Lagarith supports 4:2:0 (YV12) and 4:2:2 (YUY2). Standard HuffYUV supports 4:2:2 (YUY2) but not 4:2:0 (YV12) - to use YV12-mode HuffYUV you have
to use ffvhuff, either through ffdshow's VFW interface, ffmpeg, or other libavcodec-based encoders (mencoder, Avidemux, VLC, etc.). Using YV12-mode and adaptive Huffman tables makes it fall much closer to Lagarith, filesize-wise, than it does standard HuffYUV, and it decodes a lot
faster than either one of them.
FFV1, on the other hand, supports 4:4:4, 4:2:2, YV12/4:2:0, 4:1:1, 4:1:0, and RGB32 (which are all the ones listed in ffdshow's Configuration dialog; it may support even more colorspaces than that) - like ffvhuff, it's only available in libavcodec-based solutions. It gives similar size ratios as Lagarith, for reference. FFV1 also supports 16bpc, but I never got it to work right with the CrowdRun sequence, although that was last August/September
- it may have gotten completely fixed by now. It may also have 10bpc and 12bpc support, but I'm not entirely sure on that point (nor do I know, if it does support them, if it works as well as the default 8bpc mode does). That thread seems to hint that it does.