|
|
Post by GSV3MIaC on Jun 15, 2014 8:24:44 GMT
So do we suspect that the current approach, which is to try to get everything equal % funded ( at least that's what it looks like) is a codsup, like the LIFO secondary market buying (now fixed), or do we subscribe to the conspiracy theory that it is a deliberate attempt to shaft autobid users by putting their money where everyone else feared to tread??
I'd have a poll, Burt I'm not sure I know how. 8>.
|
|
blender
Member of DD Central
Posts: 5,719
Likes: 4,272
|
Post by blender on Jun 15, 2014 9:50:58 GMT
So do we suspect that the current approach, which is to try to get everything equal % funded ( at least that's what it looks like) is a codsup, like the LIFO secondary market buying (now fixed), or do we subscribe to the conspiracy theory that it is a deliberate attempt to shaft autobid users by putting their money where everyone else feared to tread?? I'd have a poll, Burt I'm not sure I know how. 8>. You would need to be clear on the poll question. I do not think that FC is trying to shaft anyone. Autobidders are now protected by MBR on primary market and so will get a good-enough return from their lending. I think that FC's purpose in setting the Autobid algorithm is to get the primary market loans onto the loan book, and secondarily to ensure that liquidity is not unacceptable. Personally I am happy for FC to use the Autobid algorithm to optimise (in their terms but probably including stability) the primary market as it develops. It is a lever that I would grant them, as long as I know roughly how it works and can adapt. What we have to watch is the preference for the primary market at the expense of the secondary - because FC's prime directive must be to grow the loan book. Liquidity matters to the lenders, not to the borrowers or FC (except as it affects lenders). But to get this in proportion, there is none of the bad practice here of giving new lenders better rates than existing ones, or temporary bonus rates for new accounts. Now that really was shafting people.
|
|
|
|
Post by GSV3MIaC on Jun 15, 2014 11:14:23 GMT
OK so you, if I understand right, think it's intentional, but not an issue (loading most autobid money into auctions manual bidders are ignoring)? I guess there are three survey answers already then ..
|
|
|
|
Post by GSV3MIaC on Jun 21, 2014 9:55:28 GMT
I see in the other place that FC reckon autobid users are achieving higher averasgwe rates than manual bidders?! I'd love to check the statistics on that!!
I wonder if FC could release some anonymous 'lener stats_' like the downloadable loan book stats?
|
|
jimbo
Posts: 234 
Likes: 42
|
Post by jimbo on Jun 21, 2014 10:36:54 GMT
They did say over the last 6 months, and the rate rise would have helped them out when flippers dumped their high rate loan parts at 0% markup back in March and April.
|
|
blender
Member of DD Central
Posts: 5,719
Likes: 4,272
|
Post by blender on Jun 21, 2014 10:48:20 GMT
OK so you, if I understand right, think it's intentional, but not an issue (loading most autobid money into auctions manual bidders are ignoring)? I guess there are three survey answers already then .. Yes you understand me. Autobidders by definition do not make judgements about individual loan requests and accept anything that FC has passed fit. So I have no problem with the algorithm as long as it does not favour/disadvantage particular individual borrowers. I share your surprise at FC's claim about Autobid results, though it relates to this year and not 2013. I could conceive of many manual bidders being less successful on rates than Autobid though for Autobid to be generally better seems odd. There must be an awful lot of manual bidders who do not know what they are doing. But this is the FC which, when challenged on liquidity a little while back, produced statistics for speed of sale of loan parts which excluded those which were not actually sold and was based on a LIFO Autobid algorithm, so let's look carefully at the claim.
|
|
|
|
Post by aloanatlast on Jun 21, 2014 14:58:08 GMT
I see in the other place that FC reckon autobid users are achieving higher averasgwe rates than manual bidders?! I've thought for a while that they don't understand their markets, but now it looks like they're constructing their own delusions. Obviously this isn't based on rates of loan parts purchased. Probably it's based on overall portfolio returns averaged over the 6 months. So maybe the "autobid users" are mostly people actively investing at recent rates, without a lot of old stuff, while the "manual bidders" are sitting on piles of loan parts bought last year and the year before, loads of A+ at 6.7% and pre-MBR C at 9%, and this is dragging down their overall returns.
|
|
mikeb
Posts: 1,072
Likes: 472
|
Post by mikeb on Jun 21, 2014 18:30:56 GMT
now it looks like they're constructing their own delusions.That's spelled s-t-a-t-i-s-t-i-c-s. I saw that original posting, laughed, and hit "next". I'm sure someone who enjoys banging their head against the "brick wall of spin" will have a go. Or the FCA. Remember, this is in a thread where someone is asking "is my experience of earning 4% typical?". The response used to be: If you have earned less than expected, you are just "unlucky", you are on one end of a bell curve, averages blah blah etc. etc. This statistic was wheeled out to defend the (fairly obvious) deficiencies in using Autobid. Seeing as all the "smart" investors have turned off Autobid to improve their returns, I cannot accept that FC are seriously saying that Autobid is actually doing better ... So the only (charitable) conclusion is that Autobid is doing better, on average, over a cherry picked time frame, if you look at it in the right light, than manual bidders, because some manual bidders are putting in really REALLY stupidly low bids.
|
|
is
Posts: 108
Likes: 14
|
Post by is on Jun 22, 2014 17:42:10 GMT
Polls etc. is all well and good, but surely one can expect the full description of the algorithm that they are using to invest people's money? Trying to reverse engineer it from observations is never going to give anyone 100% confidence. I think the persistent unwillingness of FC to fully and accurately present the algorithm is a sign of bad faith, which is one of many reasons I've never activated it.
|
|
|
|
Post by GSV3MIaC on Jun 26, 2014 8:54:42 GMT
Apparently 6688 was the only underfunded / unfunded auction when the money rolled in last night, so it attracted more 8% MBR autobids than I have seen for months.
Going down ....
|
|
wysiati
Member of DD Central
Posts: 397
Likes: 86
|
Post by wysiati on Jun 26, 2014 12:33:51 GMT
Apparently 6688 was the only underfunded / unfunded auction when the money rolled in last night, so it attracted more 8% MBR autobids than I have seen for months. Going down .... When I last looked late yesterday evening there was well over £50k of the £150k unfunded. This morning the top rate was mid 10s% whereas some of the other smaller £100k A loans were still in 13-14% territory. Clearly the autobid mechanism provides far from even handed treatment in such scenarios and/or can be directed (clumsily?) to favour particular loan requests over others.
|
|
jm72
Posts: 109
Likes: 2
|
Post by jm72 on Jun 26, 2014 14:43:58 GMT
On the rate - I'm guessing that once filled and rates start to come down, bid-bots kick in? But then again, if so - why haven't the others pushed down. Maybe the amount of autobid kicked out a lot of the top bids?
On selectiveness - I would think so. For example, on the large property loans, autobid seems to pick up loans with lower rates, but not so much of these - maybe in order to keep 'autobid' cash free for the 2nd / 3rd tranches?
|
|
|
|
Post by GSV3MIaC on Jun 26, 2014 16:48:48 GMT
No, when rates come down, bid-bots pack up and go home. At least smart oens do.
Autobid (of the FC sort) doesn't seem to mind what the rate is, as long as it is above the Minimum asked for by the user (which may be at or above the MBR), and as long as the loan is 'less funded than other available loans' (or some such) in %age terms .. obviously a large loan can soak up more autobid cash than a small one, and a 'solo loan' which is the only one not fully funded, seems to be able to soak up almost all of it. This in the converse of the auction which started this thread, where 'filling it up early at 14%' was enough to persuade the FC autobidder not to bother at all (whereas bid bots were quite happy to play at 14%, 13,9% and all the way down to 'unacceptably low').
|
|
|
|
Post by thesnoop on Nov 5, 2014 18:31:26 GMT
Apologies in advance for newbie ramblings and for digging up an old thread...  I'm thinking about doubling the size of my bids ( to increase return on the Due Diligence I'm carrying out in assessing each loan and walking it it manually). Unfortunately I don't have enough money to simply double the investment quantities and hold to term without hugely increasing my max exposure to each loan. I gather from various other posts that AutoBid will only buy loan parts offered at par. So the idea is to selling on the surplus, put them first on the SM at a premium while earning interest then ship them at par to AutoBid after a few repayments to keep inline with my planned diversification. That got me thinking...can I be sure that I can ship things, how does Autobid really work anyway? Does it make the best currently available deal on the SM for the lender that meets their criteria ? Say a user configures AutoBid to buy A band loans above 8%, and sets it away to invest. ( How does the tool decide whether to buy from the PM vs SM anyhow ? ). Now say there are 100 A band loan_parts being offered at par for an 11% buyer_rate, and there are also 100 A band loan_parts being offered at par for 10%. Would Autobid buy all of the 11% parts before buying the 10% parts? I'm assuming it would...meaning a rising trend in rates could prevent older parts being sold at par..? If the 100 parts at 11% in the example above were split across multiple loans and lenders, would older loan_parts have precedence over newer ones being offered at the same rate?? If yes, is this be based on an absolute time to market for each loan_parts at a given rate...Does that effectively make it a race to get you parts to market if there is a lot of competition at the highest current available rate?
|
|
blender
Member of DD Central
Posts: 5,719
Likes: 4,272
|
Post by blender on Nov 5, 2014 19:48:01 GMT
The mysteries of Autobid are known only to a few of the trusted servants of the FC temple. We know some things and we think we know a few more, but much is from observing the behaviour of a black box.
Because Autobid tries to lend money out quickly and because it can only buy once from each loan, it tends to buy a loan part which is close to the maximum for the lender (diversity setting and account size). According to FC the order in which loan parts are considered is random - ie it does not matter how long your part has been on the SM, it used to be LIFO, but no longer. Autobid seems to deal with a lender at a time and makes multiple purchases, because Autobid purchases at often close together for one lender in my sold parts. It does not (edited thanks GSV) seem to look for the highest interest rates, just an adequate interest rate.
How it balances the PM and SM we do not know. Presumably Autobid runs on some periodic basis on both the PM and the SM, and we have suspected that the balance between the two is controlled by some big lever which FC can operate according to their view of the market need. But we do not know and are not told.
I hope this helps - others will have better or different understandings.
|
|
