted for 40 more minutes. I don’t find that oil is necessary, though some may prefer the flavor. The first cut through the squash can offer clues as to what is to come. A green ring around the inside is typically an ominous sign, as are beads of liquid that appear after slicing. Sometimes extending the second roast up to one hour can help, as can resting the cooked squash overnight in the refrigerator, but it will never near the flavor and texture of an optimal squash.

Adding to these challenge is that kabocha has a close relative, the buttercup, which more often takes on this softer texture. I also find the flavor tends to be more vegetal. The classic buttercup looks somewhat different from kabocha. They are squatter with a boxier shape compared with the rounder kabocha. The base has an open ring with a pale, thin skin in the center, and the surface is geenrally smoother. The flesh tends to be a brighter yellow vs. the more orange hue of the kabocha.

The kabocha (left) has a button-like base in contrast to the open ring on the buttercup (right).

The kabocha (left) is rougher in texture and rounder in shape than the boxier buttercup (right).

While these classic descriptions seem like an easy way to differentiate these two varieites, the reality is that there is a spectrum in this family that makes distinguishing these two archetypes more challenging. Some kabocha-looking squash have a somewhat open ring at the base instead of the classic “button”. There are variations skin texture and shape in both varieites, and it’s not surprising they are often confused and mislableled. Some websites claim that anything but a pure closed button base identifies the squash as a buttercup, but I find it’s not so simple. There is considerable variety in these forms, which I suspect reflects the fact that there are not only two varieites, but a range that may be closer or further from the classic buttercup or kabocha.

I bought two squash at the same supermarket. One was clearly a buttercup, with the classic open base and squared off shape. The interior is asymmetric, with the seed-packed core much closer to the base than the more centered kabocha. The texture upon roasting was softer, with a sweet, but more vegetal taste that was a bit unappealing.

The classic buttercup has an off-center core closer to the base.

The kabocha has a more centered core, a deep orange flesh, and a rounder shape.

The kabocha, despite the slight ring around the button, was clearly a kabocha in prime form. The darker color, denser texture, and chestnut-like sweet flavor was a refreshing contrast.

Ring Molds for Burger Buns

I have always hand-shaped my burger buns, but after buying a set of ring molds for English muffins, I saw the opportunity to create perfectly identical buns. I took my usual burger bun dough and portioned it into a dozen round balls. I then lightly oiled the inside of each ring mold and placed them on a parchment-lined baking sheet.

It quickly became apparent that each ball was far too small to fill the ring mold, so I took a slightly oiled plastic bench scraper and flattened each ball into a disc, placing one in the center of each ring. I let the dough rise for a bit over an hour, brushed the tops with egg wash, then baked as per usual.

The buns did indeed generally fill the rings. Though it would be a bit generous to say they were all identical, they were at least the same diameter. Furthermore,the cylindrical base (while a bit odd looking), did make them a bit easier to split by providing some landmarks.

I’m not sure the technique worth the effort for burger buns, but it could prove useful for ensuring a fixed size when needed.

Rice Flour for Fried Tofu

The mention of tofu can induce a wide variety of reactions, often a grimace from many Americans. Deep frying is a reliable way to make tofu more appealing to many palates, but can be tricky to execute. Should the tofu be fried “naked”? Coated with cornstarch? Battered? I’ve tried a range of methods, but recently came across the suggestion to use rice flour, which reportedly helps fried tofu stay crispy long after frying.

I gave it a shot with Arrowhead Mills White Rice Flour. I cut soft tofu (not silken, but softer than the usual firm texture I buy) into approximately 1 inch cubes, dusted them well in the flour, and fried them for about 8-10 minutes in 350° F oil. The test audience was my kids, who normally turn up their nose at tofu. They gobbled up the cubes without complaint. True to the claims, the cubes did in fact stay crispy, even after being refrigerated for a few days and reheated in the oven. The white rice flour was less clumpy than conrstarch. The only potential detraction was that the Arrowhead Mills brand was a bit coarse, with an cornmeal-like texture. Next time, I’d try a finer grind.