wapping35
Member of DD Central
Posts: 385
Likes: 210
|
Post by wapping35 on Oct 22, 2014 19:16:17 GMT
Kevin sorry to be a pain, but if it is not a glitch and I am v willing to accept it isn't can I please clarify (per my email to your CS chaps on Oct 20th) how a YR at 5.90% set on Oct 15 is any different (or treated any differently) from a YR 5.90% on Friday Oct 17th.
I sent a screen shot showing this (I can send it again if needed) and indeed your CS guys saw it on my account when I called them on Friday morning on Oct 17th. (I believe I spoke with Jordan).
As I mention in the email I also did have a separate MR 5.9% on Oct 17th(which I assume was a 2dp one and thus say really at 5.89% and ahead of the 5.90%). As I say I see the logic of the explanation for that one, but it only explains part of what I saw not the whole picture.
I should say whilst I have used the word "glitch" (Sorry). I have not used BS nor lying. Indeed the explanation did explain part of the issue as I acknowledge.
All I am asking is an explanation of why the system treated a YR at 5.90% on Oct 15 as a higher rate than a YR 5.90% (my v own YR) on Oct 17th.
I believe if we can get an explanation for that it will be clearer, at least to me.
p.s. I should add if the answer is we do not know why this happened and as it has now resolved itself we cannot investigate further, that would be okay. Not ideal, but okay since I now know how to deal with the issue if it ever happens again. i.e. As on Saturday Oct 18th just reset the offer at the same rate.
|
|
|
Post by GSV3MIaC on Oct 22, 2014 21:27:43 GMT
Psychic debugging ... I guess at some point in time either the algorithm, or the hardware, or whatever changed enough to give a different answer for the contract rate equal to a visible 5.9% lender rate. That would explain everything everyone has seen I think. Whether some programmer recompiled single precision as double, or the java got updated, or whatever, I can't guess.
Unless there is an ongoing random variation, it ought be a one off issue, although just possibly ' new 6.3% money' will also, if it ever is needed, turn out to be lower rate / up the queue from 'old' 6.3% money.
Obviously they're not sorting the queue on the value you can see. 8<.
|
|
|
Post by geoffrey on Oct 23, 2014 6:51:39 GMT
It might help us to understand this issue if westonkev could explain precisely how the Market Rate is calculated to two decimal places when all of the offers are made to one decimal place. Is it calculated as a percentile? As an average of all the money offered at the last match rate and below? And for the sake of clarity, we should ignore the lower uncompounded contract rate, as that always resolves to a single-digit compound rate. Another alternative I can think of is that the borrowers see an MR calculated in a different way to the lenders' MR -- do the borrowers see a "simple" interest rate when "fishing" on the market, for example?
Westonkev, the problem is not that anyone thinks you are lying. The problem is miscommunication. We all thought we understood a simple and transparent queuing mechanism, only to discover that it is actually far more complex. Your statement that there are actually several different Market Rates is really an example of putting the cat amongst the pigeons, and needs further explanation. If it is technically complex, get one of your IT people to write a precise explanation of how the 2dp vs 1dp matching works, and how a 2dp MR is calculated, and post it here for us. We also think that the YourRate functionality has now become misleading, and needs to be suspended and probably rewritten.
Thank you for staying on the board and remaining engaged with lenders.
|
|
pikestaff
Member of DD Central
Posts: 2,136
Likes: 1,484
|
Post by pikestaff on Oct 23, 2014 7:52:18 GMT
...Westonkev, the problem is not that anyone thinks you are lying. The problem is miscommunication. We all thought we understood a simple and transparent queuing mechanism, only to discover that it is actually far more complex. Your statement that there are actually several different Market Rates is really an example of putting the cat amongst the pigeons, and needs further explanation. ... I don't think it is complex. Broadly speaking, I think it works how we were always told that it works, and in normal circumstances I think all offers at a given rate will be equal (apart from the time made) and will queue correctly. I'm not going to change my mind about that unless I see evidence to the contrary. I think there must have been a glitch along the lines suggested by @gsv3miac which disturbed things. One possibility is that the real queue is at the uncompounded contract rate, but is displayed to us at market rate, and something happened to change the roundings or the number of decimal places used which led to the contract rate being slightly different for newer bids. I can think of a few other possibilities that could give a similar effect. Any one of them could have led to the explanation we have been given, especially if it was not fully understood by the IT guys or if something was lost in translation.
|
|
spiral
Member of DD Central
Posts: 908
Likes: 455
|
Post by spiral on Oct 23, 2014 7:59:47 GMT
The problem I have with the 2DP MR is that 9 times out of 10, it would not be a 1DP rate it would have to have a second digit of 1 through 9 and only 1 time in 10 end with a 0. This means that (depending on rounding) 50% of the time MR offers would jump YR offers (e.g 5.89% v 5.9%) and 50% of the time YR offers would jump MR offers (e.g. 5.9% v 5.91%). If this occurred on 9 out of 10 days, I would have expected it to have been raised before now. I only ever use YR and my money has only ever joined the back of the queue and never jumped ahead of any money in that queue.
|
|
oldgrumpy
Member of DD Central
Posts: 5,087
Likes: 3,233
|
Post by oldgrumpy on Oct 23, 2014 8:48:36 GMT
"I only ever use YR and my money has only ever joined the back of the queue and never jumped ahead of any money in that queue."
That is what I observe now and that is what I have always observed before last Tuesday, and that is what I expect to see in the future. Old Wappers and I both (at some time last week) jumped the Q.
|
|
|
Post by geoffrey on Oct 24, 2014 14:05:20 GMT
I think there must have been a glitch along the lines suggested by @gsv3miac which disturbed things. One possibility is that the real queue is at the uncompounded contract rate, but is displayed to us at market rate, and something happened to change the roundings or the number of decimal places used which led to the contract rate being slightly different for newer bids. I can think of a few other possibilities that could give a similar effect. Any one of them could have led to the explanation we have been given... Perfectly plausible. But we'll never know unless westonkev, or one of the IT guys/gals, tells us. Their staying silent just fuels the speculation...
|
|
wapping35
Member of DD Central
Posts: 385
Likes: 210
|
Post by wapping35 on Oct 24, 2014 16:37:52 GMT
Okay guys I called RS re my email and now have an answer which I believe I understand and it does explain what I saw. It also means there was not an IT glitch but that said the process is v complicated so I can also see why I thought it was a glitch. But maybe I am Mr Thickie... My understanding is as follows it is based on my call with RS's customer services today, who I must say were v helpful and thorough. i.e. I would encourage you to call, if you need further clarification: It seems that BOTH the MR and YR each day are truncated (rounded down) to 5.90% (the rate in question) to reflect a underlying simple interest number which is in a band. For 5.90% compound interest that band is 5.75% to 5.84% simple interest. That is 5.75% gets you to 5.903% and 5.84% to 5.999% . So both round down to 5.90%. The rate we select... When I set my YR at 5.90% on Oct 15th the simple interest rate was 5.84%. On Friday Oct 17th when I also had a YR 5.90% it reflected a 5.75%. Thus my Oct 15th 5.90% was stuck at the back. 5.84% is higher than 5.75%... On Saturday when I reset it to 5.90% (the same rate) whilst I did not know it, that did actually lower my rate from 5.84% to 5.75% so I moved ahead of OG who I assume was still at 5.84%….. OG you may wish to check to see if your matched Oct 15th order achieved on Sunday 19th was at 5.84%. So as I say I think I now understand what happened and as Kevin says it was not an IT glitch. Sorry Kevin. However I should say the email on Monday was not exactly that clear, but certainly RS's customers services lady was v good at explaining it to me. Anyway I hope this helps and bear in mind the above only reflects my interpretation of what I was told but to me it does fit what I saw. W35
|
|
oldgrumpy
Member of DD Central
Posts: 5,087
Likes: 3,233
|
Post by oldgrumpy on Oct 24, 2014 17:49:29 GMT
wapping35 confirmed 5.84% Nice wasn't she? Phoned me just as I was preparing this reply. No wonder Kevin says she and Steph are the best customer service people anywhere! And we've got to be nice to Kev! Sorry I gave you a hard time, Kev - the whole thing must have been a nightmare to explain convincingly. Oh, today's 6% market rate is at underlying 5.85% simple so it has hardly changed although we are in at 6%!
|
|
pikestaff
Member of DD Central
Posts: 2,136
Likes: 1,484
|
Post by pikestaff on Oct 24, 2014 18:02:56 GMT
wapping35 I understand the explanation perfectly but that does not mean that it is right! I agree that loans with contract rates anywhere in the band stated would map to a MR or YR of 5.9%. and that if this were so we could have a range of possible values underlying any given MR or YR. However, I do not believe that this should happen. Moreover, the explanation is implausible because, if the MR or YR did move about as described, you would expect the observed behavour to occur quite frequently and not just once. I also note that the lower of the two rates is the minimum rate needed to round down, after compounding, to 5.90% while the higher of the two rates compounds almost exactly to (but very fractionally below) 6.00%. That's odd in itself because what is the likelihood of the rate on successive days being at the extreme ends of the supposed 5.9% band, unless something else is going on? So I asked myself, do we ever have loans with intermediate rates? Not usually we don't. I've looked at a sample of my loan book. Of those that I have looked at, originated on a range of different dates:
all of the 5.9% loans have a contract rate of 5.75%, which is the minimum rate needed to round down, after compounding, to 5.9%. all of the 6.0% loans have a contract rate of 5.85%, which is the minimum rate needed to round down, after compounding, to 6.0%. The same is true for different rates. All of the loans that I have looked at have the minimum rate needed to round down, after compounding, to a specified single dp rate.* This is what I would have expected. There is no good reason for the loans ever to have an intermediate rate. The anomaly is that, for some unknown reason, the system did something different, for one night only. I suspect that the market rate that night should have been 6.0%, but a fault in the roundings somewhere led to it being translated to a contract rate of 5.84% instead of 5.85% and then reported back to us as 5.9%. I think it is significant that 5.84% compounded is so very close to 6.0%. It would be easy for such an error to have slipped in. The good news is that 6.0% loans are now being written with the usual contract rate of 5.85%. Whatever the error was, it seems to have been fixed. * Edit: My original sample was taken from loans with no lender fee, but the relationship also holds true for older loans where there is a 10% lender fee, if this is deducted before compounding.
|
|
oldgrumpy
Member of DD Central
Posts: 5,087
Likes: 3,233
|
Post by oldgrumpy on Oct 24, 2014 18:21:50 GMT
As was explained to me, this can happen again, especially when a large tranche of money takes a few days to clear from the market. It is purely down to the non revealed fact that a 1dp onscreen rate can actually be derived from different 2dp set rates (simple). How much better it would be if there was just one set rate (simple) for each 0.1% division (on screen rate). But there isn't. That's that.
|
|
wapping35
Member of DD Central
Posts: 385
Likes: 210
|
Post by wapping35 on Oct 24, 2014 18:57:43 GMT
Grumpy thanks for confirming the 5.84%. For me that really closes the circle. I was dreading you saying 5.75%.!!.(and RS did not know I was going to ask that question to confirm what I was told). Sorry Kevin I am sneaky... I would agree that this could happen again and indeed I have seen (or thought I had) a few strange queue things in the past, but they did not last more than a day and I did not track them. Unlike this one (4-5 days). I think this was exacerbated by the v large influx of money from the advertising (which certainly in London is v extensive especially on the Tube). I have had a PM (and i see a like) from Kevin indicating what I have said is a fair summary, but I take Pikestaff's point and it might be useful if Kevin could confirm explicitly.. I would also suggest /recommend that parties call RS Customer Services if they wish to clarify the position or at sceptical. Hey I was. On a final point if it does happen again you have the reset option I mention (what I did last Saturday) which will "jump" the queue. So in my view worth knowing. Cheers W35
|
|
pikestaff
Member of DD Central
Posts: 2,136
Likes: 1,484
|
Post by pikestaff on Oct 24, 2014 19:03:48 GMT
As was explained to me, this can happen again, especially when a large tranche of money takes a few days to clear from the market. It is purely down to the non revealed fact that a 1dp onscreen rate can actually be derived from different 2dp set rates (simple). How much better it would be if there was just one set rate (simple) for each 0.1% division (on screen rate). But there isn't. That's that. But there is, unless something goes wrong!
|
|
|
Post by GSV3MIaC on Oct 25, 2014 7:32:49 GMT
Yep, the RS explanation doesn't explain, at least not to me, how setting a YR of 5.9% can result in different MRs on different days, unless someone is changing summat in the background which we can't see. I agree there are lots of MRs which would give the same answer, but the one that you get should be deterministic, surely, not variable.
|
|
|
Post by geoffrey on Oct 26, 2014 10:54:00 GMT
The whole Your Rate functionality needs to be looked at. It shouldn't be behaving differently from setting a manual lending rate (for which no such anomaly has been observed). Westonkev has frequently said he doesn't like YR, because it doesn't do what people intuitively think it will. At the very least it should be made to function in the same way as setting a manual rate, so that it matches all possible MRs that resolve to the single-dp rate you set as our "minimum" rate.
|
|